G. Vetter-Leak-Free Pumps and Compressors Handbook-Elsevier Science (1996)

Preface
The increasingly strict regulations to avoid dangerous and hazardous

leakages have resulted in the past few years in a continuous drive for new
developments in the field of leak-free fluid handling machines such as
pumps, compressors and vacuum pumps. Much progress has been made
in the design of pumps, with centrifugal and positive displacement pumps
participating equally.
In leak-free pump systems, employing membranes, peristaltic tube
elements, bellows, canned m o t o r - o r mag-drives, the problem of
hermeticity is primarly solved by enclosing the working volume in contact
with the fluid by means of a functionally determined separating wall - a
membrane or a can in various configurations- serving as a static seal,
leak-free for all practical purposes, against the environment.
Various safety concepts, a very innovative field, signal any impending
disturbances of the leak-free operation of these pumps.
While leak-free compressors and vacuum pumps are less common than
pumps handling fluids, the solutions for leak-free operation are quite
similar to the provisions used in the design of pumps for fluids.
The German machine industry has contributed substantially to the
technical implementation of leak-free fluid handling machinery, since
strict requirements and regulations for the protection of the environment
have been in force for some time. The machine industry has responded to
this challenge and become the vanguard of world-wide efforts in this field.
Strong tendencies towards the development of leak-free fluid handling
machines can also presently be seen in the USA and Japan.
For the design, installation and successful operation of leak-free pumps
and compressors, the suppliers as well as the users need special expertise
and skill.
The operational limits and risks have to be evaluated, taking the
background of usually higher costs for installation and for energy during
later operation also into account.
The present book was compiled based on numerous lectures on the
subjects of safety via leak-free pumps and leak-free pumps and compressors under the guidance of the editor, in cooperation with the technical academy Haus der Technik in Essen, Germany.
viii
Leak-free Pumps and Compressors
The first edition of these papers published as a book in 1989 by Vulkan

Verlag in Essen, Germany was soon out of print. This stimulated the
editor and the team of authors into publishing a second edition with a
somewhat broader scope of subjects in 1992, again with Vulkan Verlag.
This small technical book, now edited and published in English by
Elsevier Science Ltd is intended to present the state of the art for leakfree pumps and compressors and to point out the key items to be
considered in the design, installation and operation of such machines.
Several well-known experts are contributing authors and accordingly
explain the subject from various points of view, but always based on actual
experience and proof of successful operation. Principles of functionality
are explained predominantly, however, no claims for exhaustive coverage
are made.
The editor hopes that this book continues to be well received: may it
contribute towards the advancement of leak-free production technology.
Many thanks are extended towards the colleagues participating as
authors and towards Elsevier Science Ltd, all of whom have substantially
contributed to the final success of this book.
Gerhard Vetter
CHAPTER
ONE
A survey of leak-free
centrifugal and positive
displacement pumps
G. Vetter
Terms and their significance

Process technology systems have to operate without dangerous leakages.
This requirement is very important for the safety of operation and for the
protection of the environment. Seals on shafts, rods or pistons have a
potential for leakages.
In the course of the last two decades a remarkable development towards
leak-free pumps has taken place.
Pumps with a working volume permanently closed hermetically against
the environment by means of impermeable walls are called 'leak-free'. The
absence of dynamically operating seals which are in direct contact with the
fluid in question and subject to pressure represents an important criterion
for leak-free pumps, which sometimes are also called 'hermetic pumps'*.
The absence of dynamically operating seals generally involves bypassing
potential trouble spots. Leak-free pumps are however not inherently more
reliable or economical due to this design feature.
The term 'leak-free' has also to be considered against the background of
the quantitative leakage rate of a system. In leak-free pumps, leakages can
no longer be detected by the bubble test method, a condition corresponding to a maximum leakage rate of 10 -4 mbar/s. With special care during
manufacturing and assembly, integral leakage rates of less than 10-6 mbar/s
can be achieved.
'Leak-free' systems should be regarded therefore not as hermetically but
as technically leak-free [1].
*Hermes Trismegistos, a person from ancient Greek history, was considered to

be gifted with magic power and able to make treasures and vessels inaccessible,
hence 'not hermetically but technically leak-free'.

The large and increasing importance of leak-free pumps stems primarily from the requirements of users in the field of process technology.
Manufacturers have responded to this challenge by the development of
leak-free series of pumps.
In reciprocating leak-free displacement pumps an elastic intermediate
wall - diaphragm, bellows, flexible tube - serves as the displacing device
and replaces the piston seal (Fig. 1.1). The mechanical or hydraulic actuating system remains completely outside the range of the fluid handled.
The material of the diaphragm and the hydraulic actuating system limit
the range of successful operation (<200~
In centrifugal or rotary displacement pumps the working volume is enclosed by the so-called 'can'
permitting the transfer of the driving torque by means of a magnetic field
resulting from a multiphase rotating electrical system or from mechanical
rotation of permanent magnets (Fig. 1.2). Such designs inherently expose
the bearings of the rotor to the fluid to be handled. Inability of the
bearings to run dry, sensitivity of the bearings against abrasive particles in
the fluid or incompatibility of the bearings with the fluid may be limiting
factors for such applications.
Figure 1.1 Leak-freereciprocating displacement pumps.
Leak-free displacement pumps
Figure 1.2 Leak-freerotary centrifugal and displacement pumps.
Peristaltic pumps with mechanical drive mechanisms for the rollers

(Fig. 1.3) occupy a special niche of the market. With regards to the engagement of the individual rollers and their consecutive partitioning of
the flow of fluid into discrete portions such pumps do not differ basically
from mechanically reciprocating diaphragms.
Figure 13
Rotary peristaltic pumps.
For fluids possessing special properties - presence of abrasive particles,

corrosive action, toxicity, potential for explosions, annoying odours, formation of deposits, low vapor pressure - as well as a possible sensitivity of the
fluid against components of the atmosphere and high requirements with
respect to hygienic problems, the use of leak-free pumps always merits
consideration [2-4]. Frequently such pumps are the only choice to avoid
acutely dangerous situations or for reliable and safe operation of a plant.
As a matter of fact, leak-free centrifugal and displacement pumps usually require higher initial investments than conventional equipment. Verification of future economical operation may sometimes be necessary,
particularly as all leak-free rotating machines inherently operate at somewhat lower electrical and hydraulical efficiency.
Sealing concepts
All efforts to improve the operational response of seals facing rotary or
reciprocating motion in conventional machinery cannot overcome the fact
of leakage flows being unavoidable in dynamically operating seals, in most
cases such small flows are even functionally necessary for step-down of
the pressure or for lubrication.
Pumps of leak-free design, avoiding dynamic seals in contact with the
fluid, are generally only economical where conventional sealing, flushing
or lubricating systems fail or are too expensive, complicated or unreliable.
Such circumstances may occur in conjunction with:
9 high pressure to be sealed off
9 fluids with low viscosity and poor lubricating properties
9 dangerous, toxic or abrasive fluids

9 fluids with high vapor pressure, if the seals are potentially in danger of
running dry
9 very stringent hygienic requirements
9 fluids sensitive to atmospheric influences
9 fluids unable to tolerate high local friction or shearing.
In the case of piston seals within reciprocating motion, liquid lubrication
and cooling action is provided in the sealing area subject to the highest
compression on the low pressure side. The plunger seal shown (Fig. 1.4)
is for example suitable for easily evaporating fluids (such as liquid carbon
dioxide). In such an application it is essential to provide blocking by flushing (item 6), preventing boiling off of the fluid and accordingly protecting
the piston against running dry.
Other important and well-known features of such plunger seals are
hard, smooth, geometrically precise plungers joined to the rod by a connection permitting some lateral motion. Guiding bushings (item 1) assure
aligned travel in the well chambered (item 3) packing.
In extreme cases, the high pressure seal has to be protected against
abrasive components in the fluid by forced flushing of a front side sealing
gap into the working chamber.
Various configurations have found acceptance for this purpose (Fig.
1.5). Most of them generate a sufficiently large flushing flow during the
suction stroke of the pump, keeping the throttling gap reliably filled with
homogenous flushing fluid. Various systems keeping particles at a safe
distance from the seal by means of sedimentation (Figs 1.5b and 1.5c)
have also been applied successfully.
For the reverse requirement of excluding contamination or infection
entering from the outside, sterilizing flushing of the seals (e.g. with wet
steam) creates safely operating sterile interfaces (Fig. 1.6).
At rotary seals on shafts of centrifugal or rotary displacement pumps,
quite similar provisions are used to improve the operational conditions. In
most cases the seals have to seal against lower pressures however.
Figure 1.4 High pressure plunger seal.
Figure 1.6 Sterile interfaces with flushing provisions.

Single stage centrifugal pumps Usually reach about 10 bar per stage. In
multiple stage pumps the pressure is, in most cases, lowered in front of the
shaft seal to the suction side pressure level, in conjunction with the compensation for the axial thrust.
In cases of high system pressure only, e.g. for the circulation of a fluid, a
high pressure difference is present permanently at the shaft seals. Well
proven designs for shaft seals with lower leakage (Fig. 1.7) are also based
on blocking systems providing cooling action, lubrication and pressure
equalization at the seal face.
Figure 1.7 ~;haftseals for centrifugal pumps.

The flushing systems require maintenance and regular checking, some
of them are quite complex and the fluid used for flushing has to be compatible with the process or else decomposition into harmless components
must be feasible.
In single stage centrifugal pumps the reduction of the pressure at the
shaft seal by hydrodynamic design measures vanes on the back of the
impellers or a supplementary bladed rotor wheel and constitutes a well
proven practice [5]. If modern, improved materials are used for the sliding
rings (e.g. silicon carbide), unflushed dependable operation of sliding ring
seals in abrasive-corrosive suspensions (e.g. washing fluid of flue gas
desulphurization plants) is feasible.
Developments in the sealing of high pressure differences, where single
or double stage blocking via flushing represents a well proven technology
(Fig. 1.8), nowadays show a tendency towards single stage blocking systems
Figure 1.8 Rotating mechanical seals for operation at high pressure.

for pressure levels up to 300 bar [6], with skilful utilization of the capabilities offered by modern materials (SIC) and the promotion of the formation of hydrodynamic lubricating films through a wavy surface shape of
the sealing surfaces.
Leak-free reciprocating displacement pumps

In this type of pump, displacement per stroke is directly achieved via an
elastic separating wall - diaphragm, bellows, hose. The diaphragm can be
actuated mechanically, hydraulically or pneumatically (Fig. 1.9).
In case of mechanical actuation, the diaphragm is exposed to the pressure difference between the discharge and the ambient pressure as an
external load, and additionally has to tolerate continuously the stress of
alternating deformation.
With hydraulic or pneumatic actuation, the diaphragm experiences
approximately balanced forces, since the pressures are bilaterally almost
equal. The diaphragm is exposed to the stress of alternating deformation
only, which is not determined by the discharge head.
The rating of leak-free displacement pumps nowadays ranges from 10
ml/h up to 500 m3/h, with applications for elevated discharge pressures
dominating.
Principal designs
According to the applications outlined in Fig. 1.9, for functional and economical reasons the following designs are used:
Metering pumps with mechanically actuated diaphragms are applied for

low volume flows and delivery heads. At very low hydraulic powers linear
magnetic drives are preferred (Fig. 1.10) as these drives can be easily frequency controlled. Mechanical drives with spring loaded cam and lostmotion stroke control have gained acceptance up to the maximum loads
handled with mechanically actuated diaphragms (Fig. 1.11). This design is
particularly economic in applications calling for occasional manual adjustment of the dosing flow.
Dosingpumps with hydraulically actuated diaphragms are used for higher
delivery heads, and are predominantly equipped with crank drives and
steadily adjustable stroke by variation of eccentricity. The hydraulic fluid
transfers the displacement of the piston to the diaphragm. Various proven
systems for control of the filling and for safeguarding of the hydraulic system are available (Fig. 1.12).
10
Fluid conveying pumps with hydraulically actuated diaphragms and

elevated delivery heads come in two principal designs:
Linear triplex process diaphragm pumps are predominantly equipped with
FFFE-diaphragms and corrosion-resistant materials. They feature speed
Figure 1.9 Leak-free reciprocating displacement pumps, application limits,

characteristic features.
11
Figure 1.11 Diaphragm metering pump with spring loaded cam drive
(ALLDOS).
12
Figure 1.12 Diaphragm metering pump with hydraulic diaphragm position

control - speed controlled or stroke controlled (LEWA).
control via frequency controlled asynchronous three-phase motors and are
primarily suited for homogenous fluids with low viscosity (also for fine grain
suspensions) up to very high pressures (500 bar) and loads (1000 kW).
Diaphragm pumps for slurries (Fig. 1.14) are mostly equipped with
elastomer diaphragm (sometimes in the shape of a hose). They run at low
speed and have a two or four cylinder arrangement, with double acting
pistons on the hydraulic side actuated by straight thrust crank drives.
Pumping of slurries requires easily accessible pump valves of rugged
design. In slurry pumps for very large loads (1500 kW) control of the
diaphragm position is performed via electric sensors to avoid any
disturbances (Fig. 1.15).
Diaphragm design
For diaphragms and bellows actuated directly mechanically, an empirically
determined optimum for the degree of stiffness of the membrane, support
against the pressure difference, service life, linearity and elasticity of the
characteristic with respect to the discharge pressure has been attained. As
13
Figure 1.13 Large process diaphragm pump for transporting of dangerous fluids
against high pressure (LEWA).
a matter of fact, for several reasons the service life of the mechanically
operated diaphragm is limited to a few thousand hours, a condition satisfying most applications however. All kinds of elastomers (lined with PTFE
also) are used as material. Recent developments using diaphragms composed of several PTFE layers not only offer reliably longer service life
(around 10000 h), but also implement a modern safety concept.
To resist corrosion, diaphragm pumps for dosing and process technology
are almost exclusively equipped with PTFE diaphragms. Based on the
peculiarities of deformation and stress as well as for optimization of the
deflection, thin fiat diaphragms and thicker, undulated diaphragms are
nowadays successfully in use (Fig. 1.16). Clamping is performed using
limited compression, i.e. with a defined amount of compression in the
14
Figure 1.14 Diaphragm pump for slurries with a tube-shaped diaphragm
(ABEL, FELUWA).
clamping area. To assure good clamping and to avoid leakage, the clamping surfaces are usually grooved. In cases involving very high pressure, provisions for pressure equalization from the hydraulic system towards the
sealing/clamping area may be advantageous.
According to the present experience, a service life ranging from 10000
to 20000 h can regularly be expected for PTFE-diaphragms (<15&C)
even in applications involving high discharge pressures.
15
Figure 1.15 Reciprocatingdisplacement pumps for slurries (WIRTH, GEHO).

Thanks to the greater degree of deflection possible with elastomer
diaphragms, these are attractive for use in pumps for slurries which normally show only little corrosive action. A thicker diaphragm facilitates
clamping, where limited compression is being used (Fig. 1.17).
The effects of particles in the clamping region can influence the service
life of diaphragms, particularly if the numbers of coarse particles in the
fluid are increased. In such cases the use of tubular diaphragms (hoses)
with their internal flow (Fig. 1.17) may be advantageous.
Figure 1.16 Elastomer diaphragms: (a) plane-parallel diaphragm, (b) undulated

diaphragm, (c) limited compression, (d) clamping with pressure equalization on
the hydraulic side.
16
!~176
I[ c/
a)
Figure 1.17 Freely deflecting elastomer disc and tube-shaped diaphragms,
clamped with limited compression: (a) conical clamping area, (b) clamping
area with grooved profile, (c) tubular diaphragm clamping area with O-ringprofile, (d) diaphragm clamping area with O-ring-profile (PTFE- coated).
Properties and characteristics

The well-known properties of reciprocating displacement p u m p s - linear
pressure-stiff characteristic adjustable by frequency or stroke and good
energy efficiency- need not be elaborated in detail again here [7, 8]. Special
features of diaphragm type pumps with respect to leak-free design are:
(a) Insensitivity against running dry, a property not available in any other
design of leak-free pumps.
(b) Additional elasticities present in the working volume (larger dead
space, the influence of elasticity of mechanically actuated diaphragms)
cause a slightly lower volumetric efficiency in diaphragm type pumps.
(c) As a consequence of the additional elasticities, larger pulsation amplitudes can commonly be expected for diaphragm pumps, as compared
with piston pumps. The installation design must allow for these
pressure pulsations [9].
(d) For comparable conditions, the energy efficiency [8] appears to be
highest for leak-free diaphragm pumps, particularly for partial loads.
Leak-free diaphragm pumps feature lower losses due to friction at
the seals than plunger pumps and accordingly show better efficiency.
(e) Hydraulic diaphragm pumps generally demand more attention to
the suction conditions of the pump. In addition to the pressure drop
in the suction valve determining the NPSHR, other pressure drops
occur in the hydraulic actuating system. While the vapor pressure of
the fluid solely determines the limit for cavitation, release of gases
from the hydraulic fluid may be another limiting factor on the suction
side.
Leak-free displacementpumps
17
(f) If attention is paid towards suitable suction conditions (unobstructed

input of the fluid, preferably in conjunction with some positive pressure) and reliably operating precise venting systems to eliminate gas
bubbles in the hydraulic system are provided, the metering accuracy
of hydraulically operating membrane type pumps is approximately
equivalent to piston or plunger type pumps.
Leak-free centrifugal and rotary displacement

pumps
The well-known designs of pumps with canned motors or magnetic drives
(couplings) using permanent magnets are characterized by the can providing hermetic encapsulation of the pump (Fig. 1.2). Magnetic forces act
through the can on the rotor driving the pump. The can does not participate in the transport process, but may reduce the efficiency of the pump
by causing additional frictional and electrical losses. The can is subject to
the total pressure difference as a mechanical load, and therefore needs to
possess some rigidity, i.e. sufficient wall thickness, while on the other hand
the gap between the stator and the rotor of an electric motor or between
the outside rotating set of permanent magnets and the inside set of permanent magnets should be as small as possible to maintain reasonable
efficiency of the configuration.
In pumps with a canned motor, the rotating magnetic field of the stator
generated by a three-phase electrical power supply acts on the rotor of the
motor through the can. The cylindrical outer surface of the can rests snugly against the bore of the stator and can accordingly be supported by these
means even against very high pressure differences.
In the case of magnetic drives the magnetic field rotates mechanically
and acts on the interior set of permanent magnets. Some clearance has
therefore to exist between the can and the sets of rotating magnets on
either side of the can. The can is not circumferentially supported and has
to withstand the full pressure difference without showing any deformation
which might interfere with the rotation of the two sets of magnets.
For both configurations, the leak-free design inherently requires the
bearings of the rotor-impeller combination to be located in the fluid
present in the working volume of the pump. A bypass flow derived from
the discharge side of the pump passes the bearings and the gap between
the rotor and the inner surface of the can, expected to provide lubrication
and cooling of the bearings as well as of the can.
18
Principal configurations and properties

Basically all types of pumps with a rotating fluid transporting mechanism
can be operated leak-free in combination with a canned motor or a magnetic drive. The maximum power of such pumps is not inherently limited,
but most applications are predominantly located in a power range lower
than 200 kW. The overall economy of such machinery needs to be
carefully evaluated.
At the present time, canned motors are almost exclusively used in
combination with just centrifugal pumps. The possibility of supporting the
comparatively thin walled can allows operation at very high system
pressures. A partial flow of fluid (Fig. 1.18) removes most of the electrical
losses (eddy currents in the can, stator and rotor losses).
Fluid lubricated bearings cannot run dry. Evaporation of the fluid in the
working volume may cause adverse loads on the bearings and needs
therefore to be avoided.
Canned motor pumps therefore require various safety measures, with
protection against explosion being most important. Manufacturing demands extreme care and lots of experience. Particular attention has to be
paid to smooth and complete support of the can and reliable assembly of
the can (dependable and leak-free welding seams).
Even on rupture of the can, canned motor pumps stay leak-free. The
housing of the motor in conjunction with the stator can be designed to
serve as a supplementary sealing shell. Sensors provide dependable monitoring of disturbances.
Figure 1.18 Routing of the partial flow of fluid in a canned motor

(HERMETIC).
Leak-free displacementpumps
19
Canned motor pumps have a compact configuration and can be adapted to serve special applications. Such pumps are less mass produced, but
custom designed in most cases.
Magnetic drives may also be used for any type of pump with a rotary
fluid transporting element without an inherent upper limit in the rating.
A standardized conventional electric motor is used, the losses occurring
in the motor stay outside of the working volume of the pump. The losses
due to eddy currents occurring in the can have to be removed by a suitable
bypass cooling flow of fluid. Since these losses rise as the gap increases
and as the can has no circumferential support, mag-drives are preferentially used for low pressure differences. Development of non-metallic
cans (plastics, ceramic materials) not subject to eddy currents is progressing (Fig. 1.19). Usually such cans improve the efficiency, although manufacturing expenses may increase.
For explosion protection, attention has to be paid to the heating of
metallic cans, even though presently magnetic couplings are not subject to
the explosion protection regulations. In spite of that, magnetic drives
should only be operated below the temperature limits as spelt out in the
classification requirements.
Figure 1.19 Plastic lined centrifugal pump with magnetic drive and plastic can
(RICHTER).
20
In case of can rupture, the outer housing normally serves as a temporary

supplementary sealing shell, requiring a dynamic auxiliary seal (Fig. 1.19).
Permanent-magnet couplings are inherently torsionally elastic and feature a characteristic slipping or breakdown torque.
The drive has to be designed taking the moments of inertia and the
typical shock-factors into account to assure reliable start-up and operation at all times.
Efficiency
In contrast to reciprocating displacement pumps, leak-free design in centrifugal or rotary displacement pumps usually requires the expenditure of more
energy. This energy is mostly necessary to cover the additional hydraulic
losses, which may get offset partially by the absence of shaft-seals friction.
Additional energy is required furthermore due to eddy current losses in
metallic cans. For leak-free design, the total energy balance may rise by
5-10%.
In pumps with canned motor the input of heat into the fluid is still
larger, since the losses in the rotor of the motor (due to hysteresis, eddy
current and ohmic losses) are discharged into the fluid. It can happen that
about 10-20% of the total power consumed may heat up the fluid, if no
hydrostatically cushioned auxiliary cooling loop for the fluid is provided.
By and large, the overall efficiencies of pumps with canned motor or
magnetic drive show little differences in case of low fluid temperatures.
According to the latest tendencies, for higher fluid temperatures, pumps
with magnetic drives seem to show a slight advantage as compared with
pumps with canned motors.
Safety concepts
In addition to the observance of the explosion protection regulations, provisions for monitoring the temperatures, dry running of bearings, vibrations and sometimes of the actual rotor position may be necessary or are
at least recommended.
For reciprocating diaphragm pumps of high power, monitoring of the
temperature in the hydraulic fluid is advantageous and monitoring for
excess vibration recommended.
For safeguarding leak-free operation, immediate detection of diaphragm
or can rupture or of leakage through pores in these components is decisively important. In diaphragm pumps, the use of sandwich type diaphragms
(Fig. 1.20) represents the long-term proven state of the art [2]. During
operation, the layers of a sandwich diaphragm are coupled together
21
Figure 1.20 Sandwich type diaphragm with hydraulic coupling of the sandwich
layer diaphragm.
hydraulically in conjunction with a check valve (Fig. 1.20, item 4) [10]. The
rupture monitoring is performed by a pressure sensor.
Recently, sandwich type cans for pumps with canned motors or magnetic drives and various provisions for leakage detection have been introduced successfully (Fig. 1.21).
For many years with canned motor pumps the housing of the motor
together with the stator served as a secondary sealing shell to provide safety. But even when sensors for the detection of can leakage or rupture were
provided, when leakage occurred the stator of the motor was in many
cases adversely affected and the repair of corrosive damage was expensive.
In magnetic drives, the outer housing can be utilized as a secondary
safety shell only if a dynamically operated and normally dry running seal
is included on the shaft between the electric motor and the outer magnet
assembly. However, such a seal involves reduced safety and reliability.
The introduction of compact sandwich type cans therefore represents a
true step towards increased safety [1]. However, the necessity for eventual
safety provisions and the safety level required should always be critically
analysed.
22
Figure 1.21 Canned-motor pump with sandwich-can and rupture monitoring

instrumentation (HERMETIC).
Bibliography
1. Grundlagen der vacuumtechnik; Berechnungen und Tabellen. LeyboldHeraeus Druckschrift O4.1.1/1143.03.80 GK 5.D.
2. Vetter, G." Stand und trends bei der entwicklung leckfreier oszillierender verdr~ingerpumpen, Chem.-Ing.-Tech. 37, No. 3, 218--229 (1985).
3. Kr~imer, R. and Neumaier, R.: Centrifugal pumps and rotary positive displacement pumps of hermetic design, Hermetic Pumpen GmbH (1988).
4. Lehmann, W. and Vollmers, W.: Auslegung von kreiselpumpen mit permanentmagnetischen antrieben, Pumpentagung Karlsruhe (1988).
5. Vetter, G.: Pumpen for Rauchgas-reinigungsanlagen - anforderungen und
erfahrungen-, Komponenten zur rauchgasreinigung, Sonderteil BWK 10,
K20-K27 (1988).
6. Victor, K. H.: Kontakt-und verschleiBfrei arbeitende elasto-hydro-dynamik
(EIfD) - gleitringdichtungen fiir den hochdruckeinsatz, Pumpentagung Karlsruhe (1988).
7. Vetter, G.: Pumpen-bauelemente der anlagentechnik, Jahrbuch I, VulkanVerlag (1987). Fritsch, H.: Pumpen in prozeBanlagen, 3R international 27,
No. 7, 485--493 (1988).
8. Vetter, G.: Zum Kenntnisstand der rechnerischen bestimmung von druckpulsationen durch oszillierende verdr/ingerpumpen, 3R international 27, No. 7,
468-475 (1988).
9. Vetter, G.: DBP 18 00 018. Neumaier, R." DE 36 39 720.
CHAPTER
TWO
Properties and design

criteria for magnetic
drives on pumps
W. Lehmann
Introduction
The commonest method used today to seal the product side of a pump
shaft from atmosphere is by a packed stuffing box or a mechanical seal.
The radial and axial sealing surfaces have to be lubricated or cooled with
liquid, often the media being handled by the pump and generally it is
impossible to prevent the sealing fluid escaping to atmosphere through
the shaft seal. Statistics indicate that seal failures cause the highest number of problems on plant equipment and that particular attention should
be given to leakage rates.
In the past the main object of sealing a shaft was to reduce leakage as
much as possible however present demands particularly in the chemical
and process industries are to prevent any leakage whatsoever. This can be
achieved by using glandless pumps. The most common types are canned
motor pumps and magnetically driven pumps.
The state of current technology

The characteristic of the canned motor pump is the internal construction
of the motor, pump and seal. The drive is transmitted through the electromagnetic field to the shaft which is common to both pump and motor. On
a magnetic drive machine, the pump shaft is separate and is sealed by a
can. The drive force is transmitted from the motor to the pumps shaft by
the magnetic drive.
It is not the intention here to assess both systems which have been in
use for many years. However, the differences in their basic construction
have had a decisive influence on their commercial usage.
24
For example, canned motors dominated the market only whilst magnetic drives were available with bulky aluminium nickel cobalt (AlNiCo)
magnets which had to be fitted with a safety component to prevent
demagnetisation. Due to the development of magnetic materials based on
the rare earth samarium and cobalt (SmCo) it is possible to offer compact
pumps with permanent magnet couplings which, when properly designed,
do not have these disadvantages. The significance of these pumps in the
market place has increased considerably as a result of this, therefore the
physical and operational qualities of the magnetic drive pump are presented here for the plant designer and end user.
Design
Magnetic drive
A magnetic drive assembly (Fig. 2.1) comprises the following components:
9 outer magnet
9 can
9 inner magnet.
Figure 2.1 Principlesof construction of a magnetic drive system.
Magnetic drives on pumps
25
The magnet is connected to the motor shaft by the half coupling. The
motor torque is transmitted via the magnetic field to the inner magnet
which is coupled to the pump shaft.
Permanent magnet drives operate without slip, i.e. motor and pump
speed are the same. A stationary can, made of non-magnetic material (e.g.
high grade stainless steel, ceramic or plastic) fitted between the inner and
outer magnets, forms the pump seal.
Since the 1970s, samarium cobalt has been the main magnet material
used for more powerful magnetic drives. Since 1980 heavy duty heat
resistant couplings made from the magnetic alloy SmECOl7have been
developed.
Magnetic drivepumps
Most types of pumps with conventional shaft seals can now be fitted with
magnetic drives, as shown in the examples (Figs 2.2 and 2.3). Normally
not only the hydraulics but also the pump connection dimensions can be
maintained. Designer and users can therefore select according to the specific requirements of the application or a standard pump combination can
be replaced retrospectively, if necessary.
A characteristic feature of glandless pumps is that the impeller is supported inside the pump chamber. Product-lubricated sleeve bearings
made from abrasion and chemically resistant materials (e.g. SiC) have
proved successful for this design.
A partial flow through the drive assembly is controlled so that the heat
generated by the magnetic coupling is dissipated into the pumped media
by the liquid flow into, and out of the can. Different arrangements are
possible, for example externally filtered liquid can be fed into the bearing
or drive assembly.
It should also be mentioned that magnetic drives are used on agitators
and other rotating machinery (Fig. 2.4) as a safety feature. Thus a definite
contribution towards improvement of the environment has been achieved
by a development in the field of engineering.
Characteristics of a magnetic drive
Torque
To understand the dynamic and stationary operating behaviour it is necessary to have a closer look at the internal structure of the external and
internal magnetic rings (Fig. 2.5). They are fitted with individual magnetic
plates which are magnetised radially and have alternate polarities. This
r-
Figure 2.2
Types of pump with magnetic drives: side channel and in-line pumps (SIHI).
Figure 2.4 Ball valve with magnetic drive.
27
28
Figure 2.5 Modular structure of a permanent magnetic coupling: internal magnetic ring, isolation shroud, external magnetic ring.
means inevitably that only an even and equal number of plates is feasible
on the short circuit rings which bear them.
When not under a load, opposite magnetic poles of the two halves of
the coupling face each other (Fig. 2.6). The field lines run radially between the poles and are closed through the short circuit rings.
Figure 2.6 Characteristic torque of a magnetic drive.
29
If a torque is to be transferred the external magnetic ring must be rotated relative to the internal magnetic ring by a phase angle of tp. This tensions the field lines around the circumference, rather like elastic bands.
The transferable moment increases with the phase angle and in fixed position reaches its maximum value of
~ = + ~/p
the static nominal moment of a particular size. Because the structure is
symmetrical the direction and level of the torque is independent of the
direction of the load. In the range
- 7 r i p < ~p < 7r/p
the coupling is stable in operation, i.e. the moment and angle of torsion
increase evenly. This is the 'normal' operating range of a magnetic drive.
Although the external and internal magnetic rings rotate out of phase,
they rotate synchronously as required. If the load moment exceeds the
maximum value, like poles increasingly run over each other. This weakens
the radial magnetic field and therefore the transferable moment.
In practice this means that the magnetic drive is 'breaking off'. The two
halves of the coupling run with considerable slip. Since almost no torque
can then be transferred, the drive motor runs in a condition similar to
idling. The pump follows with just a few revolutions and cannot be started
synchronously again until the drive has been brought to a standstill. Under 'normal' operating conditions this state is undesirable, but at the same
time it offers an overload protection for the motor and pump in a similar
way as that for a hydrodynamic coupling.
This characteristic behaviour of a magnetic drive demands that the
pump/magnetic drive/motor combination is of synchronized design (Fig.
2.7). This not only means that the expected stationary load moments can
be transferred when the pump is in operation, but also the dynamic moment peaks, e.g. when starting. In general it can be said that the nominal
moment of a magnetic drive must be larger than the maximum expected
load moment. It must also be taken into account that increasing the size
not only causes higher construction costs but, if metal isolation shrouds
are used, this would increase eddy current losses. The latter are predominantly determined by the 'installed' magnetic field and only partly by
the respective operation point.
Efficiency
It is not possible to determine the power rating requirement of a pump with
a magnetic drive simply from the motor arrangement of a conventional
30
Figure 2.7 Volutecasing pump to DIN/ISO 2858 with magnetic drive (Halberg).
design. If we assume the same hydraulic parts and the same point of
operation, the rating balance is rather different (Fig. 2.8).
As a comparison, curve b is almost identical to the rating requirement
of the pump with a mechanical seal or stuffing box.
Although there is no friction and no flushing of the shaft seal, and possibly also no losses from the external bearings and flexible coupling, there
are however, not only the comparative losses from the sleeve bearing and
the magnetic drive, but also, if an electrically conductive isolation shroud
is used, there are eddy current losses.
According to the induction law the synchronously rotating magnetic field
induces an electrical field in the conductive wall of the isolation shroud
whose magnetic forces point in the opposite direction to the direction of
rotation. The resulting counter moment is proportional to the speed and
satisfies the following equation:
M~ = C1. n.
The following then applies to the power loss caused by eddy currents:
Pw = C . n 2.
31
_~, # ~ i ~ ~ "
I.
'"
---4--0
Figure 2.8 Influence of the isolation shroud material on the power requirement:
(a) metal, (b) ceramic, plastic.
As we can see the moment is directly proportional to the speed, whilst
the power loss is proportional to the square of the speed.
For practical use this means that the efficiency of a magnetic drive is
reduced as the speed increases:
P~
PK +C'n2
,M K
MK +C.n
As mentioned above the transferable moment and the magnetic field

change with the angle of torsion, so that some sort of dependency on the
moment to be transferred can be expected as well as the dependency on
the speed.
Using tests [1] it has been shown that the static moment of a magnetic
drive and the reaction moment of the eddy current can be described approximately by means of a trigonometric function.
The static moment can be described by
MK = 3/,,=. sinPtp
2
32
and the reaction moment by

Mw = Mw
"
9 cos p-tp.
4
The linear dependency of the maximum reaction moment has been

confirmed as being
Mwma = C . n.
The operating and limit values shown in Fig. 2.9 are produced for fixed
points in the operating range of a magnetic drive. If, for one model the
constants Mrm= and C have been determined, the efficiency can be calculated dependent upon the speed and the load point. For a particular
magnetic drive with magnetic plates made of SmCo and an isolation
shroud made of material 2.4610 with a wall thickness of s = 2 mm this
dependency is shown in Fig. 2.10. This means that the drive rating for a
required operating point can be calculated in a simple manner.
Isolation shroud
The stationary isolation shroud (Fig. 2.11) serves to seal off the outer
magnetic drive on the drive side. The magnet assembly revolves around
M. ~ /
0.70MK.x-~
o.87~: : : ~
I/
'
',
x
;'-..
',
I
0
...E_
2p
"~\
~p
I'
I"- OPERATING -'1

RANGE
Figure 2.9 Operational and limit values of a magnetic drive.
33
OPERATING RANGE
OF MAGNETIC DRIVE
81
1000
1500 1800
l/rr/n
3000
3600
Figure 2.10 Efficiencyof a magnetic drive with a metal isolation shroud.
Figure 2.11 Deep-drawn isolation shrouds of modular design.
the shroud with a safe radial clearance of about 1 mm. The shroud is normally made from a non-magnetic metal and it is produced by deepdrawing, without welding seams. By means of the flange on the pump side
the isolation shroud is attached (clamped) to the pump casing.
In standard applications, high-grade material 2.4610 (Hastelloy C-4)
has gained wide acceptance for the isolation shroud. The wall thickness
should not be less than 1.6 mm.
34
Heating
As an efficiency of 100% can theoretically only be achieved with electrically non-conductive materials, the heat loss of an electrically conductive isolation shroud must be dissipated continuously. This is normally done by
means of an internal partial flow, which is taken from the main service liquid
flow, guided along the isolation shroud in the form of a coolant flow and fed
back into the main flow at a point of low pressure. If the isolation shroud is
carefully selected to match the size of the magnetic drive, the heat loss can
be dissipated without a major increase in the partial flow temperature.
Figure 2.12 shows the expected temperature increase of the partial flow
as a function of the magnetic drive rating. With an increase in the flow
rate of the main product flow the temperature of the partial flow only goes
up marginally.
This conclusion is derived from the fact that centrifugal pumps normally show a steady drop in the head characteristic with a drop in pressure
differential and thus a reduction of the internal partial flow.
The general equation used to determine the counter moment produced
by eddy currents allows us to arrive at the steps to be taken to keep the
power loss, and with that the temperature increase, as small as possible.
As an example, the following equation governs the material and the
wall thickness of the isolation shroud in the area where the magnetic field
transfer takes place [2]:
b
Mw = C . ~5
- 100
oo
Ap
/
/
4-
-80
//-.-A~
A'O
%
-
GO
40
....~
-'~
Ap
10
m3/h
20
1-
Figure 2.12 Temperature increase of the partial flow.
i ill
15
35
where C is constant (geometry, physical data of the magnet material,

speed), b is the wall thickness of the can, and~p is specific resistance of the
can material.
This example shows that the moment, and hence the power loss, is
directly proportional to the wall thickness of the can. Unless other boundary conditions dictate a different course of action, one would normally try
to keep the wall thickness as small as possible, consistent with the mechanical load. Compared to the other casing parts, however, this would render the isolation shroud the weakest member of the assembly.
Despite an increase in initial costs, the use of a highly corrosion-resistant material at least in that part of the isolation shroud where the magnetic field is imposed, is therefore generally recommended.
With this type of material one not only achieves an increase in the corrosion-resistance, but also a reduction of the eddy current losses. Assuming otherwise identical conditions, a high quality alloy material (2.4610),
in comparison to stainless steel (1.4571), on account of its higher specific
resistance of
8(2.4610) = 1.25 flmm2/m
t~(1.4571) = 0.75 f~mm2/m
leads to a 40% loss reduction:
M,,,2.4610 -
0.75
" - - - M w ! 4571 = 0 . 6 M w
1.25
14571,
If the partial flow in the magnetic drive is expected to cause wear,

because it contains abrasive solids, this will invariably lead to larger wall
thicknesses and hence to a loss increase. At this point it is worth considering whether corrosion-resistant, wear-resistant and non-loss isolation
shrouds of ceramic materials (Fig. 2.13) are a better choice and a more
economical alternative.
Pump/magnetic drive/motor drive system

Substitute system
For the purpose of roughly calculating the required moment of a magnetic
drive, it has proved useful to reduce the pump/motor set to a simplified
mass acceleration system. Friction and venting losses may be neglected in
this case.
As shown in Fig. 2.14, the correlation between air gap moments in the
36
Figure 2.13 Ceramic isolation shrouds.
MOTOR
MAGNETIC
DRIVE
PUMP
MA
MP
MM
JAM I
j~
Figure 2.14 Mass system motor/magnetic drive/pump.
motor MM and those in the magnetic drive ]VIA may be expressed in the
following equations:
MA =MM-(J g +Jau) Or~

Oz
M . = (J. + JAM)
Oto~,
Ot
+ (J~ + J~) =='~ + Me.
Ot
For a defined time interval, it is assumed, as a further simplification, that
37
Oto~, _ Otoe
Ot
Ot
From this, the following equation can be derived to express the correlation between the two air gap moments:
[1
MA = 1
1
g~,
1 + (J p / JM),o,~
- MM
With this equation the moments actually developed in a magnetic drive

may be determined for any given load condition (Fig. 2.15).
Start-up torque
Here breakaway torque M~t4 is substituted for motor torque MM.
The hydraulic moment MpA is still zero at this point. For this duty point,
the following equation is then true for the breakaway torque MAa in the
magnetic drive"
M~ = 1 1
M~
l + ( Jp / JM),o,~
t'In
~7 ~
,,
HA
20
~0
60
80
Figure 2.15 Air-gap moments of motor and magnetic drive.
100
38
Pull-out torque
When running up the motor, the maximum or pull-out torque is reached
at motor torque MMK.The hydraulic moment already present at this point
may be derived from the relation governing the stationary duty point as
follows:
Mex =MpB" n ~
n8
From this it follows that:

1
/ {)/
. 1_ M~,
1 +(.#,./JM),.,
M~,
,,,,,,
,,
,,
nx
nn
Stationary duty point torque

This point is defined as the point where n = riB, and
M M= M ~ = M ~
M___m.= 1.
M~
Here, the maximum running torque MMB should be taken to be equal
to rated motor torque MMN.
Design torque of a magnetic drive

In order to select the correct size of magnetic drive for a pumpset, we not
only have to know the characteristic torques of the motor and pump, but
also their moments of inertia.
For the purpose of comparison, it is useful to relate the various torques
to the rated torque. For simplicity's sake, the following is assumed for
standard three-phase induction motors with an average power rating.
MMA - MMK = 2.5 MMN.
For the purpose to this review of the limiting values, the operating point
can be expressed as maximum Men = MMN.
Furthermore, it is assumed that nr-- 0.8 nn. Based on these assumptions,
39
the following equations may be derived for the given points:

Breakaway torque"
2.511_1+ (Je 1/ JM),o,.,j.
MM,
Breakdown torque:
1
[1- MMN 90.64 ])
l+(J,/Ju),~.,
2.5MMN)
MMNM~ 2.5(1=
M
0.74
].
= 23|1 MM.
L 1 +(Jp IJM ),o,~
AK
Stationary operating point torque:

-M-~ =1.
MuN
These relations are shown in Fig. 2.16 (shaded area) as a function of the
correlation between the moments of inertia of the pump and motor.
05
MA
,,
/////"/"I"
" shock
Sta"rting
/ / No
risk
of /break-off
i ~~Ar
/
/
/
/
/
/
.
Breakdown torque
2.0
""
1.5
1.0
m=lm
0.5
.ira.
~ n l ,
,llll,
mmll
Iilll
Stationary
Risk of break-off
i
re=ill
~ Starting torque
operation
i
i,
ii
/
O
n--.
~,
'I
'
(~P) total
Design range
Figure 2.16 Influenceof the moments of inertia on the required air-gap moment
of a magnetic drive.
40
It is a known fact that three-phase asynchronous motors never start up

without a shock or free from vibrations. For a number of shaft couplings
connecting the driving engine and driven machine, shock factors are given
[3] as multiples of the indicated starting torque. This multiple is of the
order of 13 for a rigid coupling, approximately nine for a flexible coupling,
and approximately three for a belt drive.
Tests on a self-oscillating substitute system (Fig. 2.17) have shown that
magnetic drives are also governed by a specific shock factor [4]. According
to [4], magnetic drives with an average power rating can be expected to
have a starting shock torque MMSof approximately
MMS =
1.5 MMA
and oscillations due to excitation equivalent at maximum to the mains frequency (Fig. 2.18). Although this is a clear indication of the attenuating
effect of a magnetic drive, the load increase during start-up has to be
taken into consideration for its design.
With a shock factor of 1.5 the following real conditions are obtained on
start-up:
M us = 1.5.2.5. MMN
MMN =3" 1 -
1]
l+(Je;JM~o,~,
This curve has also been plotted in Fig. 2.16. It results in areas of (Je/Jm)
total where the minimum required moment of a magnetic drive is
TACHOMETER
GENERATOR
TORQUE
MEASURING HUBS
NM
MAM
MOTOR
MAGNETIC
DRIVE
Figure 2.1"/ Standardtest-bed arrangement.
TACHOMETER
GENERATOR
MAp
Np
PUMP
41

3276814 TBI:
25u.~ TB2: OFF
88" 5,'11 15'11".".5 EIID OF OUTPUT (IIBR. & ENTER)
CURS
ORG
:,': TIHE
14.400 m:~
0 . 0 0 0 ms
--Rt
CHiJlrdII
q.el2 U
235,000 mIJ
- - B l CH 2.001
300.000 mV
50.000 mU
~UED CHFoIIHELS -" NO CHI~INEL SAVED
32768W TBI:
2~d.LS TB2: OFF
ill
-X:
-.-.-Rz
----B:
~P,~D
DISP--RFINGE
0.000,,,
-3.1q1.,.
0.000...
~00.000
m~
7,097 U
20.q75 U
CUR dP
88/ 5/11 15:11:35 BiD OF OUTPUT
576
(NBR. g EHTER)
"
' OR8
0.000 ms
~80.000 mO
CHI~Iilm
q 3 0 . 0 0 0 m1,.I
15.000 ~
CHFr
." NO CHFINMEL SflUEO
TIME
CH 3.001
DIFF
" EE X P
14.400 ms ~ 1
t4.577 V ~ 1
270.000 mU ~ I
CURS
'
20.800 ms
89
mU
DIFF
....i ~ x P
20.800 m~ ~. 1
7].750 mU $ 1
q15.000 mO ~ 1
Ol
SP--RI~I"N'GiE
0.000...
800.000 m~
-I.571...
3.SqB U
0.000...
20.t~75 O
CUR JP
832
Figure 2.18 Dynamiccurve of the torque at the motor and pump ends during the
starting phase.
determined either by the stationary operating point, by the breakdown
torque or by the starting shock.
To ensure trouble-free operation, the admissible torque of the magnetic
drive has to be greater than the rated motor torque by the given moments
of inertia. If a maximum torque to be transmitted is not located in the top
section of the shaded operating range, the magnetic drive must be expect-
42
ed to break-off. This can, for example, be the case during start-up, as is

shown in Fig. 2.19. As the motor is running up to its unloaded speed, the
output speed on the pump side only reaches several revolutions. Due to
this asynchronous running of the internal and external magnetic rings,
327~8H TB1,
25us TB2" OFF
X: TIME
-----R: C l - l ~
----B= CH 2.001
SRUED CHRI"tfiELS :
e~," 5/13
CURS
ORG
lq.q00 ms
0.000 ms
6.295 U
-672..500 n,U
315.000 mU
Lt0.000 mU
HO CHI:H'IEL SRUED
_-:-.2768W TBI: 25us TB2= OFF
7'20= 0 EhlD OF RECORD (HBR. g EHTER.'.,
DIFF
EXP
lq.q00 ms ~ 1
6.967 U ~ l
275.000 mU ~ 1
88/ 5/13
DISP--RFINGE
0 . 0 0 0 . . . 800.000 ms
-2..~0...
7.737 U
0.000...
20.=f75 U
CUR, dP
576
7:20t 0 Eli* OF OUTPUT (MBR. & EHTER)
np
~
X: Title
- - R " CH 3.001
~B'. CHIH~]B
SRUED CHRtftELS =
Figure 2.19
out.
,,,,,,
CURS
0RG
27.200 ms
0.000 n,s:
1.769 U
508.750 n~,'
105.000 mU
1.275 U
H0 CHRIIttEL SRUED
DIFF
EXP
27.200 ms ~ I
1.260 U ~ 1
- I . 170 U ~ 1
DISP--RRHGE
0 . 0 0 0 . . . 800.000 w,l:
-1.2.50...
3.868 U
0.000...
20.475 U
CUR JP
108S
Dynamic curve of the torque on the motor and pump sides at break-
43
non-stationary magnetic fields are generated by the poles present. If a

pump set is operated for any length of time under such operating
conditions, damage due to vibrations is to be expected, e.g. to the sleeve
bearings or the shaft/hub connections.
From Fig. 2.16 it can also be derived that by enlarging the rotating mass
of the motor, e.g. the magnet holder on the motor side, the size of the
magnetic drive can be kept smaller and, therefore, more economic. For
actual pump and motor data the design range may also start in the crossshaded area just below the limiting curve. Figure 2.20 shows an example
of a graded drive system for a two-stage side channel pump. It is recommended that pumps with a very high moment of inertia are equipped with
a soft starter in order to reduce the starting and breakdown torques so
that in the end the operating point torque is the only factor left to
determine the size of the magnetic drive. The resulting torque curves
obtained with this type of starter are shown in Fig. 2.21.
By adopting this solution, the flywheel moment on the motor side and
the magnetic drive, and the vortex current losses can be reduced to a
minimum.
Figure 2.20 Two-stage side channel pump with magnetic drive (SIHI).
44
327'6E.1.1 TBI: 200us TBZz OFF
88/ 5," 6 10:24:24
(ItBR, & ElffER)
MAM
F ........................... -
..................................................
nM
~../f
. . . .
....
CURS
oR6
oIFF
X: T I H E
2 3 0 q . 0 0 0 ms
0 . 0 0 0 ms 2 3 0 q . 0 0 0 ms
- - R : CH 1,001
17,.=;00mU
27,500 mU -lO,000 mU
~B:
C H ~
10.155 U
30.000 MJ
10.125 U
SALE0 CHattELS : f~) CHRI"gtEL S ~ ' O
3276E~4 TBI: 200us TB2: OFF
ExP
~ I
~ I
* I
OXSP--RRNGE
0 . 0 0 0 , . . 6qO0.OOO ms
-2.500.,,
7,737 U
-0.6q0...
t9.815 U
CUR JP
11520
88/ 5/ 6 10:24:24 EIO OF OJTPUT (IIBR. & EI4TER)
MAp
.
--.
--_-_
---
. . . . . . . . . .
.. _. . . . . . . . .
. . . . . .
,. . . . . . . . .
'
X:
---R=
~B:
SRt~O
.
.
.
.
C u R s
....
{ RG
O
DIF'F E X P
TIME
2176.000 ms
0.000 ms 2176.000 as ~ I
CXIB~i]l
930.000 mU
-7.500 ~1
937.500 mU t I
CH 4.001
10.000 mU
5.000 mU
5.000 m U t 1
CH~r
: I~) CHRtttEL SF~ED
Figure 2.21
starter.
DISP'RRNGE
0 . 0 0 0 . . . 6q00.000 ms
-2..'500...
7.737 U
-0 640.
19..;35 P
CUR ~P
1O830
.
.=
,"<
Dynamic curve of the torque on the motor and pump using a soft
45
Safety requirements
Technical standard
For some time to come, the main fields of application for pumps with
magnetic couplings will inevitably be found in the chemical and process
industries.
This fact must not only be catered to from a design point of view, but
the stringent safety requirements applicable in these branches of industry
also have to be met.
To begin with, the demand for the smallest possible leakage rate can be
easily fulfilled by providing the motor can with a static seal. The required
reliability of operation is achieved by designing an adequately dimensioned wall thickness for the can and selecting a corrosion-resistant material
of construction, and, secondly, by providing a large enough clearance between the inner and outer permanent magnets. With these technical features alone, it is possible to offer a product with a balanced set of
characteristics which exceed by far the reliability of operation and the
availability of pumps with single or even double-acting mechanical seals
including a sealing pressure system. These are facts to be remembered by
those demanding additional safety provisions for the area in contact with
the atmosphere located outside the can.
If required for a particular application, the safety of this area can be
increased by adding one or more additional safety features, such as the
monitoring of this area, which may or may not be sealed off with an additional auxiliary seal, by means of a sensor, or by fitting the motor with a
double-walled can or with two independent cans.
Liquid level, temperature

Many places of installation in the chemical and process industries are
subject to explosion hazards. Although the magnetic couplings themselves
are not subject to the special regulations applied to all electrical equipment, for example canned motor pumps, every care must be taken to
ensure that they do not give rise to any hazards.
As a general rule, the maximum surface temperature of the couplings
must comply with the required temperature class of the host equipment.
Even with metal cans, only a small temperature increase is to be expected in the area of the can, assuming, of course, that the machine is run
within the specified operating limits, the maximum permissible temperature of the liquid is ultimately dictated by the temperature class. Nowadays, magnetic couplings with magnets made from SmCo can withstand a
46
flow temperature of approximately 300~ maximum. However, when

determining the size of the drive assembly, the fact that the torque to be
transferred decreases superproportionally as the temperature increases as
a result of the reversible and irreversible changes in magnetisation, has to
be taken into consideration. For example, at a flow temperature of 300~
a magnetic coupling can only transfer approximately 70% of the rated
torque at 20~ This is based on the assumption that the external temperature field is kept lower by means of appropriate changes in design.
Furthermore, the end-user is interested in knowing what temperature
to expect in the event of incorrect operation or breakdown of equipment.
Operating a pump fitted with a magnetic coupling without liquid must be
avoided. In this case, the temperature in a metal can would go up very
rapidly and steeply. In order to stop damage from occurring or temperature limits from being exceeded, a liquid level indicator should be fitted,
or other adequate precautions taken, in particular in those areas which
are subject to explosion hazards.
A further point worth noting is that an extended period of improper
operation with the valve closed, or an interruption of the flow due to a
blocked or broken-off inner magnet as a result of overloading, can cause
an increase in the outer temperature of the can as well as the temperature
of the liquid inside the pump.
In order to understand these phenomena better, the author fitted a
water pump with a magnetic drive assembly installed in a pumping plant
with thermocouples. The pump was operated starting from a stationary
operating temperature of 20~ and recorded the temperatures (Fig.
2.22). The readings obtained in situ were the temperature in the zone
of maximum temperature in the can (Os), at a sensor connection (V~r)
mounted on the flange of the can, and in the liquid inside the discharge
casing (t~F).
As Fig. 2.22 shows, the difference in temperature between the three
values is not very large. The rise in temperature continues until an equilibrium is reached between the thermal energy input and the heat dissipating
through the surface. The process comes to a definite holding point as soon
as the temperature reaches the boiling point of the medium.
A comparable, yet less steep, temperature increase is to be expected as
a result of the weaker magnetic field in the event of a break-down caused
by a blocked or broken-off inner magnet.
Generally speaking, operating limits of this kind should be recorded
and identified as incidents by means of monitoring instruments installed
in the plant.
In addition to this, a measuring connection on the can allows the enduser to monitor the temperature. As the used magnet materials are
47

200-
150-
'
Flow rate
Q=o - ' t - - - - Q : % . - 100 . . . .
50-
,1
......
rl
10
20
30
~F.
40
50
60
mh
70
Figure 2.22 Temperature pattern with an interrupted flow.
characterized by a high degree of thermal stability, it is finally up to the

user to decide which means of monitoring he deems necessary.
Future developments
With the magnet materials available today, it is possible to manufacture
reliable magnetic couplings the use of which is not solely justified with critical, expensive or environmentally hazardous liquids. As almost all standard pumps can be fitted with these couplings, other possible fields of
application should be investigated on a much broader scale than in the
past, considering the economic efficiency resulting from the reduced amount of maintenance they require. In addition to pumps, leak-proof magnetic couplings are meanwhile being offered as predictable machine
elements for application in other equipment used in the process industry,
such as agitators and valves. Whereas the eddy current losses are of minor
importance in these applications, more attention should be paid to the
non-conducting material used for the cans of pump drives.
Apart from the various synthetic materials available, industrial ceramics are gaining more and more importance, not just for the internal sleeve
bearings, but also as a can material (Fig. 2.13). In particular on pumps
used for the delivery of very hot liquids (heat transfer liquids), an uncooled magnetic coupling with a ceramic can fitted in the 'dead-end' zone
(Fig. 2.23) can be considered a practical solution. Here, the author calls
on both manufacturersand end-users to advance this technology with the
necessary expertise and the responsibility demanded by the application.
48
Figure 2.23 Volute casing pump with an uncooled magnetic coupling for hot
liquids (SIHI).
Bibliography
1. B0tel, A. and Schattka, H.-L.: Untersuchung von permanentmagnetkupplungen zum antrieb stopfbuchsloser kresielpumpen, unpublished thesis paper, Liibeck Technical College (1987).
2. Petersen, A.: Wirbelstromkupplungen und diimpfungssysteme mit permanentmagneten, I/ALI/OBerichte XV, No. 2 (1969). (Eddy current couplings and
damping systems comprising permanent magnets.)
49
3. Dittrich, O.: Wirkung der kupplung beim anfahren mit elektromotor, Antriebstechnik 12, No. 4, 85-89 (1973). (Coupling response on start-up with an
electric motor.)
4. Rix, H.-P. and Mohrdieck, E.:Untersuchungen der anlaufverh~iltnisse an einem
pumpenaggregat mit verschiedenen permanentmagnetischen kupplungen,
unpublished thesis paper, L0beck Technical College (1988).
Symbols and physical sizes

A
J
M
P
b
m2
kg/m 2
Nm
kW
m
N/mm 2
n
p
1/min
u
tp
m/s
degree
p
to
f~mm2/m
1/s
area
moment of inertia
torque
power rating
can wall thickness
constant
rel. magnetic force
length of magnet
speed
number of magnetic poles
effective radius
time
circumferential speed
phase angle
efficiency
spec. resistance
angular acceleration
Indices
A
K
M
P
W
magnetic coupling
coupling
motor
pump
eddy current
CHAPTER
THREE
Zero-leakage pumps
equipped with permanent
magnetic drive
M. Knorr
Introduction
The basic idea is to achieve slipless transmission of the torque by the use
of permanent magnets through an isolation shell that would perfectly seal
the liquid chambers. Continual improvements have been made by the
application and evaluation of new technologies.
For three decades, these systems have been employed by the chemical
industry where aggressive and/or toxic liquids are frequently pumped. The
major areas of concern were mostly plant components subjected to
extremely critical operating conditions and processes.
Today, with its sophisticated magnet system and silicon carbide journalbearing configuration, the seal-less magnetic-drive pump exemplifies
operational safety and reliability. It also incorporates the advantage of a
markedly lower sales price while meeting the demand for a readily maintained pump, as compared to that of other sealing systems.
Permanent magnetic, synchronous coupling design

and operation
The synchronous coupling utilizes attraction and repulsion between the
permanent magnets of both coupling halves to achieve non-slipping transmission of the torque, free of any mechanical contact.
Figure 3.1 is a schematic diagram of the synchronous coupling. As shown,
there are an even number of single permanent magnets of different polarity
arranged symmetrically with respect to rotation.
Because the magnets in the coupling are arranged circularly and co-axially,
there are no undesired resultant axial or radial forces. This arrangement
52
0 ):))I])
Figure 3.1 Synchronouscoupling.

allows design of the isolation shell with a minimum thickness and adaptation of the torque to its required duty by changes in the system length.
The magnets of the two coupling halves have different polarity, and in
the off-load position they attract until north pole is opposite to south pole.
In order to counter-rotate these poles, the torque must increase, reach a
maximum and fall to zero until like poles are in alignment.
When operation is started under these conditions, the coupling will
either jump forward or backward, depending on the direction of start,
until it is in a stable position.
When a torque is applied and rotation is started on the drive side of the
coupling, the driven side will stay at rest until such time as the attraction
torque between the magnets is equal to the torque required to rotate it.
Once this point is reached, the driven side will rotate in synchronization with
the magnets in a slight misalignment equal to the drive torque requirement.
If the drive torque exceeds the maximum torque the magnets can transmit, the torque will decrease and the driven side will desynchronize and
stop. This implies that synchronous rotation is required for the transmission of a torque.
It should be emphasized that, with an optimum sizing of the systems
and with appropriately selected magnet materials, coupling stalling will
not cause any magnetic field failure. However, synchronization of the
coupling halves must be restored.
Permanent magnet materials which are currently used for technical
equipment components are classified into ceramic and metallic materials.
The ceramic material has no significance with regard to the transmission of torque, but ceramic is preferable for use with those pumps where
the pumpage consists of highly concentrated nitric acid.
Zero-leakage pumps with magnetic drive
53
Metallic materials are classified into two groups. The line that follows
the demagnetizing curve and notably the ratio between remanence and
coercivity are decisive for designing the optimum shape and the utilization of the density of energy. Today, it is common practice to manufacture
the magnet system from rare earth metals, e.g. cobalt permanent magnet
materials. However, the temperature coefficient of the remanence restricts
the continuous operating temperature to 480~ If continuous operating
temperatures are between 480-840~ the magnet systems are made from
AlNiCo permanent magnets. This material allows use of couplings up to
840~ operating temperatures without any excessive detailing work and
safely precludes any essential torque losses caused by the temperature
level.
According to the specification, and consistent with the operating temperature and the pressure level in the system, the system length could be
changed to allow the use of synchronous couplings made from AlNiCo
(up to 3200 ft/lb), or from samarium cobalt (up to 25300 ft/Ib) to transmit
the torque without any mechanical contact.
Seal-less magnetic-drive pump design and

operation
If a generally accepted magnet coupling is to be used in a pump, the
hydraulics and design aspects must be known. All data of the interior and
exterior acceleration and shock moments of the entire unit (with the
inclusion of the electric motor and the flexible coupling) have to be
obtained prior to use.
For direct start-up of the motor, i.e. not requiring any start-up aid, the
maximum moment of the magnet coupling must exceed that of the maximum moment (start-up and/or tilting and/or rate moment) that could
occur between the two rotating magnet systems of the magnet coupling.
Figure 3.2 illustrates a seal-less magnetic-drive pump complete with a
rare-earth-cobalt permanent magnetic synchronous coupling which is
capable of operating continuously at temperatures ranging from -150~
to 480~
Mode of operation
A conventional standard motor is normally used to drive this pump. The
required torque is transmitted to an exterior magnet carrier equipped with
permanent magnets through a shaft, which is supported by permanently
greased or oil-lubricated grooved ball bearings and a flexible coupling.
54
Figure 3.2 Seal-lessmagnetic-drive pump SLM-N.

The most essential component required by a magnetic-drive pump is a
partition wall, the so-called isolation shell. This isolation shell must separate the liquid chamber from the atmosphere. The interior magnet carrier, which is also equipped with permanent magnets, is arranged inside this
isolation shell. The interior magnet carrier is supported by a bearing
flange and journal bearings lubricated by the liquid pumped, allowing
non-slip torque transmission from the motor to the impeller. A casing
encloses this perfectly sealed unit.
Hastelloy is preferable for the fabrication of the isolation shell. By
highly advanced fabrication processes, the cylindrical portion and the bottom of the isolation shell are one seam-less piece. This procedure provides the isolation shell with an extraordinary stiffness and strength.
On-site system pressure determines the wall thickness of the isolation
shell. Due to their physical properties, metallic, non-magnetizable materials are electrically conductive. Since magnetic field lines are constantly
interfacing the isolation shell, eddy currents will occur, mostly at high
speed levels, which - compared to conventional pumps - would lead to
energy losses. This loss is converted into heat which is introduced to the
liquid pumped via a flush flow. For this reason, material of high strength
and high electrical resistance is selected. If the above measures are followed, an efficiency of 95% can be obtained in the region of the maximum
torque--even at high speed.
55
Because the pump construction is simple with few component parts,

disassembly and assembly is easy, meeting the demand for a readily maintained pump assembly with maximum interchangeability.
Journal-bearing design, life and safety

The axial thrust inherent in every pump must be eliminated, particularly,
if magnetic-drive pumps are involved. This is achieved by applying special
means or by installation of a journal bearing, depending on their direction
and forces.
Safe operation of these pumps was frequently questionable due to the
use of journal bearings made from epoxy-resin impregnated or metalimpregnated, hard-burned carbon or polytetrafluorethylene. This is due
to the different types of liquid pumped, often with very low viscosity and
hence low lubricating properties and/or with solid constituents.
Silicon carbide journal bearings have excellent physical properties
because they offer excellent resistance to wear, are extremely hard, thermally stable and chemically inert. Therefore, they are particularly well
suited to increase the service life and the operational safety of magneticdrive pumps.
The pump design described in this chapter requires the installation of
a hydrodynamic journal bearing in the liquid chamber, and normally the
pumped liquid is used to lubricate it. The journal bearing is designed to
withstand any combination of axial and radial forces enabling the pump to
operate over its entire performance range without limitation.
The design of the journal bearings complying with these requirements
incorporates an extensive proprietary knowledge, and therefore, no design details can be disclosed here. However, the flushing system is of paramount importance (Fig. 3.3). A small flow of pumpage is taken from the
pump at a high pressure area. The flush with its pressure maintained, is
fed internally into the journal bearings where the flow is separated to
lubricate the thrust bearings and to cool the area between the inner
magnet carrier and the isolation shell.
Using water/water-based liquids during normal pump operation, the flush flow temperature increase ranges from approximately 1.8~ to 14.4~
The pressure curve is shown in Fig. 3.4.
Magnetic-drive pumps, installed in process plants where the NPSHR
and the NPSHA are very narrow and where the temperature of the pumped liquid is very close to the boiling point, require that the flush system be
changed (Fig. 3.5).
In order to prevent boiling liquids from evaporating, the flush is led via
an internal system to the rear of the impeller (pressure end). With this
56
s 3:
O ~
II
.4
Figure 3.3 Siliconcarbide journal bearing design and standard flush system.
0 i scharae
Pressuf'e
5uc t i on
Pressure
Figure 3.4 Flush system pressure curve (standard design).

s.
~"
II
57
.4
Figure 3.5 Flush system (liquid gas pumpage).

flush-system configuration, the required superimposed pressure in the
isolation shell is achieved (Fig. 3.6). This implies that these pumps can
also be used if the pumpage is a boiling liquid in spite of the heat that will
be produced in the isolation shell area.
No maintenance and safe operation

Because construction is simple, magnetic-drive pumps designed in accordance with the operating data specified on the user's data sheets and
operated in strict compliance with the operation manual, are operationally safe and readily maintained units.
Operational monitoring
Prior to defining the required protective measures the following should be
noted.
Pumps must be filled with liquid before they are started in order to
ensure lubrication of the journal bearings by the internal and external
58
Discharqe _
Pressu~'e
I
o Return
p3 Pressure
Suet
1
3'
i on
Pressure
=_
Figure 3.6 Flush system pressure curve (liquid gas pumpage).

flush and to dissipate the heat from energy losses in the interior areas of
the isolation shell.
For monitoring the inlet conditions at the suction end, the following
control means have been used satisfactorily:
(a) float switch liquid level monitoring,
(b) capacitive liquid level monitoring for non-conductive liquids,
(c) conductive liquid level monitoring for conductive liquids,
(d) light beam limit value transmitter,
(e) resonance probe liquid level monitoring.
The conditions at the delivery end, particularly of the flush system, are
monitored by the following:
(a) magnetic flush rate monitoring transmitter,
(b) resistance thermometer (PT 100) on the isolation shell.
The temperature increase on the isolation shell is monitored via a resistance thermometer (PT 100).
As long as the pump circulates a liquid flow (water-based liquid) in the
range of Q rain. and Q max., the temperature increase of the isolation shell
will be in accordance with Fig. 3.7. The pump concerned is of type 40-250
(3" x 2" x 10") with 22 hp magnetic drive and 5 hp loss in performance.
59
AT
/ 2 '
-
J /
11
0
....
",,--
10 20 30 40 50
t [minJ
Figure 3.7 Temperature increase on the isolation shell.

The location of the temperature sensitive element determines precisely
any temperature variations. Therefore, the switch-over point will be set as
a function of the maximum operating temperature and the specific data of
the pumped liquid.
Vibration velocity measurements have become a standard procedure
with magnetic-drive pumps and are recorded during each trial run by a
vibration and spike energy analyzer.
The recordings allow immediate maintenance actions to be taken
depending on the actual condition of the unit concerned.
All information collected during normal operations provides a clear
picture of the mechanical condition of the pump. Mechanical vibration
measurements have proven to be a particularly sensitive and accurate
indicator of the mechanical condition of the pump. If the readings are
followed in terms of time, changes of the conditions provide early information of an anticipated pump failure before any such failure occurs.
Shock impulse measurement (SPM) is a reliable means to detect commencing anti-friction bearing failures at a very early stage. If desired, this
can be done by the installation of test nipples in the area of the antifriction bearings. The test readings- via a test instrument- will be evaluated immediately according to set evaluation rules.
The SPM method is based on the fact that anti-friction bearing failures
cause mechanical shocks and that at the point of the shock a material
acceleration occurs, the extent of which depends on the kind of damage,
bearing size and speed.
60
The front of the pressure wave will hit the transmitter, and an undamped resonance vibration is produced. The increase in amplitude is converted into analogue electrical impulses and the readings of the test results
can be taken. Nomograms, allowing instant assessment of the bearing
conditions are made available to the user.
Safety systems
If failures of anti-friction bearings or journal bearings have occurred or if
solid particles are unexpectedly included in the liquid pumped, the isolation shell could be damaged, allowing the liquid that was isolated before
to escape to the atmosphere. This risk had previously been precluded by
providing the enclosed driver casing with radial lip seals (Fig. 3.8) and
appropriate alarm signals. Even after an extended low level operation,
this equipment was able to hold back the pumped liquid that would have
otherwise leaked from the damaged isolation shell.
Nowadays, it appears that the efficiencies during motor power transmission to the pump impeller with a permanent magnetic coupling installed play a minor role.
m
V
Figure 3.8 Magneticpump SLM-N with radial lip seals.
Zero-leakage pumps with magnetic dl~ve
61
Instead, safety has become the number one priority and was originally
ensured by increasing the wall thickness of the isolation shell. In this construction, however, the air gap was increased and eddy currents inside the
isolation shell caused a performance loss. Also, any imminent leakage of
the thicker isolation shell would not be noticed.
An improvement in pump drives of this type resulted in the Sealex9
Control Double Skin isolation shell (Fig. 3.9). The 'Sealex-CDS'- system
incorporates an interior and exterior shell which are maintained at a distance allowing pressure readings to ensure safe operation. The cylindrical
shell area of the double skin isolation shell also allows the narrow distance
required between the permanent magnets to maintain high efficiency.
Since either the exterior shell or the interior shell could be damaged, each
individual shell has been designed to withstand the design pressure.
The areas of contact, at least in the cylindrical portion of the double
shell, provide mutual support, implying that the increased safety properties
of the single walled isolation shell are obtained. In addition, this arrangement features the advantage that the volume in the network of paths between the two shells would not change if the pumped liquid caused pressure
shocks and/or pulsating superimposed pressure. Both shells would expand
if they were subjected to any pressure variation, but their contact would be
maintained. This arrangement precludes any false alarm which would
Figure 3.9 Magneticpump SLM-N with 'Sealex-CDS' system.
62
otherwise be given if the volume in the gap between the two shells varied
because of pressure variation of a single or pulsating nature and if the
pump were controlled via pressure readings.
Pump and agitator design examples

The pump that is illustrated in Fig. 3.10 is a fine example of the creative
and innovative work of a magnetic pump design along with a flexible
approach to tackle any task.
Operating data:
9
9
9
9
9
liquid pumped: highly subliming smelt, 155~ setting point

operating over pressure (PN 63)
operating temperature (250~
viscosity: 31.25 mE/s
density: 1.6 kg/dm 3.
The pump design and the user's operating procedures require

compliance with the following:
9 start-up procedures
9 isolation shell temperature control
Figure 3.10 Magnetic pump SLM-N with casing and bearing bracket, heating
coil and double-skin isolation shell.
63
9 double skin isolation shell

9 complete drain prior to each new heat.
For agitator drives with permanent magnetic synchronous coupling
(Fig. 3.11), the magnet system and the design of the isolation shell are to
the generally adopted rules.
The user would neither accept any lubrication of the bearings fed from
the exterior nor the use of the liquid produced in the process. This implies
that any additional opening of the system to the exterior is inadmissible.
Figure 3.11 Agitatorwith permanent magnetic synchronous coupling.
64
The requirements imposed on the design were met, first, by the use of
inclined ball bearings made of silicon carbide with cages made of Hastelloy, and further, by the allowance for rotation of the agitator shaft below
any critical point and the bearings having no inherently stored lubricants.
This highly sophisticated solution is another example for the application
of synchronous coupling in the process industry. In order to keep the cost
price of agitator drives low, it is strongly recommended to carefully clarify
and define the maximum speed of the agitators in the prebid stage.
Figure 3.12 shows design details of a magnetic-drive pump with a
separate flush system.
For liquids containing a high quantity of solid particles, it becomes
necessary to use separate flushing systems. The separate circuit ensures
lubrication of the journal bearing and dissipates the heat generated by the
losses in the magnet system. The separate flushing system can use a
neutral, filtered liquid or a process-compatible liquid. The flush rate of
approximately 0.5-2.5 gal/h which is lost into the pumpage is continually
made up by an auxiliary feed unit. A heat exchanger is installed to dissipate any excess heat.
lllln
Heat exchanser
0 T;outiet
0 T;inlet
Figure 3.12 Magnetic pump SLM-N with separate flush system.
65
The selected inlet pressure must be higher than the pressure at the suction end and should prevent the liquid pumped from flowing backward
into the separate flushing circuit.
The seal-less self-priming side channel pump SLM-STR in multistage
construction with suction impeller (Fig. 3.13) is designed for I ma/h up to
30 ma/h and delivery heads of 100 m. The design of the suction impeller
permits cavitation-free pumping at NPSH values around I m.
Figure 3.13 Seal-less self-priming side channel pump SLM-STR with magnetic
drive.
66
Another fine example is the seal-less screw pump with magnetic-drive

SLM-LN (Fig. 3.14) demonstrating the variety of possible uses for magnetic couplings in pumps.
Figure 3.14 Seal-lessscrew pump SLM-LN with magnetic drive.
67
Conclusions
Steadily growing environmental awareness, increased demand for operational safety, availability of installation, extended maintenance intervals,
and very low service cost are some of the reasons why a great number of
zero-leakage pumps with magnetic drives have been installed.
Their advantages, compared to canned pumps, have been highlighted
here. The seal-less magnetic-drive pump is considerably less complicated
and notably lower in price than the canned pump.
Canned pumps may need an off-standard motor that can lead to problems with respect to readily available spare parts and also maintenance
and repair work. The repair work is penalized by high cost and labour.
Further, skilled maintenance personnel is required for the mechanical
and electrical components.
The objective of this chapter is to provide detailed information on the
present state-of-the-art of the magnetic-drive pump. The aspects of possible combinations of the magnetic coupling systems with positive displacement pumps, multi-stage centrifugal pumps, vertical pumps and
side-channel pumps have not been discussed herein. Slipless torque transmission by means of permanent magnets is a widely accepted concept that
should continue to develop and grow in the future.
CHAPTER
FOUR
Leak-free centrifugal
pumps in plastic
G. Naasner
Introduction
For the delivery of corrosive, toxic and abrasive media in the chemical and
pharmaceutical industries increasing use is being made of leak-free centrifugal pumps in plastic. These pumps are suitable for the delivery of such
media due to their high chemical and thermal resistance and, because
they are leak-free, they also meet the increasingly stringent environmental
requirements.
Leak-free centrifugal pumps in plastic are the exclusive domain of magnetic drive pumps. This type of pump is examined in more detail below.
Materials
Hermetically sealed plastic centrifugal pumps are either made of solid
plastic components or consist of thick-walled plastic linings armoured by
an outer metallic shell. As these plastics are exposed to the attack of the
process medium, particular requirements must be met, such as:
9 chemical resistance
9 temperature resistance
9 abrasion resistance.
In cases of higher chemical and thermal stress, fluorinated plastics such
as PTFE and PFA are employed. They provide a universal corrosion resistance and a temperature resistance for the components of up to 180~
(see Table 4.1). For lower temperature requirements FEP may alternatively be employed (but is seldom used in pump manufacturing) and in
cases of lower resistance requirements the partially fluorinated PVDE If
chemical resistance and temperature resistance allow it, UHMW-PE is also
a material well-suited for pumps; particularly on account of its remarkable
abrasion resistance.
70
~0
~)
c~
~)
I
~.)
c~
c~
. ,..~
.,..~
0
b~
.,..~
0
0
0
bX)
.,..~
0
0
0
b~
0
0
0
b(}
0
bI}
"0
0
0
bO
e~O
0
0
0
0
0
0
4--
~ p,,q
Q
~
4--.t+
.=. v, =o
Leak-free centrifugal pumps in plastic
71
When requirements exceed the mechanical limits of the selected plastic,

ceramic materials are employed (see Table 4.2). Of particular significance
here, besides aluminium oxide and carbon, is silicon carbide, which has
become the standard material for plain bearings and mechanical shaft
seals. The exceptional qualities of silicon carbide equipped with Safeglide
are more thoroughly examined below.
The conveyance of combustible or explosive media possibly requires
the employment of electrostatically conductive materials. Only the silicon
carbide materials and carbon naturally provide these properties. Fluoroplastics are, as a rule, available as a special type in electrostatically conductive versions. FEP and A1203 being the only cases where this is not
possible (see Table 4.3).
Plastic-compatible construction
Plastics have, to a certain degree, quite different properties to metals;
such as
9
9
9
9
high thermal expansion

high level of shrinkage during processing
'cold flowing' of the plastic when under stress
reduction in stability of shape with increasing temperature.
These characteristics must be taken into account for plastic components

as a decisive factor in the reliable operation of equipment.
The following construction principles are valid for high-performance
centrifugal pumps made of plastic:
9 Inside:
9 Outside"
Thick-walled plastic lining or coating of all wetted parts

with chemically and thermally resistant plastics.
Metallic armouring to support the plastic lining, to absorb
the pressure loads, to permit simple pipeline connection
and to absorb the pipeline forces.
By means of this clamshell design, the specific material properties of plastic

and metal are favourably combined (Fig. 4.1).
Following the principle of avoiding unnecessary sealing joints, the manufacturers endeavour to produce a one-piece plastic pump body wherever
possible. Such a design offers the greatest possible tightness. Static seals are
merely necessary for pipeline connections and back plates (Fig. 4.1).
The fluoroplastics employed for high-performance plastic pumps are
subject to considerable shrinkage during fabrication. This can mean the
plastic separating from the metallic wall, creating a cavity between lining
72
r~
ur
~ ~..~
N
o ,.~
t~
~ ,,~
I ++
+.
+ + + +
+ + + +
+.
++'I"
+ + + +
-I-++
++++
+ . +
+++I
++++
++
++
I ++
-,-- "I'+
I +
4-- "I--+-I"
I ++
+,+I
4" I~+I
~.~ @~-~ = ~
4.1
0
~j
"0
I-i
~b
4-i
0
0
-I--..14"
~8

0
.I
. .-~
.,-i
r.~
p~
I-,

Table 4.3
Electrostatically conductive materials for non-metallic pumps

....
Ele'ctrostatically conductive . . . . .
(conductive resistance < 108 )
standard material
electrostatically
conductive special
material
PTFE
PFA
FEP
PVDF
PE
A1203
Pure SiC
SiSiC
Carbon
i
Figure 4.1
73
One-piece pump body in metallic shell.
N0t electrosta'iically
conductive
74
and armouring which results in undefined geometries. This alteration in

shape must be prevented by dovetail grooves, boreholes or recesses in the
metallic part which provide a positive locking of the plastic lining (Fig. 4.2).
Over and above this, pump impellers have to guarantee safe transmission of the torque, even at higher temperatures. For this purpose metal
cores of as large a diameter as possible and with holes or cogs are provided as the hub of the impeller (Fig. 4.3).
Due to the further development of production techniques, it is
today also possible to manufacture impellers of fluoroplastic with threedimensional curved blades. Considerable improvements in hydraulic
geometries, with notable increases in efficiency and NPSH reaction are
thereby achieved (Fig. 4.4).
Can units
Cans are an essential element of magnetic drive pumps for the functioning and operational safety of this equipment. The cylindrical part of the
can lies, in the case of central magnet couplings, between the inner and
outer rows of magnets. In this area, metallic cans are made very thin in
order to allow as little distance as possible between the inner and outer
magnets. This distance between the magnets has a considerable influence
on the volume of magnets employed. Over and above that, the thickness
of metallic cans is a direct criterion for the induced eddy current losses
and, thereby, for the generation of heat in the can.
Such a narrow space between the magnets cannot be achieved with a magnetic drive pump made of plastic. This is because a thick-walled plastic lining is required for the inner magnet assembly and, furthermore, because of
the temperature-related expansion of the plastics, which therefore require
larger clearances for safe operation. As a side effect, these larger clearances
make the pump more suitable for handling solids! Additionally, the heat generation due to the eddy current losses of a metallic can would, under certain
circumstances, exceed the temperature resistance of the plastics employed.
For these reasons plastic magnetic drive pumps are, as a rule, equipped
with non-metallic cans. With such cans, eddy current losses and the associated problems do not occur.
Cans are of single or double design and manufactured in various material
combinations (Table 4.4). For high-performance plastic magnetic drive
pumps, the double can system should usually be employed. The inner can is
made of, for example, PTFE and guarantees the chemical resistance (also at
higher temperatures up to 180~ The outer can is made of carbon fibre
composite plastic (CFRP) and absorbs the prevailing pressures and pressure
peaks in the plant (Fig. 4.5).
75
Figure 4.2 Plastic-lined plain bearing pedestal of a magnetic drive pump with
anchorage grooves.
76
Figure 4.3 Plastic impeller with inserted metal hub with anchorage boreholes.
Leak-free centrifitgal pumps in plastic

Table 4.4
.
77
Can materials for magnetic coupling pumps

.
i,
ii,
Materials
metallic
2.4610
Hastelloy
Property
.
Corrosion resistance
Internal pressure resistance
Temperature resistance
(180~
Resistance to fracture
Eddy current losses
(heating the medium)
non-metallic
ZrO2
Zircon
Oxide
1.4571
(V4A)
.
CFRP/
PTFE
GFRP/
PVDF
+
++
++
+*
-
++
+
+
-
++
++
++
++
++
- -
++
++
++
- -
++
++
++
" *Caution with HCI, HF, HNO3, H2SO4, N a O H etc.
Figure 4.4 One-piece fluoroplastic impeller with three-dimensional curved

blades and conical cover disk.
78
Figure 4.5 Double can system with monitoring of can. (Works photo Richter
ChemieoTechnik.)
The fact that eddy current losses do not apply has, apart from the
advantages already mentioned, a positive effect on the efficiency of the
pump (Figs 4.5 and 4.6). These advantages become particu'l"ffrly obvious
when operating the pump at partial load, since the eddy current losses are
not dependent on the transmission efficiency but only on the magnet
volume employed and the type of can. In partial- load operation, the full
extent of the losses remain.
Furthermore, the double design of can systems provides the opportunity
to monitor the integrity of the gap between these two cans. For this
purpose, pressure or moisture monitors can be used. For magnetic drive
pumps made of plastics, conductive circuits are also placed in this gap (Fig.
4.5). If, in case of damage, process liquid reaches the gap, either the conductive strip is broken by corrosive attack or the conductivity of the medium causes a short-circuit. The change in the resistance of the conductive
strip can be taken as a signal for protective measures. The CFRP-can and
the external bearing pedestal housing also serve as further protective
covering against a spontaneous escape of the medium in case of damage.
Dry-running
Plain bearings in magnetic drive pumps run in the process liquid and are
cooled and lubricated by the same. The last few years have seen silicon
79
Can Comparison
- full
load
58 % Pump Power Output
Power Losses
25 % Hydraulic
RICHTERCan
15 % Plain Bearing
and Impeller Friction
PTFE/CFK
2 % Rolling Bearing
100 % Pump Power Inpu
2 % Rotting Bearing
-~
HC4-Can
~~_
15 % Plain Bearing
25 % Hydraulic
10 % Eddy Currents
Figure 4.6 Can comparison in full-load.
carbide establish itself as the material for this plain bearing due to its
universal corrosion resistance, its high resistance to wear and, not least,
thanks to the reduced price level. Besides these advantages silicon carbide
is, however, as other ceramic bearing materials, highly sensitive to dryrunning. Dry-running, one of the most common causes of damage to
80
Can Comparison
- partial
load

Power Losses
30 % Hydraulic
RICHTERCan
~/.
30 % Plain Bearing
PTFE/CFK
_
- )
4 % Rolling Bearing
4 % Rolling Bearing
1O0 % Pump Power Input
30 % Plain Bearing
HC4-Can
30 % Hydraulic
9
20 % Eddy Currents
Figure 4.7 Can comparison in partial-load operation.

magnetic drive pumps (as well as to all other pumps with ceramic plain
bearings or mechanical shaft seals), does not normally occur during regular
operation of the pump. This damage occurs during unforeseen operating
conditions or due to misoperation of the pump. According to ceramics
manufacturers, temperatures of several hundred degrees Centigrade are
generated at the running surfaces within a few seconds when dry-running
81
at 2900 rev/min. This increase in temperature leads to destruction of the

plain bearing through thermal stress, causing further damage to the pump.
For magnetic drive pumps made of plastic this point is of even greater
consequence because the outer as well as the inner bearing sleeves are
embedded in plastic, which does not allow any heat discharge through
the plain bearing pedestal or the shaft due to its insulating properties
(Fig. 4.8). Developments were required which, instead of aiming at
improving the heat dissipation circuit, aimed at a fundamental reduction
of the frictional heat resulting from dry-running. The Richter Safeglide
bearings are the result of these developments. The bearings in question are
plain bearings made up of the base material SiC or SiSiC and coated at the
running surfaces with a corrosion-resistant and friction-reducing layer.
Figure 4.9 shows the coefficient of friction measured at the conventional
SiC plain bearing. In comparison, the same silicon carbide bearing with
the modified Safeglide surface achieves the coefficient of friction shown
in Fig. 4.10. The effect becomes even clearer in a direct comparison (Fig.
4.11). This shows that with a running time of 60 s for example, the coefficient of friction of Richter Safeglide bearings is less than that of conventional SiC plain bearings by a factor of almost eight.
In order to investigate the effects of this reduction in the coefficient of
friction with regard to the increase in temperature during dry-running, the
Figure 4.8 Magneticdrive pump in plastic (close-coupled).
82
Figure 4.10 Coefficient of friction of Richter SiC-Safeglide bearings.
83
0.5
0.4
0
,N
r
0.3
0
4,a
Q
"6 0.2
Ssfeslide
o.I
I
10
I
20
i
30
I
4O
J
50
1
60
t (s)
Figure 4.11 Comparison of the coefficient of friction for conventional and

Safeglide plain bearings; plain bearing material SiC, speed 45 min-1.
test set-up shown in Fig. 4.12 was chosen. This involved measuring the
temperature of the rotor-sided plain bearing, which absorbs the axial
thrust of the pump. The measurement results (Fig. 4.13) show, when dryrunning, a dramatic increase in temperature of conventional SiC plain
bearings compared to a very moderate increase in temperature of Richter
Safeglide bearings, reaching a maximum T = 15~ after 10 min.
As a result of this development work, magnetic drive pumps made of
plastics have now become available which are indifferent to dry-running
over a limited period of time. This operating method is certainly not
recommended but, in case of unintentional dry-running, additional
protection is provided. In the case of wetted plain bearings (e.g. following
the draining of a vessel), the pump can be operated for an even longer
period of time without additional cooling and lubrication. As a result, a
large proportion of bearing and pump damage, which otherwise arises as
a consequence of unforeseen operating conditions or misoperation, is
eliminated.
Solids in the medium

Process media, for which magnetic coupling pumps made of plastic are
employed, are seldom pure. They are mostly mixtures, containing solids,
crystals, particles or other such material.
Due to their design, plastic pumps possess large clearances (e.g. between
impeller and pump housing or between magnet assembly and can), which
-1onr beQrng/
/Instrument
.'/"
~I~
'
.-;';
'
,.'-.'~-.'
--d,. ~ ~ B : .
.III~~
B_-2arlqg sleeve
./.-..'.~~ ~ . , ~
.
:%
.... . . . . . . ~ . . ~
Bec-~na5ss~
~ . ~ i X
.-j,
;K;-:..-'.,
~ - -
,/
Figure 4.12
t",
Zo..-L
~4-1i-i~ I
I
I
[eQds
/ " iMeOStjFiRQ FsIR.
~ - - " .
" ~ fPer-........
OO
oedesto[
I1
Test set-up to determine the increase in temperature of plain bearings when dry-running.
r~
Leak-free centrititgal pumps in plastic

~, 30
o,
85
I Conventional
2 Safeglide
E
0
5
t (rain)
10
Figure 4.13 Comparison of increase in temperature when dry-running of conventional and Safeglide plain bearings; bearing material SiSiC, speed 1450 min-1.
means a certain degree of solid particle compatibility. However, the admissible particle size of 0.3 mm should not be exceeded. Test bench results and
field experience is available to confirm the operating safety (Table 4.5).
Table 4.5 Permissible solids for the operation of magnetic drive pumps in plastic
- example
.......
Permissible share of Permissible General conditi0ns

solids in the medium particle size
Without flushing
3 ... 5%
0.3 mm
Flushing with filtered ~<10%

process medium
External flushing
> 10%
i
i,.
solids must not stick

together or encrust
solids must not stick
together or encrust
2 mm
2
ii
. / .
m m
i
ii ii
If the medium contains a higher solids content or particle sizes above

0.3 mm, the cooling and lubrication of the plain bearing with the filtered
process medium is possible (Fig. 4.14). In this case the process fluid is
extracted at the discharge nozzle of the pump and cleaned through a
ceramic filter. The filter inlet is at right angles to the main flow, so that any
possible contamination can be swept away by the process medium. The
cleaned medium is led via a duct to the plain bearing pedestal and from
there fed into the can chamber. The plain bearing pedestal is totally
closed, so the filtered medium is returned through the bearing clearances,
the only connection to the pump chamber (Fig. 4.15).
Neither the standard version of magnetic drive pump nor the design
with plain bearing flushing using the process medium can be employed
when the medium tends to encrust or stick. For such applications, and also
for solids content of over 10%, an external flushing of the can chamber is
86
Figure 4.15
drawing.
Flushing of plain bearing with filtered process medium - sectional
87
Figure 4.16 External flushing of the can chamber.

used Fig. 4.16). Pure, process-compatible liquid is passed from the outside
into the can chamber. By means of an appropriate difference in pressure,
it must be ensured that the flow of liquid is only directed towards the pump
chamber, in order to prevent any solids-loaded process medium from
penetrating into the magnet assembly chamber and the bearing clearances.
Summary and outlook

Leak-free centrifugal pumps made of plastic today offer the plant operator a wide range of opportunities for the safe and economical delivery of
corrosive, abrasive and toxic liquids in the chemical and pharmaceutical
industries.
By means of operation-specific designs with more advanced materials
and manufacturing processes (e.g. non-metallic can units or ceramic plain
bearings with improved dry-running properties), it is possible to
continually open up new and challenging fields of application.
Bibliography
1. Mersch, A.: Magnetic-drive plastic pump conquers new fields in the chemical
industry, WorldPumps 2, 44 (1991).
2. Naasner, G." F0rderung yon feststoffhaltigem korrosivem F6rdergut mit
Kunststoff- Magnetkupplungspumpen, Pumpentagung Karlsruhe (1992).
CHAPTER
FIVE
Canned-motor pumps:
an important contribution
to leakage-free
operation
R. Neumaier
Introduction
Centrifugal pumps are among the most frequently used machines in process plants. Their operating safety and functional reliability are highly
rated. Increasing environmental awareness, a dominating factor over the
last few years, combined with demands for increased safety and recent
incidents, have resulted in increased requests for leak-proof hermetic
pumps by the legislature as well as operators of plants dealing with hazardous and environmentally harmful substances. For manufacturers of
pumps with rotary drive systems, the only answer to these demands lay in
modified drive systems in which the shaft does not penetrate the pressurized pump case. Centrifugal and rotating positive displacement pumps
can only operate entirely leak-free in this way, as absolute sealing integrity
cannot exist between two chambers with differing pressures when a rotating shaft penetrates both chambers. Thus, the solution to this problem is
possible only if the torque required by the rotor is provided from the
outside through a rigid case by electromagnets or permanent magnets. Of
these two hermetic drive systems, the electromagnetic drive has, in the
form of the canned-motor pump, experienced widespread success and
gained increasing sales over the last three decades. This is based on the
advantages offered by this drive system, particularly regarding safety,
monitoring, pressure, temperature and output ranges.
90
Hermetic centrifugal pumps with canned-motor

drive [I]
Multistage design
The pump and drive motor are directly connected to form an integral unit
(Fig. 5.1). The magnet gap of the three-phase asynchronous motor is
equipped with a fixed and a rotating motor can. The fixed motor can functions as the hermetically sealed element of the drive. It divides the motor
into a rotor chamber which is constantly filled with liquid, and stator
chamber which remains dry. This construction is designated as a 'half-wet'
motor. Its hermetic design requires that the rotor bearings be located in
the rotor chamber; corrosion-resistant slide bearings are almost always
used for this purpose. The can enclosing the rotor provides further protection against corrosion. Heat resulting from eddy current, magnetization, copper and frictional losses is dissipated in the pump flow. In addition,
a partial flow QT is branched off from the discharge side of the pump via
a self-cleaning ring-type or flow-through filter and returned to the suction
side chamber via the annular gap between the rotor and the fixed motor
can, through the rear bearing, and a longitudinal bore through the shaft.
The front slide bearing is supplied with a separate lubricant flow from the
rotor chamber to the impeller inlet via holes in the impeller hub. The pump
motor shaft has a running clearance of 0.5-3 mm, depending on motor
size. The difference in pressure between the front and rear rotor chambers,
Figure 5.1a Hermetic multistage canned-motor pump.
Canned-motor pumps
91
Figure 5.1b Sectional view of a multistage canned-motor pump, as shown in

Fig. 5.1a.
and the balance mechanism of the wear rings on the slide bearings gives
good compensation of axial thrust, rendering axial bearings unnecessary
and eliminating increased frictional loss. Thus the axial thrust forces are
neutralized over the entire characteristic curve.
Pumps of this construction are currently used for:
9
~
9
9
pump flows ranging from 1 m3/h to 100 m3/h

delivery heads ranging from 10 m to 500 m
(with the use of frequency converters) up to 800 m
with drive outputs ranging from 1 kW to 260 kW.
Single-stage design with standard hydraulic system

according to DIN 24256/IS0 2858
With the high number of standard centrifugal pumps used in chemical
plants, the installation of canned motors (Fig. 5.2) can attain substantial
improvements in safety standards while eliminating emissions. The structural design and specific operational characteristics of these pump sets are
as described above. A positive feature of these pumps is the interchangeability of the hydraulic components (casing, impeller) with components of
conventional pumps; this enables spare parts stocks to be reduced, which
92
Figure 5.2a Canned-motor pump with 'wet end' hydraulics according to DIN
24256/ISO 2858
in turn cuts costs. Pumps with this design are currently manufactured for:
9 pump flows ranging from I ma/h to 600 ma/h
9 delivery heads ranging from 10 m to 230 m
9 drive outputs ranging from I kW to 260 kW.
Figure 5.2b Sectional view of a single-stage canned-motor pump, as shown in

Fig. 5.2a.
Canned-motor pumps
93
Canned-motor pumps for specialized applications

Fluids near boiling point- liquefied gases
A primary field of application for canned-motor pumps is the pumping of
liquefied gases in its broadest sense: ranging from hazardous to highly
toxic substances such as phosgene, vinyl chloride, ethylene, ethylene oxide,
hydrocyanic acid, chlorine, etc. to the widely varied field of refrigerants,
these including ammonia, freon, etc. These substances determine the
special emphasis placed on absolute tightness of the pumping systems.
Multistage liquefied gas pumps. The construction of these pumps is

basically identical to those described in the Multistage design section.
However, it differs in that the partial flow warmed by heat dissipated from
the motor does not re-enter the suction of the pump, but is directed to a
position of at least one stage of pressure, for example between the first
and second stages (Fig. 5.3). If the partial flow were returned to the suction, gas would form, as liquid gases are normally pumped at temperatures near the boiling point. Proper function of canned-motor pumps
however, requires that gas does not form in any part of the pump. The
effect of introducing the partial flow to a position with a lower pressure
than that in the rear rotor chamber can be expressed in a thermal balance
calculation and depends on the operating parameters.
Single-stage liquefied gas pumps. This pump construction involves the
reversal of the direction of partial flow, as the pumped liquid cannot be
returned to the suction because of heat absorbed from energy losses
Figure 5.3 Canned-motor pump for liquid gas.
94
Figure 5.4 Canned-motor pump with 'wet end' hydraulics according to DIN
24256/ISO 2858 for pumping fluids close to boiling point.
(Fig. 5.4). A partial flow QT is taken via a ring-type filter in the pump
pressure chamber, directed to a pressure-increasing impeller through an
intermediate chamber and the bore of the motor pump shaft. The impeller
pumps QT through the stator can gap into the volute of the pump; a partition is provided between the QT inlet and QT outlet to prevent short circuiting of the partial flow. In this manner, a flow of cooling lubricant
independent of the operating condition of the pump is guaranteed and
consistent cooling conditions are provided over the entire range of the
characteristic curve.
Heat transport
When thermal oils are pumped by conventional centrifugal pumps, liquid
escaping at the shaft seals results not only in a bad smell, but also in the
danger of spontaneous ignition. Fire hazards and odour problems can be
eliminated through the use of a canned-motor drive. Pumps suitable for
operation with a canned-motor drive are designed for either external or
self cooling systems.
Externally cooled canned-motor centrifugalpumps. Pump and motor are

physically separated from each other by a spacer in order to prevent heat
from being transferred from the pump to the motor. Pressure between
pump and rotor chamber is compensated through a relatively narrow,
long annular gap. There is an impeller in the motor itself, and the same
liquid in the rotor chamber is circulated through an external cooler which
Canned-motor pumps
95
Figure 5.5 Centrifugal pump for thermal oil with externally cooled canned
motor.
is either arranged around the motor or mounted separately. Heat due to

energy losses in the motor is absorbed by a cooling liquid, in most cases
water. In this manner, two pumping circuits are formed with differing
temperature levels. The operating circuit can withstand temperatures of
up to 500~ while the secondary cooling circuit remains at substantially
lower temperatures ranging from 60 and 80~ The motor windings are
designed for insulation class H. Due to the pressure compensation in the
annular gap, any exchange of liquids between the two temperature zones
is practically eliminated. The externally cooled circuit has the advantage
that standard motors of insulation class H can be used regardless of
pumping medium temperature.
Uncooled canned-motorpumps. If operating conditions do not allow the

use of cooling liquids, self-cooled motors as shown in Fig. 5.6 may be
employed by using silicon-ceramic insulated windings with high temperature stability. This system allows liquid temperatures of up to 350~ in
continuous operation. A special feature of these motors is the physical
separation of the terminal box from the stator case in order to keep the
96
Figure 5.6b Canned-motor pump with self-cooled motor, as shown in Fig. 5.6a.
temperature of the terminal box low for the incoming mains cable. Such
motors are suitable for installation in pumps which must be heated for
start-up and during operation, and for operating conditions under which
a certain temperature range must be maintained for technical reasons
(see lower half of Fig. 5.6).
Canned-motor pumps
97
Pumping of suspensions
Unintended solids in the pumped liquid, such as welding beads, etc., are
kept out of the rotor chamber by a self-cleaning filter installed at the
partial flow inlet. However, such filters are not suitable if liquids with
suspensions are to be pumped. Special measures must be taken to ensure
that the fluid used in the cooling/lubricant circuit of the motor remains
free of solids. The use of the pump type described in the 'Externally
cooled canned-motor centrifugal pumps' section, in combination with an
additionally installed metering pump (Fig. 5.7), enables suspensions to be
pumped. By metering small amounts of clean liquid (approximately
2-10 l/h) into the rotor chamber, a continuous flow in the direction of the
pump is maintained, which keeps the rotor chamber free of solid particles.
The liquid metered into the rotor chamber may be the pure phase of the
suspension, a neutral liquid which is compatible with the pumping
medium, or a substance which must be fed into the system in any case for
process reasons.
Figure 5.7a Canned-motor pump with additional metering pump for suspensions.
98
Figure 5.7b Canned-motor pump with additional metering pump for suspensions,
as shown in Fig. 5.7a.
Pumping of highly toxic or radioactive liquids

Hermetic centrifugal pumps are not equipped with dynamic seals, meaning that hazardous substances cannot escape the system. Instead, a series
of static seals are used with which the risk of leaks increases with the
increasing number of pump stages. The high number of static sealscan be
reduced to only two by means of an additional enclosure around the
hydraulic part of the pump (Fig. 5.8). The two remaining static seals may
also be designed as double safety seals. For this purpose, two O-rings with
a drain channel in between are used; the drain channel is connected to a
sensor which emits an alarm signal. This method can satisfy even
extremely stringent safety standards.
Canned-motor pumps
99
Figure 5.8 Multistagecanned-motor pump with 'barrelled' hydraulic part.
Canned-motorpumps for high differential heads

It is well-known that the efficiency of a centrifugal pump is particularly
dependent on its specific speed
nq = n * "~(Q)/H 3/4 (min-1).
If the Q/H ratio drops to a low level at the preset speed, it is a good idea
to divide the pumping head among several impellers; in other words, the
pump should be of multistage design. A range of nq = 15 to nq = 30 rpm
should be aimed at, as an increase in nq in this region results in a
worthwhile improvement in efficiency. In canned-motor pumps, the pump
shaft is, in terms of statics, a beam resting on one support, meaning that
only the dead weight of the shaft and impellers- not including hydraulic
radial l o a d s - represent a transverse moment. The required impeller
between case and impellers is dependent on the configuration of the slide
100
Figure 5.9 Multistage rotor.

bearings and the tan a which is formed by the triangle created by the bending of the shaft in relation to the theoretical center of the pump (Fig. 5.9).
Most hermetic centrifugal pumps are manufactured of austenitic materials, which requires an additional clearance between the fixed and rotating wear rings. The resulting constructional tolerances lead to gap flow
effects which may, under certain circumstances, counterbalance the advantages offered by multistage design. For this reason, such pumps which are
installed horizontally should not have more than six stages. If more stages
are needed to attain the required delivery head, tandem construction
(Fig. 5.10) may prove the ideal solution. In this case, the entire pump flow
is directed from pump I to the front rotor chamber of the motor, and from
there to pump 2 via the rear rotor chamber after passing through the
motor shaft and the rotor-stator annular gap. In this manner, hermetic
pumps with up to 12 stages with delivery heads of up to 500 m are possible.
Figure 5.10a Tandem hermetic centrifugal pump with fluid guidance via the
motor shaft.
Canned-motor pumps
101
Figure 5.10b Sectional view of a tandem hermetic centrifugal pump, as shown in

Fig. 5.10a.
If still greater heads are required, motor speed can be increased by means
of frequency converters; delivery heads of up to 800 m are then possible.
Tandem pumps may also be equipped with 'barrels', whereby, for example,
a 12-stage pump would require only two static seals (see Fig. 5.11).
Figure 5.11 Tandem hermetic canned centrifugal pump connected in series.
102
Canned-motorpumps for high-pressuresystems

An interesting field of application for canned-motor pumps is highpressure technology in which liquids or highly compressed gases must be
pumped in circuits with pressures ranging from 100-1200 bar. For such
tasks, conventional pumps either cannot or can no longer be operated
with appropriate shaft seals. The technically simplest solution to this
problem is the canned-motor pump. In order to function under highpressure conditions, the entire pump set must be installed in a highpressure enclosure adapted to the specific operating pressures and
temperatures (Fig. 5.12). The most critical component of the pump set is
the thin-walled motor can (wall thickness = 1 ram) which is subject to
both system and delivery pressure and must therefore be supported
accordingly. In the area around the stator, the laminations are used for
mechanical support of the motor can if special measures can be taken to
render the stator plates suitable for this purpose. The motor can is
reinforced with rings of non-magnetic, high-tensile materials in the area
between the stator plates and the front or rear motor cover plates to
support the high inner pressure of the stators. This enables thin-walled
motor cans to be used while keeping electric and magnetic losses to a
minimum and attaining good degrees of efficiency. Special attention
should be given to the motor terminal box. It should be designed to
withstand the nominal pressure of the system in order to prevent liquid
from escaping from the terminal box should there be a leak from the rotor
Figure 5.12a High-pressure version of a canned-motor pump.
Canned-motor pumps
103
Figure 5.12b Canned-motor high-pressure centrifugal pump, as shown in

Fig. 5.12a.
chamber into the stator chamber. If applicable, regulations pertaining to

the use of motors in explosion-proofed areas must also be observed.
Pumping of highly compressed gases. The transport of fluids in gaseous,

supercritical states, for example, used for extraction from vegetable
substances, requires systems with internal pressures of several 100 bar.
Under such operating conditions, gases assume high, liquid-like densities
due to the high levels of compression. The delivery head of centrifugal
pumps is always expressed in meters of the pumped fluid. Thus, the
delivery pressure of such pumps increases with increasing density.
Correspondingly high delivery pressures can therefore be achieved at
normal rotational speeds although the fluid is in a gaseous state. Pumps
designed for applications in this field are similar to those illustrated in
Fig. 5.12 in their basic design, but can also be constructed as centrifugal
tandem pumps as shown in Fig. 5.13. Each spacer between pump and
motor is equipped with a hydrodynamic mechanical seal. The rotor
chamber and the cooling system are filled with water or other. This provides normal conditions for heat transfer, thrust compensation and bearing lubrication. The mechanical seals located between the primary and
secondary circuits prevent uncontrolled introduction of liquid into the
pumped fluid. A compensation pipe provides pressure compensation
104
Figure 5.13b
Semi-hermetic centrifugal tandem pump, as shown in Fig. 5.13a.
Canned-motor pumps
105
between the rotor chamber and the pumped fluid (gas) to relieve the
mechanical seals. A supply vessel, which is either separate or mounted on
the pump, is installed in this connecting line in which the liquid level is
monitored by means of maximum and minimum level sensors. When the
minimum level is reached, liquid is automatically added by a metering
pump. This provides a simple means of monitoring the leakage of the
mechanical seals. These types of pump are known as semi-hermetic centrifugal pumps, as they are hermetically sealed externally, but require
slide ring packings between the internal and external circuits. In place of
the solution illustrated in Fig. 5.13, a hydrodynamic shaft seal may also be
used (Fig. 5.14). As shown in Figs 5.12 and 5.13, the motor is equipped
with a separate cooling circuit for removal of heat resulting from energy
losses in the motor. A hydrodynamic shaft seal, designed as an open
impeller with straight blades, is installed in the spacer between pump and
motor case. This prevents liquid from escaping from the motor to the
pump, or, in the reverse case, gaseous fluid from the pump from entering
the motor chamber. The advantage offered by this construction is that the
pressure compensation line with buffer vessel between the pump and
motor parts is eliminated. Also, because slide ring packings are not used,
there is a higher resistance to wear and consequently longer service life is
achieved.
Self-priming centrifugal canned-motor pumps

A number of operating conditions require that the pumps used must be
capable of pumping gases or liquid-gas mixtures. Such conditions exist for
the most part when the pump suction must be placed geodetically higher
Figure 5.14a High-pressure version of a canned motor centrifugal pump with

hydrodynamic separation between pump and rotor chamber (schematic drawing).
106
Figure 5.14b Canned-motor pump designed for 1200 bar system pressure with
hydrodynamic shaft sealing, as shown in Fig. 5.14a.
than the liquid level. Also, situations exist when tanks or tank wagons
must be emptied which, for reasons of safety, are not equipped with lowlevel drains, meaning that the tanks must be emptied from the top.
Operating conditions may also exist when fluids contain gas or have a
tendency to gas off, meaning that the fluid becomes a liquid-gas mixture.
Two systems are suitable for use under such circumstances: the side
channel pump and the impeller cell scavenging pump.
The side channel principle

The side channel pump is frequently used in process plants. It shall be
assumed at this point that the function of such pumps is generally understood. Because the duties for which these pumps are utilised generally
involve hydrocarbons, solvents and liquefied gases, it was logical that they
should also be hermetically constructed by using canned motors. As
shown in Fig. 5.15, the first stage is usually equipped with a standard
priming impeller. This is due to the more favorable NPSHR values
achieved by standard impellers as compared to self-priming impellers.
Canned-motor pumps
107
Figure 5.15 Self-primingcanned-motor pump, side-channel design.
The cell scavenging principle

A disadvantage of self-priming centrifugal pumps which operate on the
side channel principle is that the pumped liquid must be free of solids.
This is because solid particles cannot be conveyed through the narrow gap
between the impeller and stage casing. In a cell scavenging system,
standard impellers are used which offer considerably higher resistance to
wear from solid particles in the pumpage. The principle on which these
pumps function is described below with reference to Fig. 5.16.
The rotating impeller generates a scavenging flow S, generated by its
blade tips. The scavenging flow is directed by baffle plate F into the
impeller cells rotating behind the baffle plate. The impeller cells form
mixing chambers M, similar to those in a jet pump. The air/gas-liquid
mixture is drawn into and blended with pump flow S and directed to the
Figure 5.16 Functional principle of a self-priming centrifugal pump with

impeller cell scavenging.
108
Figure 5.17b Sectioned self-priming canned-motor pump, as shown in Fig. 5.17a.
Canned-motor pumps
109
diffuser. A large separation chamber is located on the pump pressure side

from which air escapes upward through the discharge. The pumped liquid
falls down to be re-introduced into the scavenging process. This pump
type is superior to the side-channel pump type in the following respects:
9 improved efficiency for pumping of clean liquids because of the
standard priming centrifugal impeller
9 better resistance to wear and dirt
9 lower noise emission.
This principle is highly suited for pumping of gaseous liquids and liquids
containing solid matter.
Safety and monitoring equipment

Canned-motor pumps are primarily used in plants in which hazardous,
toxic and/or explosive mixtures are made or carcinogenic substances must
be transported in pipelines. Such pumps should be equipped with safety
and monitoring equipment of the highest possible standards. These
efforts made for the protection of personnel and machinery are justified
under all circumstances.
The use of canned motors in hazardous locations.

Explosion protection according to European Standard
'EN' I21
In cases in which canned motors must be operated in hazardous locations,
the statutory provisions as expressed in national and European directives
apply; supervision in such cases is carried out in Germany by the
Physikalisch-Technische Bundesanstalt (PTB; Federal Institute for Physics
and Technology) in Braunschweig. The German and European explosion
protection standards stipulated for these motors are contained in the
following DIN standards:
9 DIN EN 50014:
9 DIN EN 50018:
9 DIN EN 50019:
Electrical equipment for hazardous locations,

general regulations.
flameproof enclosure 'd'.
increased safety 'e'.
These standards are also included in the directives as a VDE regulation

in terms of VDE 0022 under the numbers VDE 0170/0171, Section 1/5.78,
VDE 0170/0171, Section 5/5.78 and VDE 0170/0171, Section 6/5.78. The
110
national regulations formerly valid according to VDE 0170/0171, Section

1/12.70, were made ineffective as of May 1, 1978.
Canned motors already certified in accordance with the provisions
according to VDE 0171 may continue to be manufactured and operated
without restrictions.
European Standard EN 50014 to 50020
EEx de II C T 1 ... 6
meanings:
9 E
9 Ex
9 de
9 II
oC
9 T1 ... T6
Motor complies with European standard

Explosion protection
Combined type of protection 'flameproof enclosure' and
'increased safety'
Group of electrical equipment which cannot be used in
firedamp hazardous locations (underground working)
The highest class regarding the limit gap established by
experiment with type of protection Ex d; this is suitable
for all gases and vapors; limit gap width MESG < 0.5 mm
Temperature class for the applicable max. surface
temperature.
T a b l e 5.1
Maximum surface
temperature (~
450
300
200
135
100
85
Temperature
class
T1
T2
T3
T4
T5
T6
.,
,,,,
i,
|.,..
The lowest ignition temperature of the explosive atmosphere in question must be higher than the maximum surface temperature of the motor.
According to the new type of protection Ex e for the rotor chamber,
warming of the rotor must be taken into account. The most hazardous
accident in this context occurs when the impeller is blocked at operating
temperature. In this case, the rotor reaches its maximum permissible limit
temperature as stipulated by the applicable temperature class within the
so-called tE time.
The tE time of standard Ex e motors is known. It stipulates that the
motor must be protected by a thermally delayed excess-current release
which releases within the time specified on the nameplate.
Canned-motor pumps
$
20
lO
111
l
l
Imml l l
n a n m n l l m m
'~un u
ii
n m m am
i~-a i
i
m n n in
il~Olll
Immlm!
mmmml
"E
~tu 2
]A/IN
Figure 5.18 Minimumvalues for time tE for motors in relation to starting current
ratio IA/IN.
Therefore, depending on the starting current ratio IA/IN, minimum
values as defined in EN 50015 must be observed (Fig. 5.18).
The surface temperature of the rotor must therefore be measured when
the rotor is blocked, and the temperature of the pumping medium added
to this value. To ensure the observance of the regulations for explosion
protection and for supervisory purposes, the legislature has enforced
additional safety guidelines for canned-motor pumps. In the test certificates issued by the Physikalisch-Technische Bundesanstalt PTB (Federal
Institute for Physics and Technology), conditions are stipulated which must
be met for the operation of such pumps in hazardous locations. These are
as follows:
9 For reasons pertaining to safety, the rotor chamber must always remain
filled with pumping liquid. For this purpose, liquid-level sensors or
different measures of at least equal effectiveness must be installed in
the system to ensure that the motor is operated only at sufficient liquid
levels.
9 The coolant/lubricant flow must be monitored by temperature sensors to
prevent maximum temperatures from being exceeded. It must be ensured
that the temperature of the coolant/lubricant does not exceed ...~
Monitoring devices for canned motors

Level~temperature monitoring. To fulfil the requirements of PTB
(Federal Institute for Physics and Technology), a level/temperature monitoring device was specially developed (Fig. 5.19) which consists of a ther-
t,~
9tnmdmdMIn und k l ~ m o n
~ N 3 0
Iketa
r~
~ .......
~_~
~ ....
'r
"
~ i:z~---iI
I /
t~
Thonnoe~t T 30
N lO
OeQlINr . l k t "
W
.e'~,"
0
m
0 Nhlm~ll
Ill
9uuemTo~lm
-T30
Figure 5.19 Level-temperature monitoring device.
Canned-motor pumps
113
mostat, a level sensor and a switching amplifier. A liquid expansion or

resistance thermometer is used as a thermostat. The limit temperatures
stipulated in the technical specifications of the corresponding canned-motor
pumps can be set in the connection head of the unit using a thermometric
scale. Figure 5.20 shows the circuit diagram for the level/temperature monitoring device and a so-called sensor-controlled double motor can monitoring system, which may also be installed.
The Niveaustat (level monitoring device) consists of a float equipped
with a magnet which moves on a tube in which a reed contact is installed.
When the liquid level rises or falls, the reed contact is activated by the
magnet. The switching amplifier is designed for protection type 'intrinsic
safety' (EExia) IIC/IIB for ambient temperatures of up to 50~
An opto-electronic monitoring system (Figs 5.21a and b) may also be
installed instead of the level monitoring device described above. Optoelectronic sensors are used primarily for low liquid densities (< 0.5
kg/dm 3) and high operating pressures (> 25 bar).
The level and temperature monitoring devices for canned-motor pumps are
also highly recommended for applications in which no regulations are
enforced with regard to explosion protection. Level monitoring ensures that
the pumps can be operated only at sufficient liquid levels. For liquids
which are pumped at temperatures near the boiling point, (such as liquid
gas), the level monitor should be installed at Zemin 4" 0.5 m in the suction
line in order to ensure sufficient pressure at the impeller inlet and to
protect the pump against damage (cavitation).
Zemin
- NPSHR + Hvs [m]
NPSHR
= Net positive suction head required [m]
nv~
= frictional resistance in suction line [m]
Temperature monitoring ensures that the pump is shut down in case of

excessively high temperatures. This protects the pump from more serious
damage. The temperature monitoring device also functions as a reliable
indicator that the pump is operating within the permissible range.
Temperature monitoring can also be provided in the form of a maximumminimum circuit to prevent the pump from being started before the fluid in
the pump has reached its correct viscosity, e.g. when the solidification point
of the fluid is above ambient temperature (e.g. fatty acids). Motor protection
may also be provided by thermistors or resistance thermometer PT 100
installed in the windings. Before delivery as an 'explosion-proofed' motor,
each motor type is tested by the PTB (Physikalisch-Technische Bundesanstalt;
Federal Institute for Physics and Technology) to ensure compliance with the
4~
1
L2 z.S,
V/SOHz
~' IW 1~2
':=
220V
" 'T i~=P -~S~
~)
.EXumo
P30 "I)WS"
;L" - e
,<~ E,,
I..,,~E:
t,,.,,~ ',,,,'"'~""','
"
'
,.~
@-J
T3Q
P~
'
limml
1,41
Figure $.20 Flowdiagram for integrated electronic pump monitoring.
'</".lilt 'i"
,,Imtm
Canned-motor pumps
115
~,~ "~
I./'1
"l
J
I
34
J
, ~13,5
-55
-
e9
Figure 5.21a Opto-electronic sensors.

regulations. The motors are provided with a test certificate after the test has
been successfully passed (Fig. 5.22). The manufacturer must construct each
unit according to the tested unit and carry out exactly defined routine check
tests on each unit.
Monitoring through vibration measurement. Canned-motor pumps

operate at low vibration and noise levels. Noise levels in such pump sets including both pump and m o t o r - rate an average noise emission level
which is 8 dB(A) lower than the equivalent pumps of conventional design
- w i t h o u t drive motor. In cavitation-free operation at normal operating
values, abnormal vibration indicates disturbances. Vibration meters may
therefore be used to form an accurate assessment of the condition of the
bearings and rotor. Absolute or relative oscillation may be used for
measurement for this purpose. Measurement of the relative oscillation is
more complex, but also more informative than measurement of absolute
oscillation. Relative oscillation is measured and monitored as a rotational
movement of the shaft in the bearing. However, for assessment of the
operating condition of the pump, measurement and monitoring of
absolute bearing vibration according to the permanent-dynamic moving
coil measuring method is sufficient in most cases. For this measuring
method, transducers are mounted on the motor casing (Fig. 5.23). A coil
is fixed to an elastic membrane such that it moves into the air gap of a
permanent magnet. If vibrations now act upon the transducer from the
outside, the body of the transducer oscillates with the pump part to be
measured and monitored, while the moving coil mounted on the elastic
membrane remains motionless. Relative movement results between the
116
Figure 5.21b Opto-electronic level monitoring device, as shown in Fig. 5.21a

(Phoenix-Analytec Marketing GmbH).
coil and the magnet field, which induces voltages in the coil directly
proportional to vibration velocity. The measuring voltage provided by the
transducer is connected to an electronic unit for processing the measured
values; this unit either triggers an alarm or shuts down the pump if a set
limit value is exceeded.
Rotor position measuring device RPM. Hydraulic compensation of axial

forces is particularly important in canned-motor pumps, as these pumps
are not equipped with hydrostatic thrust bearings. The slide rings in the
bearings have locating and control functions only; they are not designed
117
Canned-motor pumps
Physikamisch-Technische Bundesanstalt
KONFORMITATSBESCHEINIGUNG
(I)
PTB Nr. Ex-91.C.3}12 X
(2)
(3) OiosoBeec,N d n ~
Spaltrohrmotor T]
v-2
(4) der F~rma

(5)
I Betdebsmltteb emwle die vemchkKlen~m zu.lemlgen
(e)
Bundemlrmtalt be~-'helnlgt als Pr0fMelle nach Artlkel 14 d

Gemeh1~ichMten vain 18. Dezember 1975 (761117/EWO) d
~fl['5 (VD.E 017010171 T l l l l i l r A I l g e m e i n e

Be=
,,A3 (VDE 0170/0171 T e ~ W
Druekfeate Kal
9A3 (VDE 017010171 Toil
. ) Erh6hte Sicher
n=~k~ ~
Eao~ ~AK e ~ e m ~ U i i l m N e .
O~ F a l r a m d ~
(7)
EEx de [[C T:]
(8) Dec I-ler~oger ~ d
a ~ ,
dab jades derart gekennzeidmme [ k i l l IrntttelIn aelner Bauart
mlt den in d
e
r
/
m
l
a
g
~
m
(~r
mid
dab dte vorgelmhdeben
(9) Oas elekldsche B4Itl~II~IIYMtMdaft mit dam hter abgednlcktl~t ~ f t l i c h e n
Uflt~ngazetchon gem&8Anhang II dot Richtlinledes Raleevom 6. Fobruar 1979 (79/196/EWG) gokormzMchnet
warden.
Im Auftmg ~,
Brnurmchweig, 26.0~t. 1991
':.
I ~ ~ e d ~ a
. . . . . . . . . . . . . . . . . . . . . . . . .
Figure 5.22
Test certificate for canned-motor pumps.
118
Figure 5.24 Rotor position measuring device RPM.
Canned-motor pumps
119
to absorb any significant axial thrust. If axial thrust compensation is

disturbed through cavitation, for example, considerable friction damage
to the pump may result. Such damage can be avoided when a rotor
position measuring device is mounted on the canned-motor pump
(Figs 5.24 and 5.13). This device registers changes in the axial position of
the pump motor rotor inductively and without contact. When the
maximum permissible deviation is exceeded, either an alarm is triggered
or the pump is shut down. The device consists of a transducer, with supply
lead and electronic evaluation system. A thin shaft is mounted on the end
of the rotor. Over the shaft, a support pipe is flanged tightly to the pump
casing onto which the transducer is clamped. The support pipe and shaft
are made of non-ferromagnetic (austenitic) materials. The unit is
designed for measuring from +/-0.5 mm to +/-1.5 mm, the common
range for monitoring of axial play.
Pressure monitoring in the motor-stator chamber. The terminal box is
sealed with cast resin and connected directly to the pressure-tight stator
chamber, in which there is a pressure switch (Fig. 5.25). If the motor can
is damaged, for example because of damage to the bearings or abrasion
Figure 5.25 Pressure switch built in the terminal box.
120
from solids in the coolant and lubricant flow, the pumped fluid may
escape into the stator chamber. This results in an increase in the internal
pressure in the stator chamber. Even relatively low overpressure
= 1.0--1.4 bar is sufficient to disable the motor by means of the pressure
switch. The motor windings are damaged by the pumpage in most cases,
but serious environmental damage is prevented as the liquid cannot
escape into the atmosphere.
Cos ~p motor load monitor. Monitoring of the inductive load of the
canned motor provides a direct source of information on the operating
condition of the motor and consequently the operating performance of
the pump set. The power consumption of the motor is directly related to
the QH behaviour of the pump (Fig. 5.26). Canned-motor pumps must,
however, be operated within a specific pumping range. To the right of the
pump maximum flow specified by the manufacturer, drastic increases in
NPSHR and overloading of the motor may result. It is also possible that
the difference in pressure of the partial flow to the flow in the rotor-stator
gap is no longer sufficient and therefore has no effect. To the left of
minimum flow, there is an asymptotic increase in NPSHR, which means
~p
0.//
NPSHR
cos'f
B
rain
Qmin
~ . ~[Qtony
I cos
max.
KW
Figure 5.26
Q-H diagram with corresponding cos ~pcurve of drive m o t o r .
Canned-motor pumps
121
that damage to the pump may occur through cavitation. In addition, if the
pump flow is below the minimum, coolant-lubricant flow is thermally
overloaded, which may lead to an excessive increase in motor temperature. A cos tp monitor can be used for monitoring the angle of phase
difference which changes with the motor load. A red LED indicates that
the cos ~0 monitor is ready for operation. The motor load monitor
operates on the static current principle; i.e. when a set value (e.g.
pumping range between Qmin and Qmax) is reached, the two output
relays for cos tp rain and cos tp max are energized, and two green LEDs
light up. The values for cos ~ rain and cos tp max which correspond to
Qmin and Qmax can be set on the unit by means of a knob (Fig. 5.27). The
set values refer to the rated voltage. If the relevant level drops below or
exceeds these values, the corresponding LED lights up. The output relay
for cos 9 min and cos ~ max is de-energized at the same time.
To prevent the cos tp monitor from reacting on starting the motor,
starting time (which is extremely short) can be set at the unit. The cos tp
monitor does not function during this period. A reaction time of 5 seconds
is also provided in order to prevent the cos ~ monitor from reacting to
temporary deviations from the set values. The function diagram of the cos
tp monitor is provided in Fig. 5.28.
Figure 5.27 Cos ~pmotor load monitor according to IEC 255, VDE 0435.
122
cos
COS ~mox
r f m,~
. . . . i-- I- -._
-
CO S ,pmo x ~
28
2~ ......
I'.
'
I
I
;
i
.'-
-~---
I
I
.11
I
Ii
II
I!
', "
i"",
I
i
L,2
'
',. .
_l
. .
L,~_I
""
I
I
| ,,,
I
I,
I
I
~-
._,,_:.i
.....
,,
I
I
I
I
I
I
I
,,
......
I
i
;
cos)0min
'
18
..
L, J
F : ....
L,i-I
tl = AnlaufQberbrOckungszeit
t2 = R e a k t i o n s z e i t
Figure 5.28 Functional diagram for cos tp motor load monitor according to
Fig. 5.27a [3].
Double wall security system D WS [4]

The increasing emphasis placed on safety requirements pertaining to
pumping of hazardous substances has resulted in an innovative safety
monitoring system which was given the name 'DWS'. The principle of this
system is based on the use of a double-walled can instead of a single motor
can. In canned-motor pumps, the cans are arranged at a minimal distance
from each other in the gap between rotor and stator, and between the
outer and inner rotor in permanent-magnet driven centrifugal pumps.
The area between the cans is connected to a sensor by a hole in the pump
casing (Figs 5.29 and 5.30); this sensor reacts in case of damage to the
inner or outer motor cans. Pressure switches, capacitive or inductive
humidity probes, heat conductivity sensors, or switching systems based on
analytic measurements can be used. A wide range of control and monitoring options is available for this purpose. For example, optical and
acoustic signals can be released which shut the pump down, and valves
installed in the pipe work can be automatically dosed. Interruptions in
operation can also be avoided by means of a standby pump being
Canned-motor pumps
123
Figure 5.29 The DWS system as implemented in canned-motor pumps.

automatically activated when an alarm message is given; the defective
pump is then shut down. The monitoring and control options can be
expanded as required.
Figure 5.30 The DWS system as implemented in centrifugal pumps with

permanent magnetic clutch.
124
DWS in canned-motor pumps. The already high safety standards offered

by canned-motor pumps with double-walled hermetic shells (motor can
and pump case) between the product and the atmosphere (Fig. 5.29) can
be heightened further by means of a second static motor can. This is of
particular interest in conjunction with highly toxic substances. In addition,
the DWS system also forms an ideal bearing monitor which functions
immediately in case of bearing damage. The DWS system can also reduce
downtime and costs associated with canned-motor pumps, as the windings
are further protected against liquid leakage, which in the past generally
led to total failure. Because explosion-proofed versions of canned-motor
pumps are most often used, conditions for repair by the operator are
substantially better.
D WS in hermetic centrifugal pumps with permanent-magnet drive. Until
recently, the scope of application for centrifugal pumps with magnetic
drives encompassed environmentally hazardous but not highly toxic substances. The reason for this was the coupling can, which acted as a single
shell between product and atmosphere. If the can was damaged, liquid
would escape from the pump directly into the surroundings. The DWS
system as shown in Fig. 5.30 allows such pumps to be used for hazardous
substances also.
Required and recommended monitoring devices

Explosion protection. Stipulated by the legislature for all canned-motor
pumps which are used to pump liquids which may lead to the formation
of explosive mixtures or operated in hazardous areas subject to the
regulations for explosion protection.
Level-temperature monitoring. Stipulated by the legislature for all
canned-motor pumps which are operated in hazardous locations.
Recommended by the manufacturer: installation also in pumps not
subject to the regulations for explosion protection to avoid operating
errors and damage.
Monitoring of motor winding temperature by means of thermistors or
RTDs: recommended by the manufacturer, as this is the easiest and most
efficient method for motor monitoring.
Monitoring through vibration measurement. Recommended by the
manufacturer for extremely expensive pumps, i.e. in very high-pressure
applications.
Canned-motor pumps
125
Rotor position measuring device RPM. Recommended by the BITC

(Bureau International Technique Chlorure; international association of
chlorine manufacturers and processors), for all canned-motor pumps
used for pumping chlorine. Recommended by the manufacturer for
expensive, complex pumps and pumps which play a major role in the
relevant process.
Pressure switch. Recommended by the manufacturer for pumping fluids
which are not electrically conductive, for example oils, and which may
enter the stator chamber should the motor can be damaged.
Cos ~p motor load monitor. Recommended by the manufacturer if
flowmeters are not used to ensure minimum and maximum flows are not
exceeded.
Double wall security system DWS. Recommended by the manufacturer
for all canned-motor pumps and centrifugal pumps with permanent magnetic drive used to pump highly toxic substances.
Bibliography
1. Kr~imer,R. and Neumaier, R.: Kreiselpumpen und rotierende verdriingerpumpen
hermetischer bauart, Werksliteratur, Hermetic (1986).
2. Faragallah, W. H.: Seitenkanal-strSmungsmaschinen; Neumaier R.: Leckfreie
seitenkanalpumpen mit spaltrohrmotor, Antrieb, 254-311.
3. Dold, E. and Sfhne, K. G.: 7743 Furtwangen, Werksliteratur.
4. Kr~imer, R. and Neumaier, R.: Die doppelwandige sicherheit, Chemische
hldustrie, No 3, (1987).
CHAPTER
SIX
Standardized chemical
pump with canned motor
in flameproof enclosures
H. Schreyer
Introduction
In terms of leak-proof pumping, operating reliability and service life,
many conventional centrifugal pump systems no longer meet today's
exacting requirements. Users of standardized chemical pumps and magnetically coupled pumps have been complaining about a number of
recurring defects.
Shaft sealing
Most conventional chemical pumps are fitted with mechanical seals.
Mechanical seals, however, are subject to permanent wear depending on
the liquids to be pumped, which limits service life and precludes
hermetical sealing, as a lubricating film is needed at the sealing faces.
Antifriction bearings
The service life of antifriction bearings depends to a considerable extent on
the operating conditions. In the case of grease-lubricated bearings, the calculated nominal service life of the antifriction bearings is often reduced even
further by the service life of the lubricant and environmental conditions.
Pump shaft
In developing a shaft for conventional pumps, designers must meet conflicting demands. On the one hand a sufficiently large shaft overhang is
necessary to accommodate double-acting mechanical seals, on the other
hand large shaft overhangs increase shaft deflection, due to the radial
forces in the pump hydraulics. This in turn increases wear at the mechanical seals and thus causes defects.
128
Operators of conventional pumps know that these disadvantages mean

short service lives, recurring costs for pump downtime, dismantling, repair
and plant downtime (life cycle costs).
As the above-mentioned elements can be dispensed with in canned
motor pumps, this pump type not only makes for a maximum of plant
reliability and operational reliability, but also allows considerable savings
in current operating expenses.
This paper describes the features of a range of canned motor pumps
with a novel concept of explosion protection, universal suitability, pump
selection, installation and maintenance.
Design
The standard design is based on a combination of the well-known hydraulic
system of standardized chemical pumps in accordance with DIN 24256 or
ISO 2858 and a canned standardized drive unit.
The driving torque is transmitted by the electromagnetic field
generated in the stator winding, which acts on the pump rotor via the can
(Figs 6.1-6.4).
The can protects the stator chamber, winding and the environment
from the pumped liquid. It is hermetically seal-welded by plasma welding
prior to assembly. To minimize eddy current losses caused by the rotating
electromagnetical field penetrating the can, the can is series-produced in
nickel-based alloy 2.4610 (Hastelloy C 4). Mechanized and fully reproducible operations in combination with suitable control and inspection
measures ensure a maximum of safety and operational reliability in the
future pump application.
Thanks to a special calibration method, the can is brought into contact
both with external supporting elements outside the can and the stator
pack during assembly. This makes it possible to absorb internal pressure
- resulting from system and pump pressure- outside the can and to establish a close contact in the area of the stator pack, so that electrical heat
losses can be dissipated by the pumped liquid.
The stator pack is located in an integrally cast motor housing which is
designed to withstand a nominal pressure of 25 bar in grey cast iron GG
25 or a nominal pressure of 40 bar in nodular cast iron GGG 40.3.
For a simpler connection between motor and hydraulic parts, the
dimensions of the motor flanges have been standardized for all motor
ratings available. The bearing bracket lantern also houses a heat exchanger which dissipates the heat losses from the rotor chamber to the
hydraulic system.
Standardized chemical pump with canned motor

129
r~
t~
..=
t~
130
Figure 6.2
(a) Production of can. (b) Finished can.
131
Figure 6.3 Can with external supporting elements.

The rotor is hermetically sealed against the pumped liquid and runs
only in radial antifriction bearings made of silicon carbide ceramics. The
pump set has no thrust bearings.
Individual pump hydraulics are connected to the selected motors with
suitably adapted casing covers.
The following material versions are available:
9 In standard form, the hydraulic components are supplied in CrNiMo cast
stainless steel 1.4408 and the motor components in the equivalent rolled
material 1.4571. The antifriction beatings can be supplied either in pure SiC
or SiSiC. The standard version can withstand a nominal pressure of PN 16.
9 The pump is also available in heat-resistant cast steel GS-C 25 and
Noridur 9.4460, a high-alloy austenitic ferritic CrNiMo cast steel specially developed for handling liquids containing chlorides. In that case the
wetted parts in the motor component are made of austenite 1.4539.
These materials can withstand a nominal pressure of up to PN 25.
9 PTFE is the standard material used for static sealing elements which
come into direct contact with the handled liquid.
9 At the moment, there are 14 motor ratings (2- and 4-pole) available
from 1 to 28 kW. The range is to be extended up to 55 kW in the future.
9 These motor ratings currently allow duty points with flows of up to 160
m3/h and heads up to 150 m.
Explosion protection on canned motor pumps

As canned motor pumps are predominantly used in the process industries
with their multitude of hazardous locations, the pump sets as electrical
apparatus have to meet the requirements of explosion protection
(Figs 6.5-6.9).
bo
r~
17
Figure 6.4 Standardized modular construction system hydraulics/motor.
14
133
Clearance
luol]
,,,,,
-"1
Figure 6.5 Scheme: Flameproof enclosure.
F-'"#----:-'---:-"--'-q'--'l
!I,
I~-- . . . . . . .
iir
i!
-'ill
Figure 6.6 Scheme: Increased safety.

In order to illustrate the difference between the concept of explosion
protection commonly applied to canned motor pumps so far and the solution advocated in this chapter, it is necessary to go into the difference
between the protection types 'flameproof enclosure' and 'increased safety'
as per European standards DIN/EN 50018 and 50019.
Flameproof enclosure EEx d (DIN/EN 50018) [1]. All parts which can
ignite explosive mixtures are located in a flameproof enclosure whose
sealing surfaces, cable entries, shaft ducts etc. are made ignition-proof by
observing certain gap lengths and widths. This design prevents the transmission of an internal explosion to the explosive atmosphere surrounding
the enclosure.
Increased safety EEx e (DIN/EN 50019) [1]. Special measures are taken
to avoid excessive temperatures, arcs and sparks, e.g. by means of reinforced bearings, reinforced terminals and specially selected insulating
materials. To preclude the risk of ignition caused by excessive temperatures, suitable monitoring equipment is required. In accordance with this
European standard, the special protection type Ex s used in former VDE
regulations exclusively for canned motor pumps must no longer be used
for the approval of new machines.
134
In terms of explosion protection, conventional canned motor pumps

used to be divided into two separate zones. The stator chamber- between
motor housing and c a n - used to be classified as 'flameproof enclosure'.
The rotor chamber, by contrast, was approved as 'increased safety'. Due
to the 'increased safety' classification, the rotor chamber had to be monitored by special safety equipment. In practice, this means that the filling
level in the suction and discharge lines must be checked before and during
pump operation to ensure that the rotor chamber is filled with liquid.
Also, the temperature in the rotor chamber must be monitored to ensure
that the admissible temperature limits are observed. To this end, a temperature sensor is installed near the journal bearing at the far end of the
motor which monitors the temperature there.
The signal lines of both monitoring systems must be routed separately
from the hazardous location to the control room and connected to the
circuit of the motor protection switch.
EEx d
EEx e
Figure 6.7 Former types of protection in canned motors.

EEx d
Figure 6.8 Protection type of the EXACHEM pump.

R
135
4OOO
r//////
Q
1330
550
2JO
. . . . . .
All
"////////""////////j
i
i
-20"C
"
~ ~
I
~
I
"
l=, l -
'~ 'r
~I ~I
"IF
I-,
4"
m
O
Figure 6.9 Electrical resistance on PTC thermistors as a function of

temperature [3].
By contrast, explosion protection on the canned motor pump described
here is based only on the protection type 'flameproof enclosure' and permanent monitoring of the winding temperature.
As on the normal EEx d standardized motor, the complete internal
chamber of the canned motor is a 'flameproof enclosure'. All gaps leading
outside, particularly the gap at the shaft duct are designed in such a way
that an internal explosion cannot be transmitted to the explosive atmosphere surrounding the enclosure. In this design, the can itself is not part
of the flameproof enclosure in terms of explosion protection, i.e. the
flameproof status is guaranteed even in the event of a defect in the can.
One of the prerequisites for Ex approval is the so-called type test, in
which a specially prepared motor is subjected to an ignition test. The
motor has to pass at least five tests before Ex approval is granted.
The second feature of explosion protection is the permanent monitoring of the winding temperature by means of one PTC thermistor for each
phase. The thermistors are embedded in the stator winding. Direct contact
with the winding makes for favourable heat transmission and thus short
response times of the thermistors.
136
The thermistor curve shows a steep rise of resistance in a certain temperature range [3], thus supplying a reliable switching signal. The thermistor measuring circuit is connected with a PTC tripping device which
actuates the motor protection switch. This device is located in the central
control room of the plant, outside the explosion-proof area.
In a simple way, the thermistors therefore perform three important
safety functions:
9 protection of winding against excessive temperatures
9 compliance with maximum admissible surface temperature as defined
by explosion protection requirements
9 tripping of pump set in case of excessive inflow temperature or zero
flow due to plant malfunction.
For each motor rating, temperature rise tests are performed which eventually define the operating limits to be indicated on the motor nameplate.
During the approval procedure, the following malfunctions are simulated, in order to determine their effect on pump behaviour:
(1) Inadmissible temperature ofpumped liquid (Fig. 6.10a). In this case
the winding temperature rises until it reaches the switching temperature
of the thermistors which then trip the motor.
In all cases examined, i.e.
9 temperature far/slightly above temperature limit
9 high/low temperature change rate
9 high/low winding temperature at beginning of test.
the thermistors trip the pump set before inadmissible temperatures are
reached anywhere on the motor.
(2) Inadmissible pump power consumption (Fig. 6.10c). If the motor is
overloaded by inadmissibly high pump power consumption, either due to
incorrect selection or as a result of inadmissible operating data, the winding temperature will rise, and the thermistors will trip the pump set before
the temperature limit is exceeded. This takes place regardless of by how
much the power input limit is exceeded, or whether power input increases
slowly or quickly.
(3) Interruption of cooling circuit (Fig. 6.10b). If the cooling circuit is
completely or partly interrupted, the temperature in the winding will also
rise to the tripping temperature of the thermistors, which will prevent
inadmissible motor heat-up.
(4) Badly vented rotor chamber. This also impairs the cooling function.
Tripping occurs in analogy with interruption to the cooling circuit.
Operation against closed discharge line (Q = 0): As the motor power
used in this duty point is low, the winding temperature limit is not achieved
~3
mI
'~, ~
i
(Do) J,
(;:)0) ,,L
=S
~"
'
....
I
I
!
t,
'
mmmammmmmmmmmmm
munmnmnmmmmmmn
mllmmmmmmmmmmnnnmn
mmLmNnnmilmmunmmmmn
lanmm~mmmnmm
m~qHmmammmmmmm
inmm~~mmw~qumumu
il
....
}-,
_m
.~.
d"
(Do) s
137
I-4
o
o
(t,}
e~d)
9 ,,,.w
C.,)
.=.
e.0..0
o
omO
.=.
,,...0
(:%,
=i
l-0
o ,,,.~
.=.
b]
9 ,.,.0
~.o
,m~
Qmm
138
at all or only very slowly, because convection and thermal conduction of

the pipelines dissipate enough heat to cool the motor. If this is not the
case, the pump set reacts as for inadmissible liquid temperature [2].
The canned motor pump design described here has obtained the following EEx approvals: EEx d e II B T4 and EEx d e IIB T3.
Because the terminal box is certified with 'increased safety', the
approval contains the EEx e sign.
Temperature class T4 can cope with an inflow temperature of T --- 100
~ class T3 with T = 130 ~ The individual motor sizes are approved for
three temperature/power levels each, which, in practice, simplifies motor
selection [3, 4].
Cooling principle
The heat losses of the motor are dissipated from the rotor chamber indirectly. Pumped liquid is routed from the discharge side of the pump
hydraulics through a heat exchanger installed in the rotor chamber, where
it is heated up. It is then returned to the zone of low pressure behind the
back shroud of the impeller and added to the main pump flow. The pumped liquid in the rotor chamber is permanently circulated by means of a
special pumping device, so the heat generated in the stator, can and rotor
is dissipated to the pumped liquid and transported to the heat exchanger
by the circulation flow. Therefore heat is exchanged indirectly to the
outside (Figs 6.11 and 6.12).
This principle has several advantages.
9 Thanks to low flow velocities, pressure losses in the rotor chamber are
much smaller than in the case of direct cooling. A relatively high
pressure is therefore applied by the pump hydraulics in the rotor
chamber, which leaves a relatively high safety margin against premature evaporation of the heated liquid.
9 If the pumped liquid contains impurities, these have to pass several
curves and bends before they reach the rotor chamber. The largest portion is bound to pass through the heat exchanger, thus reducing the risk
of contamination and wear in the rotor chamber to a minimum.
9 The heated cooling flow returning from the heat exchanger is not
added on the suction side of the pump but on the discharge side,
directly behind the back shroud of the impeller. This prevents the
pumped liquid on the suction side from heating up, evaporating earlier
and therefore increasing the risk of cavitation.
139
7
. .
5
4
3
2
50
80
I00
Temperature of the fluid tf ~
Figure 6.11 Motor power available at given liquid temperature, e.g. motors with
2.2; 4; 5.5 kW.
Figure 6.12
Routing of cooling flow.
140
Operating behaviour [5]

As this canned motor pump uses the same hydraulics as the standardized
chemical pump, the operating characteristics are identical in many aspects.
However, there are a number of particular characteristics which can be
described as follows.
Selection of a canned motor pump, just like selection of a standardized
chemical pump, is based on the pump power input required for a specified
duty point. However, in contrast with the selection of a standardized
chemical pump, allowances have to be made for rotor friction in the liquid
and eddy current losses resulting from the penetration of the rotating
electromagnetical field through the can.
Another factor to take into account is that the temperature of the liquid
in the rotor chamber and the heat exchanger is higher than that of the
main flow. When selecting the pump set, it is therefore important to bear
in mind the consequences of liquid heat-up and to prevent liquids with
steep vapour pressure curves from evaporating in the rotor chamber,
especially near the bearings. With the help of a computer this can be
verified easily during pump selection.
Tests have shown that in terms of energy balance the overall efficiency
of the canned motor pumps described in this chapter is on average 5-10%
lower than that of pumps with single-acting mechanical seals.
Other special selection features such as selection of a shaft coupling for
standardized chemical pumps and analysis of start-up conditions for magnetically coupled pumps need not be observed on canned motor pumps.
Standardized chemical pumps can only be optimally run in an flow
range of about 10-120%, due to design features such as overhung shaft
and shaft seal. Canned motor pumps, by contrast, can be operated for
relatively long periods of time both in the extremely throttled operating
range and at 30% overload, due to a short shaft overhang, widely spaced
antifriction bearings and complete balancing of axial thrust, provided the
pump inlet conditions generated by the plant do not lead to premature
cavitation. This means that in most cases, complex and costly throttling or
bypass systems can be dispensed with (Fig. 6.13).
Compared with other pumping systems, canned motor pumps are characterized by extremely long service lives. We know from experience that
the absence of shaft seals and antifriction bearings with limited service
lives and additional lubricants have enabled canned motor pumps to
achieve life cycles of several years.
Nowadays more and more attention is being paid to the sound emissions
of machines. Thanks to their liquid-cooled motor, canned motor pumps
achieve much better results than pumping systems with conventional
141
F1
80N
I
I
S1-0,1 mm
Figure 6.13
S,
$1+0,1mm
S ")
Function of hydraulic balancing of axial thrust.
electric motors. The main noise sources of the latter are the noise produced by the fans of surface-cooled drive motors and the running noise of
the antifriction bearings. By contrast, sound emission levels below 60 dB
(A) are nothing special on canned motor pumps (Fig. 6.14).
As the canned motor is installed in a flameproof enclosure and the
winding is subject to thermal monitoring, these pump sets are also approved
for frequency inverter operation with infinitely variable speed regulation.
,sw
t]
S,,~r.,,
60. OdB
r'e
s/3 .=t.
~zt.Eo2
20. OuV
,.x,
RMS
Avg.
x:
Y:
T:
A
52.8dB
~8. Ot
GO
Noise - thirdspectra
50
40
30
20
Io
31.5
63
125
250
C." $8---~ay--90 0-~."dSl~." ,.~'O
500
Ik
2k
41<
Ok
t61(
A I_
[Hz ]
Figure 6.14 Sound pressure level of a canned motor pump EXACHEM 25200/42, n = 2900 I/min.
142
Figure 6.15 Installation variant.
Monitoring (Figs 6.16-6.19)

Thermal monitoring of the motor winding is obligatory when the pumps
are operated in areas where explosion protection is required. To this end,
a control line connects the terminal box of the motor to the PTC tripping
device in the plant control room. If this monitoring device is used, overcurrent circuit breakers for the motor can be dispensed with.
Additional monitoring devices can be installed to avoid operating
errors or to protect the liquid to be pumped.
If plant conditions are such that the pump may temporarily run dry, or
if the pump may be operated against closed suction or discharge line, or
if flow may be disrupted etc., all of which might cause considerable
damage, adequate protective measures should be taken, as for all other
pumping systems.
The following protective devices are well-proven in this respect:
9
9
9
9
liquid control
flow meter on suction and discharge side
pressure control device
power control device.
The first three devices must be installed at the pump set itself and connected to the control room by means of a control line, which is often quite
143
Figure 6.16 Canned motor with inline hydraulics.

expensive. A power control device, by contrast, is installed at the power
source in the control cabinet. Power monitoring is based on various
signals such as current, torque or cos tO of the motor. This is a safe and
simple way of avoiding critical operating conditions such as dry running or
operation against closed valves for longer periods of time.
If it is essential that the liquid to be pumped does not exceed or fall
below certain temperature limits, an additional PT 100 thermistor can be
fitted to the outer wall of the can, so that the temperature of the medium
is permanently monitored.
Standardized modular construction system

The standardized connection geometry between motor and hydraulics
makes it possible to freely combine various pump and motor sizes.
This not only applies to various sizes of standardized chemical pump
hydraulics but also to pumps of in-line design. In-line pumps in particular
benefit from the compact construction of the canned motor which allows
144
azl
ot,,~
o~
145
Figure 6.18 Connection diagram for sealing liquid.

simple and inexpensive pump installation in the plant without the usual
time and cost-consuming foundation and coupling alignment.
In a similar way, the modular construction system makes it possible to
use canned motors for self-priming centrifugal pumps with side-channel
hydraulics. With just a few technical modifications, these well-proven
pumps can be converted to hermetically sealed pumps. Such pump sets are
mainly used in tank depots, for draining tanks and tank wagons etc. in the
chemical industry. They are also suitable for handling low-boiling media,
liquefied petroleum gases and liquids containing gas. These pump sets are
available either exclusively with side channel impellers or with a closed
radial impeller in the first stage, in order to improve suction behaviour.
Special versions
For special applications the pump may be supplied with an additional
connection in the area of the main front bearing for feeding sealing liquid
into the rotor chamber. The sealing liquid may either be supplied by a
separate sealing system or, in the case of solids-laden liquids, drawn off a
flow filter fitted to the discharge flange.
146
rc A.
"7
Sll
-"
K1H
K1E
_#
.
Ex
L
PE
~3
Figure 6.19 Connectiondiagram for standstill heating with PT 100 temperature

regulation.
If the sealing liquid is supplied by an external system, it is essential to

bear in mind that it will mix with the main flow handled by the pump.
If there is a risk that the pumped liquid may freeze below a certain
temperature when the pump set is switched off, a special heating system
is available to keep the pump set at the required temperature and thus
avoid freezing. In contrast with conventional canned motor pumps which
used to be heated by feeding steam or hot water to the motor, a simple
solution has been found to heat the motor of this canned motor pump
design without any technical modifications to the pump set itself. A
special electronic unit, the so-called standstill heater, is fitted in the control panel and connected to the motor power cable. During pump standstill, voltage is applied to all three phases of the motor, thus ensuring a
constant temperature in the motor. By fitting a thermistor on the can, as
described above, temperature regulation is possible.
147
This design has the following advantages. The standstill heater can be
retrofitted, without modifications to the pump set. The costs incurred at
the plant are very low, as supply lines for power or heating medium can
be dispensed with.
Bibliography
1.
2.
3.
4.
DIN EN 50014 to DIN EN 50019.

DIN IEC 34, part 5 / VDE 0530.
DIN 44082.
Heumann, A.: Sicherheit durch explosionsschutz bei spaltrohrmotorpumpen,
Presentation Dechema ( 1991).
5. Schreyer, H.: Spaltrohrmotorpumpe, Chemische Produktion 3 (1990).
CHAPTER
SEVEN
Canned motor and

magnetic drive systems:
a comparison
R . Krii m e r
Introduction
Increasing environmental awareness over the past few years has led to the
enforcement of more stringent health and safety regulations in the evaluation of chemical and petrochemical plants. Federal laws on environmental protection against harmful substances have resulted in drastic
limitation of emission levels from pumps and process plants.
Glandless, hermetic pumps which operate without leakage constitute
an important factor in the compliance with such regulations.
Essentially two different types of pump drive systems are of importance
today: canned motor and permanent magnet drive systems.
This chapter compares the two systems and points out their
characteristic differences.
Safety aspects
The primary difference between canned motor and permanent magnet
drive systems is that the canned motor is equipped with an additional safety
shell (Fig. 7.1) which is sealed to the atmosphere. The terminal boxes and
terminal connections are also gas- and liquid-tight and are designed for the
nominal design pressure of the pump. Should damage occur to the motor
can resulting from damage to the bearings or corrosion, dangerous
substances cannot escape to the atmosphere. In contrast, such damage in
the magnetic drive system constitutes a considerable safety hazard.
This risk can also be minimized to a great extent in magnetic drive
pumps by using a double-wall drive can (Fig. 7.2). However, this is possible only in conjunction with continuous monitoring of the gap between
the cans.
150
Figure 7.1 Canned-motor pump.

Canned-motor pumps are therefore recommended when the pumped
fluid has a high hazard potential. Either the lower toxic limit value of the
fluid or a classification number (hazard rating diamond according to ref.
1) can be used to assess this risk.
Figure 7.2 Magnetic drive pump featuring the double-wall system.
Canned Motor and Magnetic Drive Systems
151
Thus, canned-motor pumps should be used under all circumstances for

highly dangerous liquids (health hazard degree 4), and not magnetic drive
pumps. Such hazardous liquids include hydrocyanic acid, phosgene,
hydrogen fluoride, monochloroacetic acid, acrylonitrile, etc.
With limitations, this also applies to liquids of health hazard degree 3,
which are still classified as extremely dangerous. Examples of such liquids
are chlorine, ammonia, chlorosulfonic acid, diamide, amines, perchloric
acid, etc.
Both types of hermetic drive systems can be used equally well for fluids
classified with lower hazard classifications.
The canned motor type also proves superior to the magnet-drive type
when highly combustible liquids (fire hazard degree 4 according to ref. 1)
are to be pumped. This applies, for example, to compressed gases such as
ethylene, ethane, propylene, vinyl chloride, etc.
Also included in this category are thermal oils, for example diphyl,
which have a relatively high vapor pressure (6.6 bar abs at 360~ and a
flash point of 115~ If the can of a magnetic pump is damaged, allowing
hot diphyl to escape, a fire hazard always exists. Such accidents can only
be reliably prevented with the use of a canned motor pump.
An argument commonly used in favor of magnetic drive machines is that
such pumps must not comply with explosion protection conditions and
therefore do not require a permit from the PTB (Federal Institute for
Physics and Technology). In addition, monitoring equipment as prescribed for canned-motor pumps are not required.
Such arguments must, however, be taken with caution. Magnetic drive
pumps may also act as an ignition source for external explosive atmospheres through hot surfaces or mechanically generated sparks. Sparks
may be caused by dry running, overheating of bearings, or by grazing the
outer magnet carrier on the coupling casing.
The canned-motor pump, designed as an integral unit consisting of
pump and electromotor, offers a substantially higher degree of safety
regarding explosion protection. The liquid level and temperature monitoring devices required for this pump type prevent dry operation and
over-temperatures in both the motor and the pump. To attain an equal
degree of safety in a magnetic pump, the pump set must also be monitored for dry operation and over-temperatures.
152
Serviceability
Magnetic drive pumps clearly offer advantages over canned-motor pumps
in this respect, as only mechanical and no electric parts must be repaired
in case of breakdown.
In a canned-motor pump, damage to the motor can usually result in
damage to the windings, which must then be either replaced or rewound.
Many users with large chemical or petrochemical plants often carry out
their own stator repairs without problems. The equipment required for
winding motors or welding motor cans (TIG welding) is normally available.
In contrast, certain reservations must be made regarding the repair of
magnetic coupled pumps. If the pump is operated when dry, eddy current
losses may result in rapid increase in temperature which, in turn, leads to
overheating of the permanent magnet. If this is not recognized in time,
partial or complete demagnetization of the permanent magnets may
result. In this case, repair is no longer possible. The inner and, at the
worst, the outer magnet carrier must be completely replaced.
Temperature ranges
Standard magnetic drive pumps can be operated for liquid temperatures
up to 200~ without an additional cooling system. When a spacer is
installed between pump and drive, the operating range may be extended
to up to 400~ (without a cooling system) (Fig. 7.3).
Figure 7.3 Magneticdrive pump with thermal separation.
153
The operating temperature range for canned-motor pumps is primarily

determined by the insulation class of the motor. In pumps with a specified
insulation class, a functional relationship exists between the maximum
permissible temperature of the liquid and the output of the motor. For
example, the permissible load in a canned motor is 8 kW at a liquid
temperature of 40~ 7 kW at 70~ and only 5.8 kW at 100~ For temperatures above 100~ canned motors with external cooling systems or
motors of insulation class C must be used.
No temperature restrictions other than those pertaining to the material
are stipulated for canned motors with external cooling systems (Fig. 7.4).
The temperature in the motor cooling circuit remains almost constant
regardless of the operating temperature on the pump side. For this reason, canned-motor pumps with external cooling systems are also suitable
where extreme fluctuations in temperature may occur. One particular
example of this is combined heating-cooling circuits.
The use of magnetic coupled pumps at low temperatures (below-10~
is restricted, as ice may form on the outside of the coupling can. As a result,
the outer magnet carrier may lock, especially when starting the pump.
Magnetic drive pumps are therefore not commonly used for refrigeration
liquids. Canned-motor pumps are preferred for these applications.
Figure 7.4 Canned-motor pump with external cooling.
154
High-pressure applications
Magnetic drive pumps have two physical air gaps, one on the inside and
one on the outside of the coupling can. In addition, these pumps are not
equipped with the reinforcement rings normally used for canned-motor
pumps. For this reason, and due to the thin walls of the can, the use of
magnetic drive pumps in high pressure systems is restricted. Magnetic
drive pumps designed for higher system pressures are available, but
require extremely thick cans. As the air gap dimensions and wall thickness
increase, the efficiency of the pumps decreases rapidly, especially when
combined with higher drive powers and speeds.
This problem can be solved easily and extremely efficiently with the
canned-motor pump (Fig. 7.5). The motor can is supported from the
outside by the stator and, beyond the stator, by the reinforcement rings.
This permits thin-walled design of the motor can, which results in minimum eddy current losses. Canned-motor pumps for system pressures of
up to 950 bar have been constructed with up to 80% efficiency.
Efficiency
The magnetic drive-type pump generally achieves 5-10% higher efficiency than the corresponding canned-motor pump, resulting in lower
temperature increases in the partial flows. This can be of particular advantage when liquids with a tendency to polymerize are pumped.
Despite higher efficiency, the magnetic drive pump (including drivemotor) does not rate better in overall efficiency than the canned-motor
pump, and therefore offers no power-saving advantages.
Figure 7.5 High-pressure version of a canned-motor pump.
155
Starting behaviour
Magnetic drives have serious drawbacks in starting behavior in comparison to the canned motor.
The magnetic drive can transfer only a maximum torque. If this maximum torque is exceeded, for example when the motor is started directly,
the magnetic coupling disengages and cannot fall back into step. Magnetic
drives must be carefully designed and provided with the required torque
reserves in order to withstand the torque peaks attained during starting
and normal operation. In many cases, the maximum torque can be
increased only by enlarging the balance weight at the driving end or by
installing smooth-start devices.
Installation and space requirements

Because the pump and motor form an integral unit, canned motor pumps
generally require neither base plates nor special foundations. The units
are not at all sensitive to distortions caused by the piping.
This is of special importance if heat transfer oils are to be pumped at
high temperatures.
In magnetic drive pumps, torsion and thermal expansion of the pipe
work may lead to misalignment of the mechanical coupling and, in turn,
result in damage to the roller bearings. Because the gap between the outer
magnet carrier and the coupling can is only 0.5-1 mm, damage to the can
from the outside by the magnet carrier cannot be excluded. Under such
circumstances, the pumped liquid would escape through the ball bearings.
This constitutes a safety hazard which is not to be underestimated if
explosive or toxic liquids are pumped.
Canned-motor pumps may easily be installed without foundations, for
example using Avenarius screws (Fig. 7.6). In some cases, this may mean
substantial savings in installation costs in comparison with the costs
required for magnetic drive pumps.
Canned-motor pumps are characterized by extreme compactness and
consequentially low space requirements. This allows the pump to be
installed in otherwise inaccessible places, for example in ships or in
nuclear power stations.
Noise level
Overall noise emission of magnetic driven pumps is substantially higher
than that of canned-motor pumps. This is above all due to the electromotor (fan), the roller bearings (normally four roller bearings) and the
coupling.
156
Figure 7.6 Canned-motor pump installation without base plate and foundation.
For example, the sound pressure level of an 11 kW standard magnetic
drive pump, size 50-200, was measured at 70 dB(A), and the sound
pressure level of a canned-motor pump with the same output and size at
53 dB(A).
Thus, one magnetic coupled pump has the same sound pressure level as
50 canned-motor pumps.
Regarding noise emission, the canned-motor pump is considerably
more environmentally friendly while fulfilling the most stringent regulations stipulated by the authorities.
Cost comparison
Canned-motor pumps designed for operating temperatures of up to about
100~ hold a price advantage regarding initial costs, if it is taken into
consideration that a magnetic pump requires a base plate, a coupling and
a drive motor.
157
For canned-motor pump installations in hazardous locations, additional costs must be included for monitoring equipment.
More and more operators require temperature and dry-run monitoring
systems in magnetic drive pumps for increased pump reliability and
availability. At the same time, however, initial costs also increase.
If a double-wall magnetic drive system is required for reasons of safety,
the space between the two cans must be equipped with a continuous
leakage monitoring system, also a factor in the overall acquisition costs.
Conclusion
Figure 7.7 illustrates the scope of application of the two pump systems.
The diagram has three axes representing temperature, pressure and
hazard potential. The inner tetrahedron represents the operating range of
pumps with packing; the next larger tetrahedron represents the range of
application for pumps with mechanical seals.
Temperature
Packing
Hermetic centrifugal
pumps with permanent
Mechanicalseals
Disturbing
High
High
Hazard potential
Hermetic centrifugal pumps

with cuned motor
Figure 7.7 Operatingparameters of pumps.
[~,SSgm
158
The second largest tetrahedron represents the operating range of magnetic drive pumps, while the outer tetrahedron, representing high hazard
potential, high temperatures, high pressure, or a combination of these,
outlines the typical range in which canned- motor pumps are used.
The decision in favour of either a magnetic drive or canned-motor
pump should be based on the requirement profile. Both systems offer
advantages depending on application, pressure rating, temperature, hazard degree of the pumped fluid, explosion protection and safety
requirements.
Due to the more stringent health and safety regulations, it may be
assumed that mechanical seals are being replaced by hermetic systems at
an increasing rate; whereby
9 magnetic drive systems offer an alternative to single mechanical seals
9 canned motor drive systems offer an alternative to double mechanical
seals.
Both hermetic systems can be used to solve problem pumping applications with few and simple parts. Also both provide leak-free, hermetic
systems even under extreme temperatures, thus fulfilling all conditions
required for environmental protection. In addition, if installed and operated properly, these pumps can render technical processes maintenancefree and safe, and also contribute to increased availability and reduced
maintenance costs.
Bibliography
1. Hommel: Handbuch der gefi~hrlichen gitter. Springer Verlag, Berlin, 4th edition
(1987).
CHAPTER
EIGHT
Reciprocating metering
pumps in leak-free design
G. Vetter
Introduction
Metering pumps reproducibly feed a dosing flow, which can be adjusted
across a certain range, against pressure into a system. The characteristic
relation between the manipulated variable and the volume flow is
influenced by various parameters, such as, for example, the pressure difference, the temperature, specific density and viscosity of the fluid. A linear
and only slightly pressure dependent characteristic is the prime prerequisite for accurate dosing with pumps (Fig. 8.1). Generally all displacement
pumps (Fig. 8.2) are suitable for this task.
As internal leakages through mechanical clearances cause errors in the
transfer of a geometrically defined volume of displacement into the effective dosing flow, the tightness of the working chamber is very important in
this context. Rotary displacement pumps with inherent clearances
between the rotor and the stator only show sufficient reproducibility and
linearity of the characteristic for fluids with high viscosity.
I?
I
. . . .
. . . . . . . .
:._
Figure 8.1 The influence of the discharge pressure on the volume flow/pressure
differential characteristics: (l) reciprocating displacement pump, (2) rotary
displacement pump, (3) centrifugal pump.
160
Figure 8.2 Variousdesigns of positive displacement metering pumps: (a) geared

pump, (b) progressing cavity pump, (c) peristaltic pump, (d) plunger/diaphragm
pump.
Reciprocating dosing pumps with intermittent pressure change in the
working chamber, piston seals sliding almost without any clearance and
pump valves cyclically sealing statically show negligibly low internal leakage and accordingly attain the maximum attainable standard in the reproducibility of the dosing flow. Such pumps are therefore commonly known
as metering or controlled volume pumps. This is also due to their compact
adjustable stroke design [1].
Pumps in which the working chamber in contact with the fluid is
hermetically sealed against the environment solely by means of static seals
are called leak-free. In reciprocating metering pumps this requirement is
met with elastic walls in the form of diaphragms in various configurations,
which directly cause the displacement of the fluid.
For the handling and processing of dangerous fluids, leak-free,
hermetically sealed metering pumps are absolutely mandatory to avoid
emissions. This technology, by the way, also represents the current state
of the art [2].
As experience shows, leak-free metering pumps exhibit very favorable
properties as far as reliability of operation, low maintenance requirements,
long operational life and metering accuracy are concerned, and accordingly
Reciprocating metering pumps
161
have found acceptance in the field of less dangerous fluids also. The qualities just mentioned are particularly attractive if problems with plunger or
piston seals at high discharge pressure have to be avoided [3-7].
Since on the other hand leak-free dosing pumps require higher initial
investment, the possibilities for using special flushing, sealing or
lubricating flows in conventional plunger pumps should also be
considered [8, 9].
Review of the principal designs

The range of applications for metering pumps (Table 8.1) extends from a
few ml/h up to several m3/h and discharge pressures of several thousand
bar. A sizeable variety of designs for diverse special applications has
therefore to be explained briefly.
Mechanical actuation of the diaphragm

Direct mechanical actuation of the diaphragm is used less for reasons of
safety or prevention of emissions, but rather for simplification of the
configuration, reduced demand for maintenance and lower frictional
losses, since plunger seals are avoided. For the production of compact
metering pumps of low power rating and with a favorable price tag,
magnetic linear drives or spring-loaded cam drives are the first choice. To
reduce manufacturing costs, the drive spring should be of moderate
load only.
For the bottom of the power range (about 10 W hydraulic load), the
magnetic linear drive (or solenoid) has proved to be the most suitable
concept (Fig. 8.3). This type of drive is not only economically priced and
compact in design, but also offers favorable possibilities for the incorporation of the dosing pump into automation systems: adjustability of
stroke length and frequency, batch metering, proportioning, availability of
selecting single strokes and disturbance control.
Metering pumps with magnetic drives usually are equipped with short
stroke (stroke lengths of a few millimeters only) diaphragms made of
elastomers, lined with PTFE and stiffened by means of a disk shaped steel
core. Deflection of the diaphragm therefore takes place mainly in the
circular zone in the vicinity of the rim.
Metering pumps with a spring-loaded cam drive (Fig. 8.4) make economical use of mechanical actuation of the diaphragm at much larger
hydraulic loads (< 200 W). For partial stroke the kinematics of the lostmotion principle induces shock-wise flow, which limits the application
range to a maximum 200 W. Spring-loaded cam drives have no bottom
162
limit however. In compact configurations with built-in electric motor and

actuation of the driving pin by means of a swash plate, these drives
address the same power range as the magnetic drives.
Spring-loaded cam drives with lost motion stroke adjustment are ideal
for manual infrequent stroke setting. Electronic or pneumatic control of
stroke length or frequency are much more difficult to implement than in
magnetic drives.
Reciprocating meteringpumps
163
Figure 8.4 Diaphragm metering pump with spring loaded cam drive and lostmotion stroke length adjustment (ALLDOS): (1) suction valve, (2) diaphragm,
(3) cover, (4) discharge connection, (5) motor, (6) stroke length adjustment,
(7) reduction gear, (8) clutch.
164
Figure 8.5 Small diaphragm metering pump with spring loaded cam drive
and stroke length adjustment (ALLDOS): (1-4) diaphragm pump head,
(5 and 6) built-in motor, (7) stroke length adjustment (rotatabel cam), (8) clutch.
Elastomer diaphragms are usually used for stroke lengths up to 10 mm.

Due to corresponding support elements, the diaphragms perform a rolling motion.
If straight thrust, crank-operated drives are used, the power range of
application of mechanical actuation of the diaphragm can be extended
upwards. Series of metering pumps with mechanically actuated PTFE
sandwich diaphragms (Fig. 8.6) are presently available up to 2 kW. Their
universal use is supported by a comprehensive safety concept [6]. Due to
the large forces the actuating rods would be subjected to, further increases of the load range for mechanical actuation of the diaphragm do not
seem to be likely.
PTFE bellows represent a special configuration of the diaphragm, i.e. a
serial arrangement of several individual diaphragms (Fig. 8.7). Even
though bellows show good characteristics due to the inherent radial
stiffness, and permit the utilization of the full stroke of standard drives
for metering pumps, their use is primarily concentrated on the narrow
field of glass technology. Bellows made of PTFE are too expensive,
complicated and severely limited in application (pressure, temperature).
Future chances for developments in this direction therefore appear to be
improbable. Mechanically actuated diaphragms are loaded by the
discharge pressure. The diaphragm is not only subject to stresses by
deflections, but also to stretching deformations which generate elasticities
in the working chamber (Fig. 8.8).
165
The volume flow characteristics are therefore (see Properties section)

elastic with respect to the pressure differential and frequently not exactly
linear (as an exception, bellows have a linear characteristic). This is a
result of the dependency of the displacement, with respect to the degree
of support and the length of the stroke.
Among other factors, the metering accuracy, which is based on the
calibrable relation between the metering flow and the manipulated variable (e.g. the length of the stroke) is exposed to a considerable potential
for disturbances.
Experience shows the service life of mechanically actuated elastomer
diaphragms to be less than 5000 h. For some fluids, elevated temperature
and operation at the upper limit of the acceptable pressure range, service
life can possibly be much lower, a condition which may be fully acceptable
in individual applications.
Diaphragms lined with PTFE are good for shocking surprises if the
PTFE-lining develops pores or gets damaged. Bellows made of PTFE
usually show a reliable and fluid-independent service life. Like all other
mechanically actuated membranes, they are sensitive against overloads
however.
Figure 8.6 Mechanical actuation of a sandwich type diaphragm (LEWA):

(1-7) eccentric drive with stroke length adjustment, (8 and 9) actuating rod,
(10) oil reservoir, (11) working chamber, (12) diaphragm, (13 and 14) pump valves.
166
Figure 8.7 Bellows-typepump head in glas technology (LEWA).

Sandwich type diaphragms made of FITE (Fig. 8.6) combine universal
chemical and physical durability and long service life with rather good
characteristics.
Summarizing, the features of metering pumps with mechanically actuated diaphragms may be characterized as follows:
Lower initial cost, mechanically simpler, lower metering accuracy and
less reliability than found in pumps with hydraulically operated diaphragms.
Utilizing past experiences, the optimal solution for each specific application has to be selected.
Hydraulic actuation of the diaphragm

For diaphragm metering pumps of higher power (>0.5 kW) and for
higher discharge pressures (20-500 bar), hydraulic actuation of the
diaphragm now dominates worldwide. Such an arrangement permits the
bj
a] . . . . . . . . . . .
167
Pt~Pz
Figure 8.8 Elasticity in the working chamber of pumps for: (a) mechanical
actuation of the diaphragm, (b) mechanical actuation of bellows, (c) hydraulically
actuated diaphragm.
translation of the long plunger stroke into a comparatively small deflection of the membrane. Since the pressures on either side of the diaphragm
are almost equal, the diaphragm is in a balanced condition and does not
have to support external forces.
Hydraulic position control of the diaphragm, keeping it at a safe distance from the limits of the working space during regular operation,
(Fig. 8.9c), thus protecting the membrane against damage [7-10] represents the current state of the art.
.
,L___
r;lii
,!!I
/,f l
L_I
,m il kN
al
b/
[__ H ......
cJ
ii-
Figure 8.9 The effects of particles for various types of diaphragm control:
(a) initial position of the diaphragm after installation, (b) perforation of the
diaphragm, (c) safe distance of the diaphragm via position control.
168
Figure 8.10 Hydraulic position control of the diaphragm by means of a

gate-valve (LEWA): (1) piston, (2) diaphragm (sandwich type), (3) gate-valve,
(3a) groove, (4) snifting valve, (5) relief valve, (6) venting valve, (7) diaphragm
rupture monitoring device, (8) reservoir for hydraulic fluid, (9) fluid working
chamber, (10) hydraulicworking chamber.
Diaphragm position is controlled using mechanical gate valves in the

hydraulic system. The gate-valve 3 (Fig. 8.10) activates the snifting valve 4
for under pressure only after a groove (3a) in a pin pushed by the
diaphragm during the suction stroke provides the necessary hydraulic connection. This action takes place in the rear centre of the diaphragm only.
In the design shown in Fig. 8.11, the diaphragm pushes a sensing disk,
enabling a release of the snifting valve in the rear centre only.
In the system shown in Fig. 8.12, the load of the spring of the snifting
valve is reduced in the rear centre by the diaphragm to an extent
permitting the leakage replacement by under pressure in this position.
All proven hydraulic systems with position control are safeguarded by
pressure limiting valves. Since all experience shows the impossibility of
complete avoidance of gas bubbles in the hydraulic system, hydraulical
diaphragm metering pumps are equipped with automatic gas venting
valves (with the occasional exception of micro-dosing pumps). Volumetric
venting systems are advantageous for this purpose.
169
Figure 8.11 Hydraulic postition control of the diaphragm by means of

mechanical gating pins (BRAN & LOBBE): (1) diaphragm, (2) sensing disk,
(3 and 4) gating pins, (5) sniffing valve, (6) compression spring, (7) piston,
(8) relief valve.
Figure 8.12 Hydraulic diaphragm position control by a diaphragm shifted

equalizing valve (DOSAPRO): (1) sensing disk, (2) equalizing valve, (3) hydraulic
working chamber, (4-6) valve components, (7) discharge connection, (8) venting
valve, (9) piston, (10) reservoir for hydraulic fluid, (11) diaphragm.
170
Figure 8.13 Diaphragm under pressure controlled diaphragm pump with metal
diaphragm (LEWA).
The formation of gas bubbles in the hydraulic system can be reduced by

sufficient suction pressure and low friction piston seals.
Hydraulic diaphragm position control is presently mainly proven for
PTFE and elastomer diaphragms. This design covers however the
majority of all applications to a pressure level of 500 bar (< 150~
For the less frequently used metal diaphragms (Fig. 8.13) - pressures
up to 3000 bar, temperatures up to 200~ a need for special materials for
the diaphragm, absolute absence of pores, lowest leakage rates- hydraulic
diaphragm control is predominantly achieved with snifting at under
pressure via a plane, rear support stop for the diaphragm and a curvedshape diaphragm support in front.
This type of hydraulic system, which historically marked the start of the
development, is sensitive against dirt particles in the fluid (danger of
perforation of the diaphragm, Fig. 8.9b), and consequently demands
cleanness of the fluid to be handled (Fig. 8.9b) and large membrane
171
Figure 8.14 High pressure diaphragm head with position controlled metal
diaphragm (BRAN & LUBBE): (1) diaphragm, (2) leaf spring, (3) hole plate,
(4) sensing pin, (5) sniffingvalve, (6) venting valve, (7) relief valve, (8) control pin,
(9) compression spring, (10) plunger.
dimensions. Recently, hydraulic diaphragm position control has also been

implemented in high pressure diaphragm pumps for metal diaphragms
(Fig. 8.14). Since the deflection of the diaphragm is small, adjustment of
the control system requires high precision [10].
Metal diaphragms are predominantly used in micro-dosing pumps. The
stiffness of the working chamber and the resulting metering accuracy
constitute the most important reasons for this application.
In the recent years intensive development has taken place in microdosing, as documented in [11] and [12]. In addition to new special designs
of plunger pumps for small dosing flows with low pulsation in the field of
high performance liquid chromatography (HPLC), compact high pressure
diaphragm metering pumps with magnetic linear drives have recently
entered the market, parallel to the already well proven types with springloaded cam drives (Fig. 8.15). The vertical configuration with magnetic
linear drive (Fig. 8.16) not only permits a compact design, but also simple
means for automatic venting and protection of the hydraulic system
172
Figure 8.15 Micro-meteringdiaphragm pump with lost-motion stroke controlled

cam drive (LEWA): (1) drive unit, (2) hydraulic reservoir, (3 and 4) plunger/seal,
(5) diaphragm cover/valves, (6) diaphragm, (7) snifting valve, (8) relief valve,
(9) check valves for circulation and venting.
against excessive pressure. Applications can be found in high pressure

research projects and pilot plants, with the pressure range extending up to
700 bar. In comparison with metering pumps with magnetic linear drives
and mechanical actuation of the diaphragm, micro-metering pumps with
hydraulic actuation feature an order of magnitude increase in dosing
accuracy.
Stroke adjustment by means of the hydraulic system (Fig. 8.17) reduces
the mechanical complexity of the systems and accordingly lowers the
manufacturing expenses. Just as in all systems with lost motion control of
the stroke (e.g. Fig. 8.17a), hydraulic stroke adjustment also involves shock
type displacement mechanics during operation at partial stroke (Fig. 8.18).
Combining the tasks of lubrication of the drive system with the demands
of hydraulic actuation of the diaphragm (foaming of the liquid, wear in
the hydraulic system) is not always in the best interest of optimal metering
accuracy. The compatibility of the hydraulic fluid with the requirements
of the process served by the pump poses another problem and has to be
carefully investigated.
The following brief comments apply towards details of hydraulic stroke
adjustment.
173
Figure 8.17b, adjustable pilot sliding valve. The sliding valve 7, travelling
synchronously with the piston 1, closes the access to the hydraulic system,
depending on the position of the stroke control valve body 6. Starting at a
defined minimum stroke, the position of the control valve piston is proportional to the effective displacement stroke.
Figure 8.17c, adjustable control sleeve. The position of the control
sleeve 4 (densely hatched) on the piston determines the fraction of the
stroke of the piston during which the radial openings in the hollow piston
are covered or uncovered, and accordingly the effective stroke of the
piston.
Figure 8.17d, adjustable piston closing plug. The position of a spring
loaded piston closing plug 4 is adjustable within the hydraulic cylinder by
means of the mechanism 4, 5. The piston, initially open and ineffective,
meets the plug after a selected part of the stroke and then becomes
closed, effective and starts the working part of the stroke.
Figure 8.16 Micro-diaphragm metering pump with stroke frequency controlled

vertical solenoid drive (LEWA): (1) solenoid drive, (2 and 3) piston and corresponding seal, (4) diaphragm, (5) diaphragm cover, (6) valves of the pump.
174
Figure 8.17 Diaphragm metering pumps with hydraulic stroke control: (a) spring
loaded cam drive (LEWA), (b) adjustable pilot sliding valve (DOSAPRO),
(c) adjustable control sleeve (ALLDOS, DOSAPRO), (d) adjustable piston
closing plug (BRAN & LUBBE).
175
Diaphragm design and the safety concept

Mechanically actuated elastomer (and PTFE-lined) diaphragms are empirically and individually designed to meet the intended conditions for
installation and operation. The service life of the membrane is not only
upper dead point

. . . . . . . . . . . . . .
,'i
t,,i
_ L."!
!',,____
I hKIh~o- tO
I
=v...~_',.
I
i
I
!
I
I
I I
llrt.
I
!
I
I
! ~ ,,,r
ivZ3
i|
,!
,.
'
lllk'Km=<=yK~oo i
I/
,'
-%
'~
%
li#=lil.r
I
-,
I
x%
I
I
',
II
i\
II
I
I
I
I I
il
i,
II~ ,i,
1
I
~.;--.-IH~,!-~,
~!l,i_,
~il . I I , I !.I
~I I . .
t , ,I i lI
%;I I
il!iliY
%i ! I I I#
'4JJ I
~,i i
h./h.,o o - 1.0
fi,./Vm
[---~tu -::
~
#I
n_in,,oo=1,o
o.$
'= . . . . .
Ii i I
i,I
d%,,,.. ~,# I
~o -
%i"
I
I
I
I
I
I
I
",,',
~11
'I
l-
1.o
0.5
//o.zs / x
/I
I ,"I~
rtrll1rlH .[ v,<. .
a/
#1
t
I
,1
s"
0.5
,
i'.
'I
'
I
I
I
I
!
I
~/v.,..=t.0,
.........
IN
o5\I
11"_"~ z :.~. ~C-
f
Figure 8.18 Displacement kinematics of metering pumps with lost motion
control (a) and steady stroke adjustment by eccentric system (b). hx stroke of the
piston (hKs0 A_50%); vr piston speed; q~crank angle; dVKvelocity shock.
176
influenced by the type and degree of the periodic deformation during

operation, but also by the friction in the area of support and by the fluid
handled itself.
PTFE-bellows are likewise optimized empirically, with the influences of
the fluid and in the clamping and support area playing a lesser role
however. The efforts for development of higher displacing intensities
meet narrow limits within the present state of the art. A service life of
several thousand hours is acceptable for most applications.
Recently dosing pumps with mechanically actuated PTFE-diaphragms
have become available. Such diaphragms are almost completely resistant
to chemical attack and their service life therefore can reach around
10 000 h.
Monitoring of an enclosed check volume behind the diaphragm or

bellows usually serves as a safety concept (see also Fig. 8.7). If a dynamically operating seal forms part of the wall of the check volume, suitable
lubrication for that seal has to be provided. A level sensing switch is
usually satisfactory as a sensor.
In hydraulically actuated diaphragm type pumps, PTFE-diaphragmsplane-parallel or corrugated- are used predominantly. This design is compatible with almost all fluids encountered and covers an extensive
temperature/pressure range (150~ 500 bar).
The diaphragm deformation reaches about + 10% of the clamping
diameter (Table 8.2). Disk-shaped plane-parallel membranes can be
simply manufactured in uniform quality (also as a composite). The sheets
out of which the sandwich-layers are produced should have low thickness
(0.5-1 mm), and occasionally problems result from the presence of
micropores.
Due to larger possible thickness, corrugated diaphragms show better
performance with respect to micropores, their production in reliable quality is more expensive however. Additionally, corrugated diaphragms can
tolerate significantly lower compression at the rear working chamber
limitation during high suction pressure level.
Since hydraulic diaphragm position control is employed in the majority
of pumps and consequently the diaphragm is kept at a safe distance from
the inner surface of the front cover (Fig. 8.9), fine-grain suspensions can
be handled without any problems.
Clamping of the diaphragm is performed with 'restricted compression'
(clamping space slightly smaller than diaphragm thickness) [2], with suitable grooving of the clamping area supporting tightness and clamping
(Fig. 8.20).
The safety concept is mainly implemented by means of the sandwich
diaphragm configuration [13], with hydraulic coupling of the layers
Table 8.2
Permissible
Leformation of diaphrams based on ex ~erience

i
di~k ~ape~
d ~ ~aped
d~t shaped
cU.~ s ~ q ~
pfa~ r~'~id
pfane l~anel
mmigated
calot~ simpe
(.)
(.)
o)
Cost
low
low
high
l~mutac'mrn~ qm~ty
ve~ good
rnoclmtety ~ o d
-/i
moderate
very
Match with working volume
favorable
favourable
un~vourable
favomable
eomaty
medium
not detectable
m~demle
not detectable
Sandwich
yes
dere~
yes
Surface Sem~t~ty
low
low
low
se~tive
simpte
~mievo
r,~,~
synth, rubber
500 / 150
200 / IO0
Material
p Ixtr / *C
Thickne~ mm
0~-I
t~
low
~/nth. mbbe~
15110 / 150
200 / I00
3-5
1-2
3-5
Deflection h / d~
< 0.08
< 0.15
< 0.06 .. 0.12
< 0.2
Pm<le~formaUon
calo~ (automatic)
calot~ ( a u ~ a ~ c )
mlotW(autc~aiic)
alotte
-,
~ c)
178
(Fig. 8.21). The two diaphragms are then hydrostatically coupled by a

liquid film during the suction, while they are in physical contact during the
discharge stroke.
In case of rupture, the pressure increase in the sensor (Fig. 8.21, items
4 and 5) is used to initiate a warning signal (Fig. 8.19a). The overall pump
system remains leak-free and can be decontaminated without any danger.
Sometimes the sandwich layers are also coupled by vacuum and rupture
monitoring can be performed adequately.
The application of metal diaphragms is limited to special cases. As
compared with PTFE-diaphragms, metal diaphragms tolerate much less
deflection. Dimensioning is based on the calculable tolerable stress.
Clamping and sealing is achieved by means of positive high pressing, often
supplemented by an O-ring. The predominantly applied under pressure
control systems (Table 8.1, line 3) make metal diaphragm pumps sensitive
against particles (Figs 8.9 and 8.13). These problems can be solved in the
future by position-controlled metal diaphragm pumps (Fig. 8.14).
Monitoring for diaphragm rupture by means of a sandwich arrangement
is increasingly used for metal diaphragms also. However, the radial passages in the clamping area, necessary for the connection to the sensor, are
troublesome (stress accumulation at notches, uneven distribution of the
clamping compression). For the comparatively frequent use of these pumps
Figure 8.19 Various safety concepts: (a) sandwich diaphragm, (b) monitoring
pressure increase in the reservoir (tight capsule), (c) monitoring of level increase
in sealing capsule.
179
f
\
\
2-..
4
~-~'Ub/
Figure 8.20
for liquefied gases, rupture monitoring by the pressure in a sealed reservoir

has proved to be effective (Fig. 8.19c). In individual cases, overflow valves
have to be installed to prevent excessive pressure in the reservoir.
Special operating conditions

Pumps with hydraulically actuated diaphragms can also successfully be
used to solve problems in special applications.
The use of hydraulic interconnecting pipes in the so-called remote head
configuration* keeps fluids with a temperature exceeding the permissible
level at a safe distance from the diaphragm mechanism (Fig. 8.22). Such
an arrangement extends the initially permissible upper temperature
(< 150~ towards higher values. The high temperature in the valve head
1, gradually decreases along the interconnecting pipe 2, towards the
membrane drive, section 3.
In rare cases (e.g. handling of radioactive fluids) the overall safety
concept may make it advisable to provide remote actuation of the
diaphragm pumps. The pump section containing the driving plunger for
the hydraulic fluid then stays completely out of the dangerous area
(Fig. 8.22) while the valve head of the pump section containing the driven
diaphragm is located inside.
*Also called 'hydraulic linkage'.
180
In high pressure applications, the discharge stroke of the diaphragm

always starts from the rear centre. During replenishment of fluid leakages,
the full system pressure is then acting on the diaphragm as the pressure
difference, pushing the diaphragm against the safety support. Planeparallel PTFE diaphragms, which show a favorably low sensitivity against
notch stress, tolerate pressure levels up to 300 bar quite well, if the
support surface is smooth and has only slight grooves. As experience
shows for pressure loads of this magnitude, metal diaphragms require a
pressurized leakage replacing system to substantially reduce the pressure
difference during leakage replacement [14].
Good reasons lie behind the preferred use of hydraulic metering pumps
for microdosing: absence of any particles resulting from wear of plunger
seals, elimination of non-reproducible maintenance-dependent elasticities
of the working volume, general exclusion of stuffing box problems. The
combination of top precision valves made of modem ceramic materials
with speed controlled stepping motor drives and automatic compensation
of the influences of pressure differences on performance permits the
Figure 8.21 Sandwichdiaphragm with hydraulic coupling of the two diaphragm

components: (1 and 2) diaphragm, (3) spacer, (4) check valve, (5) pressure sensor,
(6) intermediate space (oil-film).
181
Figure 8.22 Diaphragm pumps in a remote-head configuration (LEWA):

(l) valve head, (2) pipe connection, (3) diaphragm drive section, (4) diaphragm
valve head.
microdosing of fluid flows also at low pulsation levels in the ml/h range
against pressures up to 600 bar [15].
Properties
Characteristics
Metering pumps occupy an important position in the automation of
production processes. In this role they are either directly responsible for
the quantitative dosing flow, or serve as a control component in flow or
process control loops. Generally, metering pumps are integrated into the
process via various interfaces in the fluid, energy and automation systems
(Fig. 8.23).
The different metering pump features (Fig. 8.24) are characterized by
the stroke frequency n and/or the stroke length hr as manipulated
variables to adjust the metered flow rate.
182

~
|111
F
i
'
'
!
t_
I
4
|
Figure 8.23 Metering pump within the process: F, E, Amfluid-, energy- and
automation periphery, C interfaces for communication.
An important characteristic of reciprocating metering pumps is that

fluid and operational parameters show relatively little influence on the
flow rate performance.
Accordingly, the flow rate can be described by the linear correlation
with the manipulated variables n, hK.
m =Ar'n'hr'i'Q'rlv
(1)
where (n speed, hr stroke, Ar plunger cross section, i number of cylinders)

[16-181.
If the working chamber is internally tight, the volumetric efficiency ~V
takes account primarily of influences of elasticity.
For leak-free diaphragm metering pumps with adjustable stroke, in the
case of constant middle position of the plunger stroke the efficiency can
for example be expressed as
(2)
In addition to the relative dead volumes in the fluid and hydraulic section
(err, eru), the compressibility of the hydraulic and the discharged fluid
(~r Ir the relative stroke hK/hKloO,the pressure difference z~ and the
elasticity factor ;Lof the working chamber also show some influence.
In pumps with mechanical actuation of the diaphragm, the elasticity of
the working chamber, represented primarily by the diaphragm as an
elastic wall exposed to the pressure difference Ap, constitutes the most
important factor.
183
"~1 ,,-, IVl..___J

,
,4
" i,
bl
al
F .... l--:c--r--r-!
..... "-*
c/
T"
~,
dl
r-
T'
eJ
,m
,m
,m.
"r---'l---[---I--*-"
,IW
-,
'l
Figure 8.24 Principal configurations of reciprocating metering pumps: (a) compact

(magnetic linear drive), (b) compact (mechanical eccentric drive), (c) modular
multi-component (speed/stroke control), (d) multi-cylinder single component
(speed control).
In pumps with hydraulic actuation and rigid pump design, for higher
pressure differences in most cases the elasticities of the fluids involved are
much more significant than the elasticity of the working chamber.
Typical characteristics for reciprocating piston and diaphragm type
pumps (Fig. 8.25a) show a linear course, proportional to the manipulated
variable n, with an offset from the origin, to the manipulated variable hK.
The metering error, by the way, increases as the manipulated variable
decreases, and theoretically reaches infinity at the place where the
characteristic intersects the abscissa Y (n, hK). This is particularly the case
for stroke adjustment at the position hK = hro. This condition limits the
control range R < hK]00 - hKo. The course of the characteristics can be
deduced directly from equations (1) and (2).
184
al
bl
2_
Y(n, hK) I
I
.~..
Y (n)
,,
L
v
no
Y
Figure 8.25 Meteringpumps characteristics: (a) reciprocating metering pump Y
(n, hK), (b) rotary metering pump Y (n), SD metering constancy resp. error; R
range of adjustment.
For comparison, Fig. 8.25b shows the performance of rotary metering
pumps (e.g. gear pump). The characteristics with speed n show an offset
of the linear course (no) due to internal leakage flows through clearances.
The internally tight (no clearance) reciprocating pumps show
proportionality as well as linearity with respect to the manipulated variable n however.
The observations described above point out important design parameters for optimization: rigid design of the working chamber and reduction of the dead volumes in applications with higher pressures.
Comparison of plunger and diaphragm metering

pumps
For a comparison, principal features such as the characteristics, the
metering accuracy, the suction conditions (NPSHr), reliability and
investment have to be considered.
Under comparable conditions, diaphragm pumps with mechanical
actuation usually show characteristics influenced more by elasticities due
to pressure and are not exactly linear. The dosing accuracy is therefore
moderate, but often quite sufficient (+ 2-5%). In comparison with other
types, the suction conditions pose no special requirements. Positive
185
pressure at the intake of the pump is recommended to assist a stable rolloff of the diaphragm at the supporting plate. Even in comparison with
plunger metering pumps the investments for the initial installation may be
lower. Compact design of the pumps is feasible and for light duty,
particularly for intermittent operation, the reliability is satisfactory.
For bellows type metering pumps with mechanical actuation, similar
considerations apply. The characteristics of such pumps are linear however. Bellows type pumps cover the narrow range of processes involved in
glass technology.
If the elasticity of the working chamber is optimized comparably,
diaphragm type pumps with hydraulic actuation show characteristics
similar to plunger type pumps operating under the same conditions [14].
Due to the larger hydraulic dead spaces, the volumetric efficiency of
diaphragm pumps is however always slightly lower, a feature becoming
increasingly apparent for higher difference pressures. For this reason,
diaphragm type pumps exhibit stronger pulsation [ 19] and demand special
attention for the analysis of pressure vibrations and the details of the
installation [20]. No basic differences as compared with plunger pumps
exist however.
At first glance, diaphragm metering pumps appear more prone to
disturbances increasing the metering error. In addition to the normal
fluctuations in the specific density and the compressibility of the fluid
handled, non-scheduled elasticities caused by gas bubbles in the hydraulic
system and by influences of the venting system [21] may appear. Using
such pumps in practice demonstrates however a decrease in the potential
of the hydraulic system to cause problems as compared with the potential
for disturbances resulting from piston seals in contact with the fluid.
Accordingly, the metering accuracy of diaphragm metering pumps at least
equals the accuracy of piston type pumps or usually even surpasses it. As
mentioned before, hydraulically actuated diaphragm pumps are therefore
used as metering pumps for micro-dosing.
During operation of metering pumps with hydraulically actuated
diaphragms, particular attention has to be paid to the suction conditions,
since in addition to the pressure drops in the pump valves, pressure drops
also occur in the hydraulic system. As experience shows, the minimum
absolute pressure in the hydraulic system should not drop below
approximately 0.6 bar to keep the formation of gas bubbles in the
hydraulic fluid to a minimum. It is essential to rea|ise that the hydraulic
fluid will be saturated with the gases contained in the air at normal air
pressure. The slightest drop of the pressure acting on the hydraulic system
will therefore immediately trigger the formation of gas bubbles. Usually
the requirements and conditions on the suction side of the pump are
186
therefore not exclusively determined by the NPSHr (based on the vapour

pressure of the fluid), but also by other influences. The rule to operate
hydraulic diaphragm pumps best at sufficiently high positive pressure at
the intake applies quite generally. Some designs even demand a
substantial booster pressure on suction for satisfactory operation.
If compressible fluids have to be discharged against high pressure,
substantial heat due to compression will be generated and the resulting
rise in the temperature of the fluid has to be taken into account for
definition of NPSHr [19]. If the result shows the pressure of the warmedup fluid to equalize the vapor pressure of the fluid, vigorous subcooling of
the fluid at the suction will be required.
The high reliability of hydraulic diaphragm metering pumps is today
well accepted. The prime advantage of such pumps is the absence of any
plunger seals in contact with the fluid to be handled. Even dry operation
will not cause harm in the case of interruption of the fluid flow. Modern
diaphragms show a service life well exceeding 10 000 h. As the service
requirements of such pumps are much lower than for piston pumps in
general, the ever increasing share of leak-free pumps in the market comes
as no surprise. The advantages described are however partly offset by the
higher initial investments required.
Bibliography
1. Hydraulic Institute Standards for Centrifugal, Rotary and Reciprocating Pumps
14th edition, Hydraulic Institute, Ohio (1983).
2. Vctter, G.: Stand und trends bei der entwicklung leckfreier oszillicrender
verdr~ingerpumpcn, Chem.-Ing.-Tech. 57, No. 3, 218-229 (1985).
3. Fritsch, H.: Pumpen in prozel3anlagen, 3R international 27, No. 7, 485-493
(1988).
4. Kraus, H.: Hermctisch dichte dosicrpumpen-entwicklung fiir cincn sicheren
bctricb, Pumpen - Bauelemente der Anlagentechnik 1, ed. Vctter, G., Vulkanvcrlag (1987).
5. Vetter, G.: Reliability and future development of high pressure diaphragm
pumps for process service, Proceedings of the 5th Pump User's Symposium.
Houston (1988).
6. Lewa Ecodos, Vorbildliche sicherheit beim, Niederdruck-Dosieren, LewaDruckschrift D1--001d (1990).
7. Fritsch, H.: Proze6membranpumpen - stand der technik und entwick
lungstendenzen, Chemie-Technik 16, 70-77 (1987).
8. Vetter, G., and Hering, L: Oscillating displacement pumps, Encyclopedia
Fluid Mechanics, Houston, Texas, Gulf Publishing Company (1986).
9. Vetter, G.: Abrasive suspensionen - f6rderung bei hohem druck
beherrschbar, Chemische Industrie 10, 944-950 (1986).
10. Membrandosierpumpen mit frei schwingender Metallmembran Kat. 2.8 (No
PC 46231 T1; 9104) Bran & Liibbe-Druckschrift.
187
11. Fritsch, H.: Leckfreie mikro-dosierpumpen, fachveranstaltung F 50-209-1238, Sicherheit durch leckfreie Pumpen, Haus der Technik (1988).
12. Lewa Druckschrift No. D1-410d, Die magnet-membran-dosierpumpen mit
explosionschutz (1988).
13. Vetter, G.: DBP 180018.
14. Vetter, G. and Hering L.: Leak-free pumps for chemical process industries,
Chemical Engineering 22, 149-154 (1980).
15. Lewa-Druckschrift No. D1-255d, Die intelligenten pr~izisions-membrandosierpumpen (1988).
16. Vetter, G.; Fritsch, H.; Miiller, A." Einfliisse auf die dosiergenauigkeit
oszillierender verdrangerpumpen, AuJbereitungstechnik 1, 16-27 (1974).
17. Vetter, G." Dosierpumpen seminar VDI-bildungswerk 4120--23, Diisseldorf
(1992).
18. Fritsch, H.: Genau und zuverl~issig dosieren mit dosierpumpen und
dosiersystemen, Chemie-anlagen + verfahren 6, 19-29 (1992).
19. Vetter, G.: Supercritical fluids pumping at high pressure, Proc. of the Int.
Syrup. on Supercritical Fluids, Nice, 587-594 (1988).
20. Vetter, G.: zum kenntnisstand der rechnerischen bestimmung von
druckpulsationen durch oszillierende verdr~ingerpumpen, 3R international 27,
No.7, 468-475 (1988).
21. Vetter, G." Ausf0hrungskriterien und st6reinfliisse bei oszillierenden dosierpumpen, Pumpen-bauelemente der anlagentech nik, Jahrbuch 1, ed. Vetter, G.,
521-529 Vulkanverlag (1988).
CHAPTER
NINE
Leakage-free metering of
fluids in fully automated
processes
H. Fritsch
Introduction
The entry of process automation into nearly all areas of the process
industry mandates that all modern metering devices can be integrated
into the overall process via simple interfaces and that they meter with high
reliability within a given error gap span. Because any fault which is not
recognized in time can cause high costs (spoiled product and its disposal)
the reliable and fast recognition of slowly building up metering errors
is very important. Increasing environmental awareness and tightening
of the laws passed to reduce harmful emissions have led to a strong
increase in the importance of leakage-free systems also in the metering
technology.
If limited to the metering of fluids the classical problem solvers are
either metering systems consisting of pump, flowmeter, controller and
adjusting device (Fig. 9.1b) or metering pumps which combine the main
single functions required for metering in one compact device (Fig. 9.1c).
Thus, the metering pump is also a 'metering system'. Both solutions today
cover a wide range of requirements. In particular, the metering pump,
because of its high accuracy, safety and reliability, has acquired a wide
application range for itself in the last decades.
Since there are no technical devices which can guarantee an absolute
reliability, it is tried, by skilful combination of instrument properties,
to reduce the possibility for errors (by several powers of ten) so that
an extremely high degree of reliability can be achieved. To allow
an optimum selection of such metering systems for the different
applications, it is necessary to first get a survey concerning the properties of the different metering devices and of the single components
used.
190
Accuracy and reliability of metering devices

How susceptible to faults metering systems or metering pumps are, can be
estimated only when the influence of the most important parameters on
accuracy and reliability are known. Figure 9.1 shows several principles of
fluid metering with and without measuring redundancy as used today. For
all non-redundant simple metering systems, the accuracy and reliability of
the metering process depends on one device (flowmeter, metering pump
or material value sensor). A metering error is not discovered unless a
Figure 9.1 Principlesof fluid metering (selection).
Metering of fluids in automated processes
191
check is carried out somewhere else in the process. In spite of this, such
simple devices are widely used. The reason for this: there are many metering duties having favourable operating conditions (harmless fluid, low
pressure and temperature fluctuations, narrow range of the metered flow
etc.) and so offering a high reliability of the single 'measuring' component. One trusts the flowmeter, the metering pump or some sensor
because, by experience, one knows that metering errors practically do not
occur and that the higher investment costs to achieve measuring
redundancy are not justified.
Parameters influencing the metered flow and the

metering accuracy
In metering systems, it is assumed that the pumps used must be monitored, that is, that they meter reliably when operating in a closed control
loop only (Fig. 9.1b). Such metering systems are used preferably at low
discharge pressures, medium to high flowrates and small to medium
control ranges. An important point which must be considered here, is to
assure a stable control within the application limits specified. Instabilities
of the control circuit can build up mainly at low flowrates (lower limit of
the control range) or when the pressure level fluctuates specially when
pumps having a very straight characteristic curve (Fig. 9.2) are used.
P
zxp
Figure 9.2 "I~ical characteristic curve (V.-p lines) of reciprocating positive, displacement pumps, rotating positive displacement pumps and centrifugal pumps.
192
It must also be noted that the reliability of the metering process

depends on the reliability of the single instruments flowmeter and controller. A disturbance at one of these instruments is not detected as long
as the control circuit functions. This then can lead to metering errors
which cannot be estimated in their magnitude.
Since the error possibility of a controller usually is much lower than the
error possibility of flowmeters, the evaluation of the metering reliability
can be limited to the flowmeter. When highest reliability is required, it is
possible to employ redundant pick-up of the value measured, for example
by using two measuring instruments, measuring independent of each
other and working on different measuring principles (Fig. 9.1e).
The following known equation is valid for the metered flow (volume
flow) V of a metering pump [1]:
V= Vh . i. rlv.
(1)
The volumetric efficiency Ausgabe rk, - relevant for positive displacement pumps only,- is usually determined as the product of the degree of
elasticity r/e and the quality factor r/c [1].
rlv = rle" Oo.
(2)
For most pump types which can be used as metering pumps the degree
of elasticity as well as the quality factor allow the analytic formulation [2]:
O E = I - ( A ha~176B)
(3)
rio =1-( C p n,OOn+~ )
(4)
Since typical metering pumps (= reciprocating positive displacement

pumps) have a design parameter C of almost 0 and the backflow losses VR
with properly designed valves are negligibly small, quality factors which
come near to the ideal value of r/c = 1 can be achieved. The formula for
the volumetric efficiency of reciprocating positive displacement pumps
therefore is:
qv--qE = 1 - ( A
h~~176
(5)
193
In this the parameters A and B are depending on the design of the pump,
further on the compressibility of the fluid pumped and the hydraulic fluid
(see Table 9.1). Equation (5) shows that the hydraulic efficiency of a
reciprocating positive displacement pump becomes lower the higher the
discharge pressure (p) and the shorter the stroke length (h) set is. For r/e
= 0-V also becomes 0. Each metering pump therefore has a limit stroke
length ho > 0) where it stops to deliver fluid. The metered fluid in the
pump fluid chamber in this case is only being compressed to the system
pressure and then decompressed without any actual delivery taking place,
Fig. 9.3 (line 1'-2'). From equation (5) the following is valid for the limit
stroke length ho (rle = 0))
ho= AP h~oo
l+ Bp
(6)
This means that not the total stroke adjustment range hi00 but only
the stroke length range h loo - ho (Fig. 9.3) is available for setting the
metered flow.
It should be noted, that for reciprocating positive displacement
pumps, the reduction of the flow on account of the volumetric efficiency
is mainly caused by the elasticity of the pump operating chamber and
therefore, practically, has no influence on the energetic efficiency of the
pump [3].
Rotating positive displacement pumps do not have valves VR = 0 and
elasticity influences generally are negligible. (A = 0, B - 0). This leads to
the following equation for their volumetric efficiency assuming that the
flow in the gaps is laminar [3]:*
r/,, -- 1 - C p-p-.nl~176
1/ n
(7)
Contrary to reciprocating positive displacement pumps the volumetric

efficiency of rotating positive displacement pumps has a direct influence
on the energetic efficiency, because leakages (backflow losses) between
suction and discharge always are energy losses.
The question whether a rotating positive displacement pump can
be utilized as a metering pump is mainly determined by its clearance
losses (C, 17, p) and the adjustment range (nl00#lmin), (Fig. 9.4). In
particular with viscous fluids at low discharge pressures rotary positive
*With turbulent flow, e.g. very low viscous fluids, rotating positive displacement
pumps normally can not be used as metering pumps anymore.
194
centre position of plunger
hw-----~
~he
discharge
ssure r~
lion
u;ure Pl
~he
hw
operating cycle of
metering pump
1 - 2 compressing
~/in
'7
2 - 3 conveying
D~ of vetodty
3 - 4 expanding
,~n,ne
4 - 1 suction
~ ~ -
lime
Figure 9.3 Influence of the elasticity degree on the progress of the pressure in
the pump working chamber and the metered flow.
displacement pumps can meter reasonably accurate and quite reliably.
The flow of a rotating positive displacement pump assumes the value 0
when Cpnloo/rln = 1 (or even can be negative for Cpnlodrln > 1).
For each pump an error curve showing the dependence of the volume
or mass flow on the usual fluctuations of the operating conditions
(pressure, temperature) [4] (Fig. 9.5) can be plotted.
195
Metering error in dependence to volume flow (I = metering pump,

II = other types of pumps).
Figure 9.5
Its a typical characteristic of a metering pump (Fig. 9.1c) that its metering error ~ is lower than the permissible metering error Y.perm. within the
desired adjustment range R (Fig. 9.5, error curve I). Pumps with an
extended error gap width must be used for metering purposes only when
they are integrated into a closed loop control circuit (Fig. 9.1b).
Designs and characteristics of glandless metering

pumps (diaphragm pumps)
To achieve a high metering accuracy, pumps with firm pressure characteristics are required. The pressure dependance of the pump characteristics is mainly determined by the basic design principles and the quality
of workmanship of the metering pump. For diaphragm pumps two basic
principles are distinguished: designs with mechanically (directly) actuated
diaphragms and those with hydraulically actuated diaphragms. The latter
not only possess characteristics which show hardly any pressure
196
dependance but they can handle far higher pressures (up to 3000 bar) and
power ratings (up to 1000 kW).
Figure 9.6 compares characteristics V(p) of different diaphragm pump
designs.
The output of diaphragm metering pumps is adjustable by means of two
adjusting values - the stroke length h and the stroke frequency n. The
output has a linear correlation to both adjusting values (Fig. 9.7).
With an extremely wide adjustment range at a very high metering, accuracy is possible when both adjusting values, the stroke length and the
stroke frequency, are fully used (over 1-100).
The linear correlation between the metered flow and the adjusting
|
1,0
o 0,9
.>
. > 0,8
_o 0,7
E
= 0,6
for
( ~ high pressure diaphragm pump

laboratory with hydraulically actuated
metal diaphragm.
= 0,67 I/h, Pmax= 400 bar.
o
9
._>
high pressure diaphragm pump with

hydraulically actuated
PTFE - sandwich diaphragm.
= 150 I/h, Pmax= 350 bar.
0,5
diaphragm pump with hydraulically

actuated metal diaphragm.
~/= 0,25 I/h, Pmax= 50 bar.
diaphragm pump with PTFE-sandwich
diaphragm and mechanically driven
diaphragm.
= 12 I/h, Pmax= 20 bar.
( ~ diaphragm pump with mechanically

driven elastomere diaphragm.
= 12 I/h, Pmax= 12 bar
10
20
I
30
.I
40
50
I
60
I
70
p [bar]
Figure 9.6 Characterisitics of different designs of diaphragm pumps.
80
197
[1 (A
strokefrequency n ~
~/= Vk.i.n (1-C~)
nloo ~ h o ~ - -
~ ~
strokelength n ~
V=f(p, n)
hloo
,
~--no-~
speed n ~ 1 1 ~ nloo
speed n ~
nloo
Figure 9.7 Volume flow depending of the correcting value for several pump
types.
values further allows a simple electrical or pneumatical control and their
easy integration into closed loop control circuits.
Diaphragm pumps with mechanically actuated

diaphragm
In the lower to medium output range (approximately 1 l/h to 2 m3/h) and
low pressures (up to approximately 20 bar), this principle has found a
wide market. Consequently the number of products offered is high as are
the differences in price and quality.
Further development of the diaphragm technology has led to
diaphragm pump series (Fig. 9.8) offering a number of advantages which
are not standard in this price range.
9 Diaphragm operating times of at least 5000 operating hours at maximum pressure, full flow and maximum operating temperature (80~
9 Suitable for sterilized metering of perishable fluids (CIP up to 150~
9 All diaphragms consisting of solid PTFE which means high chemical
resistance at a very low diffusion rate.
9 High pressure firmness of the diaphragm pack. That means high
metering accuracy.
198
Figure 9.8 Diaphragm pump series with mechanical (direct) actuated sandwich
diaphragm (performance range: tr = 2 to 1500 l/h, Pm~x= 20 bar).
The sandwich diaphragm of this series consists of altogether four single

layers (Fig. 9.9) whereby the so called protection diaphragm is not
exposed to a pressure load under normal operating conditions. After
indicating a diaphragm rupture it can be operated at full pressure over
several days in emergency operation, that means a replacement of the
diaphragm pack is not necessary immediately but can be carried out when
the process allows for it.
199
Figure 9.9 Diaphragmpump head with mechanical actuated diaphragm.
Diaphragm pumps with hydraulically actuated

diaphragms
A main feature of this design principle is the internal safeguarding against
over or under pressure by pressure limiting valves in the hydraulic circuit
(Fig. 9.10).
Two systems are market leaders:
9 Designs where the operating range of the diaphragm is limited by two
back-up plates and controlled by the integrated pressure limiting valves
(Fig. 9.11).
9 Designs with position controlled deflection using no front back-up
plate in the operating chamber (Fig. 9.12). This has the advantage that
even difficult fluids such as suspensions or viscous fluids can be metered
without problems.
200
pressure relief valve

(pressure limiting valve)
/
hydraulic pressure chamber
I
=.
9
~ ~
~
I~
~
hydraulic supply chamber

rating range of the
"~V/! diaphragmfor partial stroke
(independentof drive
ent principle)
~-:~ ~ r a t i n g
range of the
I I~
~
~/J plungerfor partial stroke
~
"~~~'//I
(independentof ddve
~,~
~ e l e m e n t
pnnc,ple)
constant plungercenter point
working chamber
Figure 9.10
------ h 100---refill valve

(sniffing valve)
constant front dead point
Principle of a diaphragm pump with hydraulically actuated
diaphragm.
Each diaphragm pump nowadays is using a sandwich diaphragm as a

standard (Fig. 9.13). This prevents contamination of the metered fluid
with the hydraulic fluid or any contact of product and atmosphere.
Operating principle of the sandwich diaphragm. Two diaphragms placed
on top of each other 1, 2 are coupled by a film of fluid 3 and a non-return
valve 4. The coupling fluid can be selected freely, of course considering
the operating conditions. The filling process is very easy. A surplus of
coupling fluid is filled in via the valve 5. The proper volume is adjusted by
itself because the surplus escapes through the non-return valve 4 and
venting screw 7 when the pump is started up. In the event of a diaphragm
rupture the metered or the hydraulic fluid will enter the rupture sensor via
the non-return valve (pressure sensor) 6 which triggers a signal.
Reliability of metering pumps

Because of the many parameters which can influence the metering error,
it is difficult to make a safe prediction how reliable a metering pump will
work within a given error gap width. Clean fluids, because of the very
201
Figure 9.12 Diaphragm pump head with position controlled diaphragm.

stable characteristic curves of the pumps (Fig. 9.2) which are not influenced by external ambient conditions, can not only be metered accurately
but are also extremely reliable as many examples in practical use prove.
202
Figure 9.13 Diaphragm rupture monitoring by means of sandwich diaphragm.

There are however, a number of process parameters (e.g. gas or solid
contents in the fluid pumped, build-up of layers or sediments from the
fluid pumped in the pump working chamber, excessive fluctuations of
the pressure or temperature level, high fluctuations of the viscosity,
extremely low flowrates etc.) which can influence the reliability of a
metering process.
Depending on how such disturbances influence the reliability of the
metering process in each specific case and what consequences result from
this, it is recommended or must be requested even, to monitor the metering pump by integrating it into a control circuit.
Integration of the metering pump in control

circuits
Contrary to metering systems of conventional design in which the pump
must be operated in a control circuit to be able to maintain the metering
accuracy required (Fig. 9.5fli), the monitoring of metering pumps within
203
the permissible adjustment range is not required usually (Fig. 9.5/1). If

metering pumps are integrated into control circuits, then it is for the
reasons given before.
An important criterion for evaluating whether or not the investment of
operating the metering pump in a control circuit is justified is that considering all ambient conditions, the sensor used for actual value pick-up
(flow meter, material value sensor) has a lower susceptibility to disturbances than the metering pump itself. Practical use shows, that this
requirement cannot always be met, or at high cost only.
For such cases, the combination of a metering pump having 'measuring'
properties with a flowmeter offers an elegant and economic possibility to
reliably detect deviations from the permissible metering error. This internal control system is explained in detail in the section, Metering systems
with programmable monitoring of the metered flow, below.
Simple control circuits using meteringpumps

Such control circuits are designed like conventional metering systems
(Fig. 9.16). The metering pump takes on the function of the conveying
element (pump) and the adjusting device (stroke adjustment) only. It is
not required to have measuring properties. In the case of direct control
flowmeters, indirect control material value sensors are used as measuring
instruments (Fig. 9.1d).
For flexible adaption of such control circuits to the varying metering
applications by combination of the single modules such as pump,
adjusting device, controller and flowmeter, a universally usable controller
was developed which contains the complete software pack for all kinds of
flow-controlled metering systems and combinations of instruments
(Fig. 9.14). Contrary to standard PID controllers, the adaptive 'FIS
dialog' controller allows for the programming of the pump characteristics
via a few parameters only. The actual value curve is established
automatically and continuously adapted to the current operating
conditions. By this, possible changes of the desired value are approached
directly and accurately in the shortest time (Fig. 9.15).
There are several reasons why frequently metering pumps and no other
(cheaper) pumps are used even for closed control loops:
9 There practically is no alternative to the metering pump (= reciprocating positive displacement pump) for low metered flows and high
discharge pressures.
9 The straight Q-h curve of the metering pump (Fig. 9.2) assures a high
control stability within a very wide adjustment range.
b~
t,
ir~dukl~ve ~ ~,
tlo~, meter
~;~
linear
?,
meter
-~..~.wJ
inputs
outputs
4 digital-inputs
4 digital
outputs
de.coupled
optoelectronically
5V,
decoupled
optoelectronically
.~
..~
interfaces
1 interfaces
,~.)
Figure 9.14
rso.
3045 *u/h~
2 analog-outputs
0/4-20 mA
(resolution 12 bit)
4 analog-inputs
impnnted current
0/4-20 mA
(resolution 12 bit)
RS-232-CL
Scale
[~
measur[t~ ) odr~Ete
pumps
b,
frl~9
oval w~/ee
adjusting
controllers
flowmeters
,n
weight
C~{{9
The universal controller FIS dialog.
..........
' ~r eaT F1 C~
~,,iBK
FIS dialog
4 relay-outputs
230 VAC/1A
30 VDC/1A
Q
( ~
r'~
[ ~
:~
~.,
O
:~
205
Figure 9.15 Time and control behavior of control circuits.
Figure 9.16 Modular metering system in a closed control loop consisting of:
solenoid metering pump, scale and FIS dialog controller. Adjusting value is the
stroke frequency (= pulse sequence) of the metering pump.
206
9 The adjusting device (electric or pneumatic stroke actuator) built into

the metering pump allows for a compact design of the total system.
9 No pump other than the metering pump has a wide spectrum of fluids
which can be metered.
9 Metering pumps with pump heads in diaphragm design are 'environmentally safe' (zero leakage, high energetic efficiency meaning low
energy consumption, low speed meaning low noise).
It should be noted that maintaining the metering accuracy specified in
the case of simple control circuits is (Fig. 9.1b) depending on the accuracy
and reliability of the flow meter only. When the flowmeter measures
incorrectly, metering errors will result which cannot be recognized even
when e.g. the metering pump operates accurately and without errors. In
spite of this basic deficiency, metering pumps are being increasingly
integrated into control circuits. The reason: today, the market offers a
wide range of volume and mass flow measuring instruments which make
it possible to select the fitting (sufficiently accurate and reliably functioning) make for the individual application. Figure 9.16 shows an example of
such a metering system used in chemical laboratories.
To be able to use the controller for internal monitoring of the metered
flow as well, it offers the option to programme the permissible error gap
width. If this is exceeded, an alarm is triggered. This makes it possible to
design very compact and economic metering systems with 'measuring
redundancy' as described in the following.
Metering systems with programmable monitoring of

the meteredflow
For this system, the basic idea is to use the 'measuring properties' to
produce measuring redundancy. The hook-up with the flowmeter is
possible in two ways:
9 In a closed control loop: the metered pump checks if the values measured by the flowmeter are 'plausible' (Fig. 9.17).
9 As an open system (control operation): the metered flow is determined
by the metering pump via the adjusting value 'desired value' in the first
place and only checked by the flowmeter (Fig. 9.18).
In both cases, error is signalled when the 'controlling device' (metering
pump or flow meter) moves outside of the error gap width programmed.
The combination metering p u m p - flow meter has the quality of a real
measuring redundancy (Fig. 9.1e) provided that the metering pump with
207
Figure 9.17 Metering system with programmable monitoring of the metered

flow in a closed control loop.
regard to its accuracy and reliability is comparable to standard flow
meters within the adjustment range required.
Problems of measured value pick-up and

evaluation
Measuring time cycle

The measuring time cycle plays a big part (Fig. 9.19) in the sequential
measurement of pulsating metered flows. Since the measuring time
normally depends on the pulsation frequency of the metered flow, one
must, in case of extreme pulsation of the flow as may be the case for
Figure 9.18 Metering system with programmable monitoring of metered flow in

an open loop (control operation).
208
solenoid operated metering pumps, calculate with a maximum possible

measuring error of +Vh (Fig. 9.19). To maintain a metering accuracy of +_2:,
in this case a minimum measuring volume (-mass) is required of
VMmin
(mM rain
I
Vh/~
(/'/nZ
(8)
mh/Y~ = th /nY~).
II
I"m - - ~ "
I !
~ " ~
I
i~.-.
r-!
II I
I:III
v~
">""
ii~i~ I ,
I
v.
I
I
i
I
I
!
III
arM=vutCt=mMt~
VMmln = Vh IT- '- V ' I n s
!
~ . . . . . ~I"~ mM
I .~ I I
!i
! I,,
"l
I I r~-~ i
!11
"
i=L
;t
~
A
'
I-""~'~
.
I
!
t
-- -- "> -- "I'~ '
.~.~_-._1/,
_.~ . . . .
.------
At.
mM mln ! mh/1;'_ ==i~1 / i1~.,
mkllnlum . a s u r l n o
ey.
nme Im' pulsat~nQ . o w
_,
.....
6train = R/nT.,
.met
\\
I
V/(t); m (t)
ltrne t
v
-3
,-
AN
--,
Figure 9.19 Measuring time cycles in dependence to the time related volume
respectively weight decrease for sequential metered flow measurement: (a) metering pump with harmonic kinematics, (b) metering pump with non-harmonic kinematics (solenoid operated pump), (I) linear piston meter, (II) measuring burette,
(III) scale/weighing cell.
For the measuring cycle time the following is generally valid
209
ArM = VMII2, = mMIm.
(9)
From this the minimum measuring time cycle for (extremely) pulsating
metered flows under consideration of the control range R results
/~tM min ----- R/nZ.
(10)
The measuring cycle time becomes longer the lower the pulsation
frequency of the metered flow (stroke frequency of the metering pump)
and the higher the metering accuracy requested are. We will demonstrate
this on a concrete example.
A metering pump has to be adjusted in a range of R = 10:1. The
accuracy (t:) required is 0.01. The maximum stroke frequency (n) is
180 min -l = 3 Hz
tMmin = lO/nr, = 10/3.0.01 = 330 s = 5 min 30 s.
For processes requiring a fast response of the metered flow to a given

change in the process profile, high control dynamics are required. This
also means short measuring cycle times can be strongly reduced (in the
best case by the factor ) when the measuring cycles are synchronized with
the pulsation frequency of the metered flow (Fig. 9.20).
This is carried out relatively easily for the sequential weighing (e.g.
using a contactor), by using the pump speed for synchronization. A
synchronization is not possible when the measuring marks are fixed, such
as, e.g., for the linear piston meter or the measuring burette (Fig. 9.19).
Resolution of the value measured

The resolution of the :value measured (N) sometimes is important for
sequential weight measurements, especially for wide measuring ranges
(M) and low metered flows (m). The smallest unit (M) in which the
measuring range (M) can be divided is
A M = M/N.
(11)
In order not to exceed the metering error, a minimum amount of (mmin)

must be measured
mmi n = AM/~
MlY.aV.
With equation (9) one receives the measuring cycle time
(12)
210
i \,
v.
AV - fl Vh) I
'I
i
I
I
I
I
I
I
I
f
I
I
i
|
I
I
I
I
I
I
I
I
I
I
I
n~t
,!
I
i
I
I
I
I
I
I
I
I
I!
Zo~t
....
Figure 9.20 Synchronization between pump speed and time when measurement
is taken.
AtM
rain =
mmin/ m =
M/7_,Nm.
(13)
Here, an example should also show to which consequences this can lead
in borderline cases: A mass flow of 0.1-1.0 kg h -1 (control range 10:1)
must be metered. The measuring range of the weighing cell used is 05 kg. Metering accuracy required: E = +1% = _+0.01. Measured value
resolution of the controller: 12 bit (N = 4000). In the most unfavourable
case (lowest metered flow m) it follows:
tM m i n
M/Y_,Nm = 5/0.01 94000.0.1 = 1.25 h
this means a measuring cycle time which is much too long for most
metering duties. To make it shorter either the measuring time must be
reduced or instruments having a higher resolution of the value measured
must be used.
Bibliography
1. Vetter, G., Fritsch, H. and MOiler, A.: Influences on the metering accuracy of
reciprocating positive displacement pumps, Aufoereitungs-Technik 1, 16--27
(1974).
2. Fritsch, H. and Jarosch, J.: Parameters influencing the accuracy of metering
pumps, DIT booklet 3, Synopsis no. 1464, 141-243 (1986).
3. Fritsch, H.: The volumetric and energetic efficiency of reciprocating positive
displacement pumps, Chemie-Technik 20, No. 12, 44-51 (1991).
4. Baule, B.: The Mathematics of the Natural Scientist and the Engineer, 2, 45.
Metering
of fluids in automated
processes
Symbols
AK
.-
C
h
hlo 0
ho
i
=
=
"-
Vh
V
VR
VM
6
=
=
=
=
=
=
~v
tie
716
Z
=
=
=
~TF
"-
eTH
eT
-_=
r.
._
displacer cross section

design parameters
stroke length
maximum stroke length
limit stroke length
number of pump cylinders
mass flow
stroke frequency
discharge pressure
adjustment range
stroke volume
volume flow
backflow in valves per operating cycle
measuring volume
characteristic value for the stroke adjustment system
dynamic viscosity of fluid to be sealed
output efficiency
degree of elasticity
degree of quality
maximum metering error
mean metering error
dead space in wetted chamber related to stroke volume
dead space in hydraulic chamber related to stroke volume
dead space of plunger pump related to stroke volume
compressibility of fluid conveyed
compressibility of hydraulic fluid
elasticity constant of pumphead
density of fluid pumped
211
CHAPTER
TEN
Process diaphragm
pumps
H. Fritsch
Importance of process diaphragm pumps for the

chemical process industry
For the preservation of our environment it is essential to keep the emission of harmful substances from production plants to an absolute minimum by applying modern techniques. This requirement is especially valid
for pumps in high pressure applications in the process industry. The higher their performance ranges and the more dangerous the fluid metered to
the environment, the more difficult it is to cope with the problems of leakage and its proper disposal. In recent years there have been extensive
efforts to promote the development of leakage-free process pumps towards
higher performance ranges.
The legislative body has reduced the maximum permissible emission
values (MAK-values) for a lot of the reagents pumped in chemical processes and, in the meantime, has set them so low that leak-free pumps
have become inevitable. The consequence is that the classical high pressure plunger pumps have had to be replaced by diaphragm pumps in high
pressure processes. Diaphragm pumps, however, are now increasingly
applied for less dangerous fluids. There are plausible arguments for this:
9 high availability and operational reliability compared with the plunger
pump because of its nearly wear-free hydraulic plunger sealing, long
diaphragm life, a leakage-free, reliable indication of diaphragm rupture
and an integrated total protection against over pressure and vacuum
9 low operating costs due to the low maintenance requirements and low
energy consumption. The hydraulic diaphragm pump has the highest
energy efficiency of all known pump designs as a result of the extremely
low friction and leakage of its plunger seal operating in the hydraulic oil.
These characteristics have resulted in increasing decisions in favour of the
diaphragm pump when investments are planned for plants. Its higher
214
price compared to other pump types often pays for itself after a short operating period due to the higher availability (minimum downtime of the plant)
and the low operating costs.
Also, when modernizing existing plants, older plunger pumps are often
replaced by diaphragm pumps in order to decrease costly down times as
well as costs for maintenance and repairs or to avoid intolerable leakages
- above all in the case of fluids which are difficult to pump e.g. liquefied
gases, liquids containing solids or chemically aggressive fluids.
When talking about leak-free pumps one normally refers to designs not
employing a dynamic seal (e.g. plunger seal or slide ring sealing) between
the fluid-filled (wetted) chambers and the atmosphere. Because not even
static seals are absolutely tight, the so-called leakage rate which can be
measured using highly sensitive leakage detection methods [1, 2] is used
as quality characteristic for hermetic pumps. To avoid misunderstandings:
the leakage rates of diaphragm pumps are lower by several powers of
ten (10-4-10-7) than those of comparable plunger pumps.
In recent years intensive research and development on dynamic seals
has been carried out with the goal to reduce the leakage rates to acceptable
values.
These activities however revealed that a dynamic seal with very low
leakage rates is subjected to a relatively high wear, which leads to frequent
shut-downs (replacement of the seal) and so to a lower availability of the
pump.
Designs of process diaphragm pumps

Two design concepts have established themselves on the market. The
modular design, consisting of individual elements covering the lower performance range mostly and the compact triplex design being successful
specially for high flow rates. The modular design (Fig. 10.1) originally
developed for metering pumps and still superior has the advantage that it
can be modified to handle the many different applications and customer
requirements. The higher price and space requirement is their biggest
disadvantage when compared to the compact design (Fig. 10.2). The
modular design is still preferred for some applications in process technology because its special advantages can be used. Generally however a
strong trend towards the compact triplex design can be noted.
Modular design
Today process diaphragm pumps based on the classical modular concept
are built up to performance ratings of about 30 kW (power per pump
Process diaphragm pumps
215
Figure 10.1 Modular concept for reciprocating positive displacement pumps

(metering and process pumps).
element). To keep the flow pulsation within limits they usually are built as
units with three elements (i = 3). For pumps with adjustable stroke length
this means that at least three pump drive elements must be equipped with
Figure 10.2 Comparison of size between modular and triplex design (power
approx. 70kW).
216
Figure 10.3 F High pressure process diaphragm pump of modular design with
stroke adjustment.
stroke actuators (Fig. 10.3). For economic reasons alone speed adjustment is a viable alternative. As an example Fig. 10.4 shows a typical high
pressure diaphragm pump of conventional modular design with fixed
stroke drive elements and variable speed drive.
Figure 10.4 Six cylinder process diaphragm pump (boxer design) of modular
design with speed adjustment.
217
Triplex design
Triplex diaphragm pumps are built up to performance ratings of 1000 kW
(Fig. 10.5). One of the main reasons for their fast acceptance by the
market was the technically excellent and economical speed adjustment
using inverter controlled, asynchronous motors. As stated earlier (Fig.
10.2) a triplex pump not only requires considerably less space than a
modular pump, it also has a lower noise emission and on top of that is
Figure 10.5 Triplex process diaphragm pump with speed adjustment (inverter
controlled, asynchronous motor).
218
more energy efficient. To assure optimum economics during manufacturing and when stocking spare parts, special attention was given to the
requirement to use as many identical parts as possible in the output range
in which modular and triplex pumps overlap. Figure 10.6 shows the
different designs of a modular drive element compared to a triplex drive
unit. One can see that the idea to minimize parts (as many identical parts
as possible) was maintained here.
Figure 10.6 Pump drive elements in modular and triplex design.
219
v/,,~2///H///, >-,q
Opposed (boxer) design
parallel arrangement
Proces s diaphragm pumps of modular design
|
I, L L A
-ll
!-!
o o VPH//)2H/'2Z/~
f- v~ _!_
! . . . . .
yzllZZ,,"H/)2HHZ/, r r
F ....
- ~ . . . _ _
parallel arrangement
T-arrangement
angle.__art a_ngement
Process diaphragm pump of compact . design
__
....
Figure 10.7 Comparisonof different designs of process diaphragm pumps.
Diaphragm pump designs in comparison (Fig. 10. 7)

Apart from the modular and triplex design there are other compact
diaphragm pumps whose main area of application is conveying of waste
water, sludge and slurries, but which are also used in the process industry.
The main properties of the different designs of process diaphragm pumps
are compared to each other and evaluated according to different points of
view in Table 10.1. It immediately shows that the space required by a pump
is not just determined by its size but also by the assembly space required to
install replacement parts. Figure 10.7 clearly shows that diaphragm pumps
with their pumpheads installed in parallel next to each other need
220
10.1 Evaluation of different diaphragm designs based on

application-relevant criteria
Table
.
evaluation
criteria
design acc. to figure 7

,
....
space required
installation expenses
(hook-up to
the plant)
++
++
++
++
++
0
0
++
.
......
energetic
efficiency
, .
possible
adjusting range
++
,.,
..
start-up under load

.
++
flexibility (adaption
to changing operating
conditions)
++
compatibility between
modular and compact
design
++
++
noticeably less space than the other designs. Compared to the compact
designs with speed adjustment the modular design with stroke adjustment
has the advantage that the pump can be started under full system pressure
beginning from zero stroke and so can be softly switched into the running
process. Extremely wide adjustment ranges (up to 1:100) are possible with
the modular design when, in addition to the stroke length, the stroke frequency is being used as the adjusted variable.
Technical designs
Diaphragm pump heads

Only diaphragm pumps with hydraulically actuated diaphragms are of
importance for high pressure process technology.
For safety reasons designs employing a sandwich diaphragm are used
nearly exclusively today, the principle of the mechanical-hydraulical
diaphragm coupling [3] is used by most manufacturers in the meantime.
221
It is available in two versions: up to pressures of approximately 400 bar

and 150~ with synthetic diaphragms (Fig. 10.8), for higher pressures or
higher temperatures using metal diaphragms (Fig. 10.9). The two designs
not only differ in the control of the diaphragms working area (synthetic
diaphragms: oscillating freely, position controlled; metal diaphragms: by
support plates and pressure controlled valves) but also by the design of
the sandwich diaphragms. The version using synthetic diaphragms consist
of two single diaphragms only (Fig. 10.8). On the other side an intermediate ring is required in which the channels for leakage indication are
located. The metal diaphragm consists of three single layers, the one in
the middle contains the channels leading to the diaphragm rupture sensor
(Fig. 10.9).
Figure 10.8 Standard diaphragm pump head with PTFE sandwich diaphragm
(pressure level 350 bar).
222
Figure 10.9 High pressure diaphragm pump head with metal sandwich
diaphragm.
Drives
In contrast to centrifugal pumps, reciprocating positive displacement pumps require a nearly constant, not speed related, mean torque. Figure 10.10
shows the mean torque measured, based on discharge pressure and speed,
for a triplex pump with a hydraulic power of 65 kW [4]. In the case of speed
controlled pumps it must be noted that for sizing the motor and the
223
operating range
discharge pressure p
1,0C:3
0,5-
0,2
0,4
O,e
0,8
1,0
stroke frequency n ' nipo
Figure 10.10 Mean torque of a triplex diaphragm pump depending on load (p)
and speed (n).
A
~ A
0"
90'
180"
270'
3e0,
crank shaft angle (r

Figure 10.11 Driving torque based on time (t) for high pressure triplex
diaphragm pump (efficiency r/v = 0.89) ( ~ measurement,--- calculated).
frequency inverter the peak-torque resulting from the pulsating output of
the pump [4] (Fig. 10.11) and not the mean torque must be used.
For pumps with constant speed the torque pulsation usually is characterized by a pulsation factor Cp which must be considered when selecting
the motor (Fig. 10.12). Figure 10.12 shows that the motor for a single
cylinder pump (i = 1) must be considerably bigger than for a triplex pump
(i = 3) of the same power.
Problems can arise when start-up against full system pressure is required.
To overcome the so called break-away torque ML, motor and frequency
inverter may have to be oversized considerably (Fig. 10.13).
This disadvantage can be avoided if the pump is equipped with a startup arrangement, installed either in the product side (Fig. 10.14) o r - if the
fluids metered are critical, which often is the case when diaphragm pumps
are used in the hydraulic circuit of the diaphragm pump heads
(Fig. 10.15).
224
break down
..............
II
time t
motor characteristics
related to pump speed
a(n)
selection criteria for the
motor
progress of torque M (t)for a single

cylinder pump (i - 1) and triplex
pump (i = 3) at anefficiency of "qv
approx. 1
[numberofcylindeisi [I' I 2" >3 '
MN~_ ~i=- , , I
M~=Cp-M m
pulsation factor cp.
"'
,57 1,121 1,01
Figure 10.12 Selection of asynchronous motors for reciprocating positive

displacement pumps.
to
__o
i..
without start-up~c---p permissible continuous

arrangement ~
motor torque
E
O)
l;
Mt
with start-up
arrangement
.._
start-up arrangement
closes
c)
required driving
torque of pump
Mu)
permissible adjusting range
no
speed
nloo
Figure 10.13 Torque characteristics M (n) of asynchronous motors with frequency inverter driving a triplex diaphragm pump.
225
F;---/7
I11/
!//f. I
vr-r'J
Jl
I
I
I to reactor
i
I
I
I
lp
from suction tank

Figure 10.14 Product side start-up arrangement for a big process diaphragm
pump.
Application areas of process diaphragm pumps

and area of overlap with other pump types
The process diaphragm pump has its main application range at lower to
medium flow rates and high discharge pressures (Fig. 10.16). If, for
instance, a toxic fluid must be conveyed at a flowrate of 1 m3/h against 300
bar hardly anybody would use anything except a diaphragm pump. The
higher the flow rate and the lower the discharge pressure become however, the more the diaphragm pump must compete with other pump
types. The output diagram (Fig. 10.16) shows that the area in which
leak-free pump types compete with each other is rather wide. Which
pump, in the end, is most suitable for a specific application, often can be
decided by a careful analysis of the costs and benefits. Due to the number
of different makes it is very difficult to establish binding rules for the
evaluation of the different pump types. In spite of this it is possible to give
a rough evaluation based on the different cost and performance
parameters which can assist the user to inform himself. Table 10.2 is an
attempt to do this. It should be noted, however, that only pumps offering identical standards of safety and also having the same performance
range have been compared (double walled can of centrifugal pumps
and rotating positive displacement pumps, sandwich diaphragm of
reciprocating positive displacement pumps). Fig. 10.17 and Table 10.2
show that one of the outstanding features of the diaphragm pump is its
excellent energy efficiency. Whereas leak-free centrifugal pumps and
226
Figure 10.15 Start-up arrangement in the hydraulic circuit of a diaphragm pump
head.
rotating positive displacement pumps have a noticeably lower efficiency
than the non-leak-free designs because of the eddy current losses in the
double walled can of the magnetic coupling or the canned motor, the
opposite is true for reciprocating positive displacement pumps. The
reason is that in the leak-free diaphragm pump the plunger seal operates
in oil, and so has much lower friction and clearance losses than a
comparable (non-leak-free) plunger pump which is already known for its
good efficiency [5].
In addition to this one should note that high efficiency when converting
energy (for pumps the conversion of electric into hydraulic energy)
not only has economical advantages but also has several ecological
benefits:
227
3000
.....
nq= n
),
I000
laboratory
metering pumps
multicylinder
-I~rocess
modular
metering pumps
low pressure
diaphragm
metering pumps
/o
r162
ItI
_L__x__a_
I0
10z
i,i
I03 ml/h
I
lO
I/h
lO z
I
m z
Figure 10.16 Performance range of reciprocating positive displacement pumps

and areas of overlap with other pump types.
9 Saving of primary energy to reduce the consumption of energy resources.
9 Lower heat radiation to the environment (atmosphere, cooling water).
A pump with an efficiency of 90% for instance only transmits 1/4 of the
heat to the environment than a pump with an efficiency of 60%.
9 Reduced drive power and so a lower energy consumption when manufacturing the driving machines.
These properties of the diaphragm pump are of benefit to the environment and should be an additional reason to decide in its favour in cases
of doubt.
I'J
Table 10.2
Evaluation* o f different leakage-free process pump types based on important application criteria
Evaluation criteria
centrifugal pumps
positive displacement pumps

rotating
discharge pressure
[bar]
<10
< 50
> 50
reciprocating
discharge pressure
[bar]
< 50
> 50
purchase price
++
energy consumption
flow pulsation
-IF
++
++
pressure stiffness (metering accuracy)

adjustment range
suitable for
low viscous fluids
++
++
++
high viscous fluids

fluids carrying solids
++
++
0
0
++
++
++
++
++
-H-
++
++
0
++
discharge pressure
[bar].
< 50
> 50
I
+4-
I -
*Valid as a rough guideline only. In the individual case there can be considerable deviations in ,ositive as well as in negative directions.
229
leak-free version
pumps)
I
I
ff
1F
leak-free s "
version
. (permanent magnet and
canned motor pumps) .....
0,5
"
-
I Reciprocating positve displacement pumps

(metering pumps)
a) big process pumps
b) metering pumps in the mean performance range
II Rotating positive displacement pumps
III Radial centrifugal pumps
IV High pressure centrifugal pumps (multi-stage
design and high speed pumps)
Figure 10.17 Comparison of the efficiency of different pump types.
Bibliography
1. DECHEMA information sheet ZfPI: Dichtheitspruuefung an apparaten und
komponenten von chemie-anlagen (1977).
2. Fritsch, H. and Horn, W.: Dichtungssysteme fuer oszillierende verdraengerpumpen, Handbuch Dichtungen 1, Verlag und Bildarchiv, W.H. Faragallah
(1990).
3. Vetter, G.: DBP (German Patent) 180018.
4. Jarosch, J. and Fritsch, H.: Betriebserfahrungen mit grossen triplex-membranpumpen in hochdruckprozessen, Pumpentagung Karlsruhe (1988).
5. Fritsch, H.' Der volumetrische und der energetische wirkingsgrad von verdraengerpumpen, Chemie-Technik, 20, No. 12, 44-51 (1991).
CHAPTER
ELEVEN
Diaphragm compressors
M. Dehnen
Design and operational characteristics

Nearly all chemical processes as well as process engineering involve compressing gas from a lower to a higher pressure level. Selection of the type
of compressor depends on the parameters of the process (type of gas,
capacity, pressure difference to be overcome). For higher pressure differences (above about 40 bar) only piston-type compressors can be used
(Fig. 11.1).
|
|
~6
bar
,r
,
2-
103
6:
li
im
2~Z
6.
4-
f tI
2-J
I
~ ....
J~
. . . . . i ....
6
//
!'/"
y.l . . . . . . . . . . .
~1 . . . . . ...............
....
_.J
2~"
:=
~a
1'
102
103
104
10
.
---------- Rolary compressors

. . . . . . . Turbo r
....
Screw comoressore
II
J.
I
_J
I0 s
10em31h
Rston compressors
Oiophrogm co.~'eesofe
Figure 11.1 Present applications of different compressor types [l].
232
j Oisplacement compressors
I
--!
!
I
Displacemeni
motion
I =, ,,
I o, ou,o,i~ I
....
Flow of
medium
Type of
machine
....
....
, ,
,met,, lonal J
j meridional J
,a,i
~ compressm
Piston-and
I
Ij
Screw
Root'sblower
confessor
rotary compressor
I
I
-t
I ro'o"ngi
diophrogmcompressor
Wor*~-. roiling~tonand waterring compressar
Figure 11.2 Classification of displacement compressors as per their displacement motion and flow of medium [1].
The diaphragm compressor can be looked upon as a modification of a
piston-type compressor (Fig. 11.2). Diaphragm machines are applied,
wherever
9
9
9
9
the process medium has to remain absolutely free from any lubricant
the gas has to meet the highest level of purity
no leakages- or only very small o n e s - are allowed
a very high corrosion resistance is required for all medium-contacted
components of the compressor
9 no elastomers are allowed to come into contact with the process gas.
Special design features of diaphragm compressors are:

9 their compression chambers remain absolutely free from any lubricant,
i.e. the process gas will not come into contact with any of the lubricants
used
9 in contrast to other types of compressors, no abrasion occurs from piston rings or stuffing boxes. There is no need to provide scavenging or
buffer gas devices. The process gas leaves the compressor with exactly
the same purity level, as prevails at the compressor inlet
9 they are hermetically sealed. The whole compression chamber is statically sealed to the atmosphere. Leakage rates of 10-4 mbar l/s are quite
usual and are achieved without any special action; higher sealing integrity
of 10.8 mbar l/s can be obtained with corresponding modifications in the
design
233
9 in diaphragm compressors the process gas will come into contact only
with metallic materials. For special process duties, different material
qualities are options.
An exception to this is the diaphragm plate itself, which - in view of the
high mechanical requirements - can only be produced from a spring-hard
Cr-Ni steel quality of high strength and elasticity (quality 1.3410).
Furthermore, this material withstands most chemical substances. Also for
the compressor valves, the various possible combinations of materials are
limited for functional reasons.
Function and operation

The gas is compressed in a double concave chamber by an oscillating
sandwich-diaphragm (Fig. 11.3), which is hydraulically set into motion
from one side. The diaphragm seals the gas chamber hermetically against
the drive unit. The diaphragm is clamped at its periphery between the
diaphragm cover and flange with a perforated plate is set into oscillating
motion by the hydraulic pressure. The displacement of the plates causes
the gas chamber to be enlarged then reduced with every cycle. As the gas
chamber begins to grow, the gas is sucked in via the suction valve, which
is installed in the cover. As the gas chamber is closed the gas is compressed through the discharge valve - also installed in the cover- into the
discharge tube.
The oil pressure, which is required to flex the diaphragm plates, is generated from the crankcase by the piston moving to and fro. This piston
displacement is nearly equal to the displacement inside the diaphragm
head. With the pressure stroke, the piston presses the hydraulic oil into
the diaphragm head and there through the perforated plate to the rear of
the diaphragm plate. Hereby the diaphragm plate is bent against the concave surface of the cover. In its return movement, the piston draws the
diaphragm back against the concave surface of the perforated plate.
With one rotation of the crankshaft the piston runs one complete
stroke, so the frequency of oscillation of the diaphragms corresponds to
the speed of the compressor. Medium- and large-sized machines have a
speed between 450 and 250 min-I; smaller compressors, with crankshafts
directly coupled to the electric motor, operate at approx. 720 min -1.
Design of the diaphragm head

Main components of the diaphragm head are: the diaphragm cover, the
triple sandwich-diaphragm, the perforated plate and the flange. The compressor valves (Fig. 11.4) are installed in the diaphragm cover, arranged
t~
Florae
4~
0~I beck ~le~
Cylinder
r
. . . .
et~
=a.
0{{ sur#e ~nk
Norar~ra volve /
Oil ~}e~ie~
Figure 11.3 Diaphragm compressor.
O~I ~ce
Oil cooI~g c~
Cem~nze~[o~ p~mp ~
235
thrustpiece
bolt withnut
spring
\ ~
diaphragmcover
discharge
suction
valve
valve
Figure 11.4 Compressorvalve in plate design.
one beside the other. They are sealed by metal sealing rings and are held
by thrust pieces.
The diaphragm set consists of three separate, non-profiled plates, which
are clamped-in gas-tight at the periphery between the cover and the perforated plate. Sealing to the atmosphere is achieved by metal O-rings.
Only the following components of the diaphragm head contact the gas"
~
9
~
9
the
the
the
the
diaphragm cover
gas-side plate of the Sandwich-diaphragm set
compressor valves with thrust pieces and sealing rings
metal O-ring for sealing the diaphragms.
Design of the hydraulic drive

The diaphragm head is screw-connected to the cylinder via the flange.
During the compression stroke, the piston, which moves to and fro inside
the cylinder, presses the hydraulic oil from the cylinder into the flange.
There the oil flows through the perforated plate to the reverse side of the
diaphragm. The perforated plate acts as distributor in order to achieve a
uniform pressure load by the oil on the diaphragm plate.
236
With every compression stroke of the piston, a small quantity of oil

creeps back at the piston sealings into the crankcase. In order to avoid this
reduction in oil volume causing a steady decrease in efficiency, the leakage has to be compensated for permanently.
This is effected by a compensation pump, which is directly driven by the
crankshaft, and at every suction stroke of the piston it feeds a small quantity of oil into the chamber behind the diaphragm. In every case, the
injected oil quantity has to exceed the oil loss of the leakage, which
means, that before the compression stroke, the diaphragm head is slightly
overfilled. This excess oil must be removed from the system now. For this
purpose, an oil overflow valve is provided at the highest point of the oil
chamber on top of the flange. This overflow valve feeds back the excess
oil, pumped in by the compensation pump. When the compressor is operating with a constant or nearly constant discharge pressure, the opening
pressure of this overflow valve is firmly adjusted by a pretightened spring
(Fig. 11.5). In this case, the oil pressure must always be approx. 10 %
higher than the maximum allowable gas discharge pressure.
If the duty of the compressor will involve a variable discharge pressure,
e.g. when filling pressure vessels, an overflow valve is used, which is controlled by the gas pressure (Fig. 11.6). As a function of the varying discharge pressure, the valve spring is pre-tightened by a stem and thus the
oil pressure is limited to approx. 10% above the momentary gas pressure.
This ensures, that the oil pressure is steadily adapted to the slowly increasing gas pressure. This measure contributes to increasing the life of all
components, since it avoids overloading the compressor by operating it
over a longer period with an unnecessarily high oil pressure.
With the forward movement of the piston- even before reaching the
front dead point - the diaphragm is pressed against the diaphragm cover;
the oil quantity which is compressed with the remaining stroke, flows back
through the oil overflow valve into the oil tank. Simultaneously, all air will
be entrained, which has gathered at the highest point from degassing of
the hydraulic oil. The excess oil quantity flows into the surge oil tank,
where it degasses and from where also the compensation pump sucks in
oil. This particular arrangement ensures that the hydraulic system is
always filled up to the optimum oil quantity and that the oil chamber is
totally air-bled.
In larger diaphragm heads, the hydraulic oil needs to be cooled. By
using a cooling coiled tube filled with flowing cooling water, part of the
compression and friction heat can be dissipated. Special types of mineral
oils are used as the hydraulic medium; at the same time, these oils serve
to lubricate the crankcase.
These oils have to meet the following requirements:
237
process gas
connection
t
I
piston
thrust piece
corle
oil outlet
l--q
!
oil outlet
seat
lifting lever
oil inlet
|
Figure
spring
11.5
Oil overflow valves -
loaded.
oil inlet
Figure 11.6 Oil overflow valves - gas

pressure controlled.
9 high lubricating properties

9 low compressibility
9 reduced tendency to foaming.
For special applications and duties, different hydraulic media have to be
used, such as special synthetic oil for oxygen compressors, or water with
rust inhibitor (the latter requires a separate loop of crankcase lubrication).
238
Multi-stage designs
Should a single-stage compression not be sufficient, two diaphragm heads
are connected in series. In this case, the piston is designed as a stagepiston (Fig. 11.7). The resulting differential surface of the piston generates the hydraulic pressure for the second diaphragm head, which means
the compression stroke of the first stage acts as the suction stroke for the
second stage. For three-stage or four-stage compressors, a duplex crankcase is provided, with the first and second stages on one side, and the third
and fourth stages on the other side.
Thermodynamics and design

The precise design of the compressor guarantees a high efficiency, long
service life of the individual components and, consequently, a high reliability and availability of the machines.
In order to meet the intended duty, the compressor has to be designed
exactly in accordance with the technical parameters pre-determined by
the end-user. Only this will ensure, that the compressor complies with the
technical requirements and also meets economic expectations.
The following parameters must be known for designing a diaphragm
compressor:
9
9
9
9
9
type of gas (also specific gas data)

requested suction volume
suction pressure
discharge pressure
gas temperature at suction inlet.
Based on these gas data, the effective suction volume (related to suction
pressure and temperature) is calculated by means of the equation of state.
The total pressure ratio of discharge pressure versus suction pressure
determines the necessary number of stages and consequently the interstage
pressure ratio of the individual compression stages of the compressor.
Depending on the type and quantity of gas, diaphragm compressors
allow stage pressure ratios of up to 1:20 (compared to piston compressors,
the max. ratio is 1:6). The stage pressure ratio is limited by the allowable
gas temperature at the end of the compression (max. however 250~ and
by the clearance. Since, at the end of the compression stroke, the
diaphragm is pressed against the cover, no dead space will result here and
only the clearance in the compressor valves and in the valve pockets has
to be taken into account in the design.
The gas compression is very close to the ideal isothermic curve because
of the good thermodynamic conditions in the compressor. These result
o
t~
o
o
u
qr-
.$.
r-
"13
0
W
C~
0
n
o
239
Id
1"4
d=
240
from the excellent emission of the compression heat to the atmosphere

via the large diaphragm cover and by the gas cooling, which is already
effected during the compression by the cooled hydraulic oil (very good
heat flow via the thin diaphragm plate).
Figure 11.8 shows the pV-diagram of a compressor.
p~
vh
1-2
2-3
3-4
4-1
Vs
Vh
: compression curve
: d i s c h a r g e curve
: back expansion curve
9 suction curve
:clearance
: p i s t o n displacement
Pi : suction pressure
Pz : (stage-) discharge pressure
P~ resp. Pz" : p r e s s u r e decrease r e s p . - i n c r e a s e
caused by throttling and friction
losses (impact pressures)
Figure 11.8 p, V-diagram.
Determination of the main dimensions

The necessary piston displacement of the first stage is dependent on:
9 the pre-determined, requested suction volume
9 the pre-determined suction pressure
9 the pre-determined suction temperature
241
9 the deviation from the perfect gas law, applicable to the type of mixture
of gas to be compressed
9 the isothermic efficiency.
Should a further compression stage be necessary for the total gas pressure
ratio, interstage cooling is provided behind the first stage. If the gas
contains condensing particles, the condensate has to be drained by
interstage separators.
The piston displacement of the second stage is determinated in the
same way, however, under consideration of the discharge pressure of the
first stage and the calculated gas temperature prevailing at the outlet of
the interstage cooler.
Efficiency
Volumetric efficiency (feeding). The volumetric efficiency is defined as
effective suction volume/piston displacement. It is dependent on'
9 the back-expansion of the gas, compressed in the clearances
9 the sub-expansion of the gas during suction
9 increase in discharge pressure, and decrease in suction pressure, caused
by impact pressures in the discharge and suction inlets
9 heating of the gas, when entering the warm diaphragm head.
Indicated isothetrnic efficiency. The isothermic efficiency takes into

account flow and heat losses; the latter having greater influence on the
isothermic efficiency. It is defined as:
/~isi -" Pisothermic /eindicated 9
Mechanical and hydraulic efficiency. All mechanical and hydraulic losses

are taken into account by these efficiencies. The hydraulic efficiency
specifically applies to diaphragm compressors and considers the losses of
oscillating oil quantities as well as the leakages which are lost via the
piston rings.
Actual isothermic efficiency. The actual isothermic efficiency comprises
all losses of the compressor and thus indicates the total losses as:
Oise =
/lisi * //mechanical * /'/hydraulic.
The quality of the compressor is determined by this efficiency.
242
Monitoring of the diaphragms

The metal diaphragms, depending on the operating pressure and when
working with clean gas, uncontaminated by foreign particles, will achieve
a service life of approx. 3000-15 000 h.
The system of using triple sandwich diaphragms with diaphragm failure
indicators prevents penetration of the gas into the hydraulic part, or oil
into the gas part respectively.
The sandwich diaphragm consists of three superposed diaphragm plates,
of which the middle plate is flitted. In case one of the diaphragms, either
oil side or gas side, fails, the corresponding medium penetrates between
the gas and oil side diaphragm plates.
A pressure switch connected to this chamber (Fig. 11.9) and/or a contact pressure gauge (Fig. 11.10) will stop the compressor. This ensures,
that gas and oil will never come into contact.
Slilfed centerdiaphragm
Gas-sidediaphragm~
Gas space~
~
/ O i l - s i d e diaphragm
Perforatedplate
~ .
/
/ ;
Oil space
Flange
i'
i
Diaphragmcover- /
Collectionchamber/
P..u~. , w , ~ h ~ _ _
(alarm or shut
/
i0 [
i ol
~"~" SliHedcenter
diaphragm
Figure 11.9 Diaphragmfailure device (with pressure switch).
243
Contoct pressure gouge

(alarm or shutdown)
t. f'
/.~'"
Collection chamber \
['Ti
"",,,
/
c""
, ~ ..... ~
Sllfled center
dlophrogm
Flonge
Perforolsd plole
'~176
0,.,,,o,o.
Oil-slds dlophrogm
Figure 11.10 Diaphragmfailure device (with pressure gauge).
Types of construction
Diaphragm compressors are preferably of horizontal design. Vertical
diaphragm compressors require less space for installation, however, they
have the disadvantage of insufficient air-bleeding, as the overflow valve
cannot be mounted at the highest point of the hydraulic system. Figures
11.11-11.17 show the different types of compressor design.
Application and operation

In all sectors of gas compression, diaphragm compressors have gained a
strong following in view of their operational advantages. Wherever
244
Figure 11.11 One-stage compressor, vertical design.

flammable, explosive, radioactive or toxic gases, or gases of the highest
purity have to be compressed, where particularly stringent measures have
to be taken into account for environmental protection, they offer
considerable advantages compared to the classic piston-type compressor.
Diaphragm compressors meet the rigorous requirements of the food,
pharmaceutical and nuclear industries.
For different applications, the following variety of conceptions are
available:
9 basic compressors, for compressing gas, single or multistage, from a
relatively low suction pressure to a higher discharge pressure
9 booster compressors, for compressing from a high suction pressure to
a still higher discharge pressure
9 transfer compressors, for pumping gas from one tank into another. A
special feature of this type is the decreasing suction pressure at
simultaneously increasing discharge pressure
9 gas circulators, these are under a higher static pressure and feed gas
within a system in a loop circuit. In these cases, the differential pressure
between suction and discharge pressure is normally not very high.
Drive mechanism
Larger-sized diaphragm compressors are nearly always motor-driven by
an asynchronous motor via a V-belt transmission. The drive pulley on the
crankshaft of the compressor is designed as a fly-wheel.
Smaller-sized compressor can also be directly coupled to the motor
shaft.
--~ _ ~
245
__., \
=
Figure 11.12 One-stage compressor, horizontal design.
Figure 11.13 Two-stagecompressor, horizontal design.
i-
l~!
,!
Figure 11.14 Two-stagecompressor, L-design.
246

l
"1
'fB~
"
Figure 11.15 Three-stage compressor, L-design.
Capacity regulation
According to the specified theoretical data, the compressor capacity can
be regulated either manually or automatically via a speed-variable drive
or via a bypass system. Single-stage diaphragm compressors can also be
regulated in capacity by adding the clearance volume. The classic type of
regulation, i.e. via valve lifts, is not applicable for diaphragm compressors.
Figure 11.16 Three-stage compressor, horizontal design.
247
......
,--
........... ::'::--1 ..... " "

Figure 11.17 One-stage compressor, duplex, horizontal design.
Operation
Starting and stopping of a diaphragm compressor system should preferably be done by an automatic control unit. The actuation of the valves and
fittings pertinent to the compressor, such as inlet, outlet, relief and bypass
valves has to be effected in a timely adjusted sequence. In view of safety
aspects of the system, this is of particular importance for compressors
used in production systems (Fig. 11.18).
Operational safety
Diaphragm compressors can be incorporated as reliable equipment within
the safety philosophy of complete units. With the monitoring of the
diaphragm regarding possible failures (if necessary, with a second, independent switch) and the cooling water flow control ensuring an automatic
stop of the compressor in case of diaphragm failure or cooling water shortage, sufficient safety for most of the systems is achieved.
For additional requirements, the compressor may be equipped with:
9
9
9
9
limit
limit
limit
level
switch
switch
switch
switch
and indicator for suction pressure and temperature

and indicator for interstage pressure and temperature
and indicator for discharge pressure and temperature
for hydraulic oil.
A compressor unit can be controlled either in the conventional way by

a relay control, or by a memory programmable control (SPS). This
control unit can be easily incorporated into modern process-control
systems.
248
u
"0
9~ -
W
L..
GI
~x
u
;3
to
.g
:3
0
(~
0
01
ot
o
v
ot
=6,_
w~
(u
u
:3
.~_ ,e..>
N
a;
>_,
o
Ih
0
OI
,~..~
249
Summary
Diaphragm compressors have been in use for decades in industrial applications and there they meet the specified demands for operational safety,
availability and economic efficiency.
In the design of production plants, they help to comply with environmental requirements, they simplify production processes, where the purity of the compressed gas is an important factor and their very smooth
running is a special merit.
Their suction capacities cover a range of some l/h and several 100 m3/h.
Discharge pressures of up to 3000 bar can be achieved with diaphragm
compressors.
Bibliography
1.
2.
3.
4.
Kiittner, Kolbenverdichter,Springer-Verlag (1991).

Baehr, Thermodynamik, Springer-Verlag (1984).
Fr/Shlich,Kolbenmachinen, Springer-Verlag (1968).
Werksinterna der Firma Andreas Hofer Hochdrucktechnik GmbH.
CHAPTER
TWELVE
Liquid ring vacuum

pumps and compressors
with magnetic drive
W. Lehmann
Introduction
Liquid ring pumps belong to the group of positive displacement pumps.
The characteristic feature of this pump type is the energy transmission
from the impeller to the fluid pumped by means of a liquid ring. This
operating principle and the simple construction of the pump allow for its
broad application, both as a vacuum pump and as a compressor for handling and compressing almost any gases or vapours. Small amounts of
liquid may be delivered with the gas or vapour flow. Depending on the
requested leak tightness requirements, the liquid ring pumps are equipped with a canned motor or a magnetic drive (Fig. 12.1).
Design and principle of operation

The basic design of a liquid ring pump is shown in Fig. 12.2. The rotating
impeller, which has straight or curved blades, rotates eccentrically in the
pump casing, and transmits the energy to a liquid ring rotating concentrically in the casing. This arrangement causes a crescent-shaped space to
form in the area bordering on the hub, with individual cells between the
blades which are sealed by the liquid ring on the outside, and on the
lateral cell walls (port plates) by narrow clearances sealed by a liquid film.
The suction and discharge ports are located in the lateral port plates, or,
on other systems, in the area of the impeller hub.
Due to the relative eccentricity of the liquid ring and the impeller, the
liquid flows out of the blade cells and back in again like a piston with each
revolution of the impeller. Through the piston action, the working volume
is alternately increased in the area of the suction port, causing the gas or
vapour to be drawn in, and decreased in the area of the discharge port,
252
Figure 12.1 Liquidring vacuum pump with a magnetic drive (SIHI).
thus compressing the gas or vapour, which is then either discharged

direct, or passed on to the next pump stage on a multistage pump design.
The service liquid plays a major part in this process. Apart from transmitting energy and sealing off the impeller blade cells, it serves, among
others, to absorb and dissipate most of the heat generated during the
compression process, by condensation of the vapour-bearing gases during
the suction action, or by the possible (and, in a number of cases, desired)
chemical reactions taking place between the gas and the service liquid in
the pump.
Since part of the service liquid is discharged continuously from the
discharge port along with the flow of gas to be pumped, the heat produced
is eliminated continuously. For this reason, the pump has to be fed with
'fresh' service liquid, e.g. from the water main, the water supply system
and/or a water separator. This way, the volume and the temperature of
-t>
"1
I"11"
Liquid ring vacuum and compressors with magnetic drive
1!!
253
O
e,t
e,i
254
the service liquid contained by the pump remain constant, thus allowing
for practically isothermal compression.
Depending on the type of operation, one distinguishes between freshliquid operation, combined-flow operation and bypass-flow operation.
Although closed pumps may, in principle, be operated in any circuit,
closed circuits are, in this context, of greater significance.
Both end-users and manufacturers have at their disposal decades of
experience with this pump type.
Pump requirements/safety regulations

The needs for extracting and compressing gases and vapours in practical
applications are manifold. Ongoing developments in the field of process
engineering against the background of the need to cut the cost of energy
and a growing awareness of ecological problems have led to the technological advancement of proven compressor and vacuum pump designs.
This has led to the development of machines having an extremely high
degree of leak tightness and, therefore, capable of meeting the stringent
requirements of the Emission Protection Act, the Water Resources Act,
and the Accident Ordinance regulations.
Where vacuum pumps and compressors have to cope with unusual
operating or ambient conditions, they have to meet the safety requirements dictated by their intended use [1]. Particularly in the fields of chemical, process and apparatus and plant engineering, there is a frequent
need to compress gases and gas/vapour mixtures characterized by at least
one of the following properties:
9
9
9
9
9
9
9
chemically unstable
explosive
toxic
carcinogenic
malodorous
corrosive
erosive.
In addition, the gases and vapour mixtures handled and extracted in the
field of nuclear power engineering are radioactive.
As a first step, national and international rules and regulations, such as
UW Verdichter (compressor association), ISO 5388 - safety rules, and
DIN 31000/1 - general guidelines for the design of technical products
from a safety viewpoint, lay down the standards for the safe design of
technical products as well as emission limits. More detailed information
on the subject of applications and products, however, is to be found in
255
publications on test results [2-4]. These reports should be evaluated on

the basis of the following distinguishing features:
Chemically unstable, explosive. Avoidance of sources of ignition; limitation of the effects of an explosion on the casings; use of materials which
do not react with the product pumped and prevent corrosion of pump
components; leak tightness of casing partition lines and shaft seals, in
particular on vacuum pumps, also in reverse direction (from the outside
to the inside).
Toxic, carcinogenic, malodorous. Minimization of the leakage flow from
the casings and shaft seals to an acceptable level; the product pumped and
the service liquid have to have a neutral effect on the materials of
construction.
Corrosive (chemical erosion). The use of corrosion-resistant materials;
allowance for material erosion through corrosion attack when designing
the wall thicknesses of pressure-retaining components. A point to be
remembered is that externally, the parts are subject to corrosion by the
atmosphere.
Erosive (mechanical wear). If the product pumped contains solids, the
material of construction selected has to be wear-resistant; allowance is to
be made for material wear by erosion when designing the wall thicknesses
of pressure-retaining parts.
Radioactive. Minimum leakage rates, also in the event of unusually high
pressures and temperatures and incidents; the use of materials which are
resistant to corrosion and show no changes in their physical properties
when affected by radiation; testability of components, avoidance of welds;
long inspection and repair intervals; provision for repairs in the hot zone;
chemically and physically pure surfaces.
Safety of operation. Ultimately, this requirement can only be met by a
leak-proof pump which by design has an extended availability. By implication, emphasis is therefore on longer intervals between repairs.
Operating behaviour and influencing factors

Leaving the condensing behaviour of liquid ring vacuum pumps out of
consideration for a start, this pump type is to be regarded as a displacement compressor. In real terms, its suction capacity cannot be defined
256
until all influencing factors, such as volumetric efficiency, vaporisation,

condensation, temperature, density, viscosity, gas solubility and entrainment of liquid in the gas flow, etc. have been considered. These factors
cannot be discussed in more detail in the scope of this chapter. For a
standard pump with mechanical seals, they result in the familiar characteristic curves (Fig. 12.3).
300
......
250
Af
~ ~ - -
'
1750 llmin
200
' ,I~7
//
150
.~f
--~.~
1~so1/nln
L, /
._~ 7
IU
Q.
~ 50
C
~0
,i~
,
,
10
90-
6
4
....~
l/)
.a
1450 llmin
~
o
a.
LEMILEL 250-
0
L,O
GO 80
120
200
t,O0 oCKzr 900
suction pressure ~
Suction capacity and power absorption

The data are applicable for:
Compression of dry air, temperature 20~
thin black curves
Compression of water vapour saturated air, temperature 20~
thick black curves
Water as service liquid, temperature 15~

Compression pressure abt. 1013 mbar
Figure 12.3
Characteristic curves of a s t a n d a r d v a c u u m p u m p (SIHI).
Liquid nng vacuum and compressors with magnetic drive
257
Liquid ring pumps with a canned motor

It cannot be said that the number of tests conducted for the purpose of
establishing the possibilities of applying vacuum pumps or compressors
with a canned motor (Fig. 12.4) is lacking. On the whole, the effort has
only met with success in the case of nuclear power stations.
Figure 12.4 Liquid ring pump with a canned motor: (1) gas pump, (2) liquid
pump, (3) suction and discharge connection, (4) steel casing, (5) canned motor,
(6) stator tube, (7) rotor (SIHI).
258
The reasons for their successful application are found, in particular, in

the leak tightness requirements with leakage rates of 1 x 10--6 mbar l/s
(helium leak tests), and the materials quality assurance measures in line
with current rules and regulations, e.g. AD-Merkbl~itter (AD instruction
sheets - pressure vessels), intercrystalline corrosion test, radiographic
examination and crack test connected with it. Acceptance and functional
testing, and establishment of the corresponding certificates, were finally
conducted under the supervision of the responsible T 0 V inspection
agency in the presence of the customer.
The design of a pump with a canned motor is shown in Fig. 12.4. The solid
connecting body incorporating the suction and discharge connections (3)
carries the canned motor on the left, and the hydraulic stages (1) in a steel
casing on the right. A separate side channel stage (2) supplies the motor
and the bearings with the required flows of lubricant and coolant. The
casing components are pressure-tight. At the motor end, the windings are
fitted with thermal detecting elements in order to monitor the pump set
as a precaution.
The applications vary according to the specific power station design, but
from the point of view of the environments of use, there is an equal distribution of vacuum pumps (Fig. 12.5) and compressors (Fig. 12.6). Taking
vacuum pumps as an example, the products handled mainly include nitrogen, oxygen, D2-vapour, deuterium, water vapour, cracker gases, helium,
hydrogen and noble gases. Special attention has to be paid to the pump
bearings.
The effectiveness of liquid ring vacuum pumps very much depends on
an efficient sealing action between impeller and casing. Very narrow axial
clearance gaps are, therefore, essential. This problem is solved by arranging
h,'/h
120
m
100
_>,
"~
8o
m
....
o.
60
~f
.o
.J
~mmmm=Mm
AO
20
78
2
910
Suction pressure
7 E g 103mbo~
Figure 12.5 Characteristic curve of a vacuum pump with a canned motor.
259
Liquid ling vacuum and compressors with magnetic drive

m)/h
210 .-
I ! 1I
210 9
200.
abs
"
190
. i
180
9 170
E
160
11,0
Compression pressure
bar
obs.
Figure 12.6 Characteristic curve of a compressor with a canned motor.
separate single bearings on the shaft and pump stages (materials: carbon,
carbides and nitrides). In order to minimize the loads acting on the bearings, the stages are either of double-acting design or designed as opposing
stages. Within this context, it should be mentioned that a closed circuit
(Fig. 12.7) should also be provided for in these circuits.
The original aim of introducing canned motor pumps to chemical plant
and the chemical process industry has only met with limited success.
Liquid ring pump with a magnetic drive

On this type of pump, the stationary end of the pump on the drive side
ends in a magnetic drive. Fig. 12.8 gives a closer look at the individual
coupling elements. To get a better impression of the real design, Fig. 12.9
shows a pump with the can removed. On the left, it shows the pump itself,
reliably sealed by the can, shown here as a sectional view.
The inner magnets mounted on the pump shaft rotate without touching
the can.
The motor, in this case mounted on the right, actuates the outer magnet
carrier mounted on the motor shaft. The pump and motor shafts operate
synchronously with the motor speed, only offset by a load angle. Due to
the modular design, the coupling may be selected in accordance with the
required load torque.
260
t~
i
l"
L'Q:~
loyal
r ~
~Ot
I
pump
Figure 12.7 Closed circuit (SIHI).
Figure 12.8 Sectional view of a magnetic drive.
261
Figure 12.9 Liquid ring vacuum pump with a magnetic drive (SIHI).
As the hydraulic parts are always more or less the same, we will not
discuss the characteristic curves in more detail (Fig. 12.3). However, it is
important to note that all drive motors are offered with temperature
monitoring.
Also in this case, the pump rotor has to be supported completely in
sleeve bearings. This point is significant, because the necessary working
clearances between the rotor(s) and the port plates have to be set at very
close range. Bearing arrangements of SiC (Fig. 12.10) have been applied
with great success; these can be inspected from the outside with a minimum of effort.
Depending on the pump type and the drive rating, the bearing arrangement to be selected varies. Fig. 12.11 shows a two-stage design, which is
mainly intended for use in the fields of chemical and process engineering.
Single-stage variants (Fig. 12.12) are of simpler design and, as a rule, have
a flanged-on motor.
With increasing drive ratings and larger-sized designs, the free shaft
end (Fig. 12.13) also gains in importance. This is increased further still, if
the downstream control chamber also has to be covered by the monitoring
system.
The power range of vacuum pumps with a magnetic drive is illustrated
in Fig. 12.14. Apart from so-called dead-end operation, there are magnetic
drive variants in operation whose inner magnet is not just surrounded by
262
,fl
Figure 12.10 SiC-bearingarrangement on the non-drive end.
the liquid, but cooled as well, similar to the bearings facing in the direction of the flow. On this design variant, the flow of service liquid is positively influenced by parallel bores.
As the majority of cans are made from metal (non-magnetic) materials,
the eddy current losses in the can have to be dissipated. The magnitude of
these losses is determined by the rotor speed and the mass of the magnets
263
Figure 12.11 Two-stagevacuum pump with a magnetic drive (SIHI).
installed. The higher the drive speed to be selected, the higher these
losses will be. As the example in Fig. 12.15 shows, the losses are in the
region of 11% at a speed of n = 2900 1/min. In order to be able to accelerate the quite considerable flywheel masses when a pump is started, e.g.
primed completely, certain power reserves in the magnetic drive have to
be provided for. Primarily, their purpose is to facilitate start-up, but in the
case of a breakdown, they serve as a rupture joint, comparable with a
hydrodynamic coupling 'breaking off'.
Any comparison between canned motors and magnetic drives is incomplete without evaluating the pros and cons of the two systems. Although
all comparisons are lame by definition, an effort is made as shown in the
following table (Fig. 12.16).
Applications
Liquid ring pumps have been used successfully for compressing gases and
gas-vapour mixtures for a great number of years.
The majority of pumps used in nuclear applications are equipped with
a canned motor with a leakage rate of less than 1 x 10-6 mbar l/s. In other
fields, the magnetic drive is the leading driving method. Their leakage
rates are in the region of approximately 1 x 10-~ to 1 x 10-3 mbar l/s.
to
G~
g~
Figure 12.12
Vacuum pump with a magnetic drive and a flange-mounting motor (SIHI).
3.
c~
e%
t,,J
Figure 12.13
Vacuum pump with a magnetic drive and a free shaft end (SIHI).
266
Figure 12.14 Power requirement of pumps equipped with a magnetic drive

(SIHI).
....
MAGNETIC DRIVE
CANNED MOTOR
Eleclric power hpul

100%
Electric power ~put

100%
. . . .
motor tosses
-~ ........ 10%
I
I
I
I m~
,~
e d o ~ - o . ~ t losses
7%
11%
......
I
I
!
motor losses
14%
friction losses
/.%
friclk:~ losses
5%
75%
7/.%
mechaa~ o u ~
mechor~a o u ~
Figure 12.15 Power comparison: pumps with a canned motor vs. magnetic drive.

Canned motor
Magnetic ddve
Component
267
Properties
IEC .......
Motor
Special design .....

synchronous
Speed
cold:
~=qc
Efficiency
hot:
nM ,.z,nc
Total operating cost
lower
Nominal pressure . . . . . .
lower
higher
!
more favourable
favoun
S~ety
.....
less
more
not required, except for

motor
}
I
required
lower
higher
Repair cost
Temperature
....
approximately the same . . . . . .

I
Starting torque ,,
requiring attention
Viscosity ....
Noise level
higher
non-problematic
less to equal
|
,,
,
lower
Figure 12.16 Drivingsystems comparison: canned motor vs. magnetic drive.

The application examples range from the extraction and compression
of dry chlorine gas to phosgene-bearing gases. In addition to these, the
author's company has had a number of enquiries for a wide variety of
substances also listed in the table included in the Accident Ordinance
(1988 ed.).
Safety
Once the customer has decided to opt for a leak-proof pump, the question
is how to monitor the leak tightness. The systems available are divided in
those monitoring the pump itself, and the other type which monitors the
area surrounding the pump.
Monitoring of the pump itself. Only by monitoring the magnetic drive or

canned motor direct, by means of an integrated liquid level regulator
(float switch, measuring probe, etc.).
Monitoring of the area surrounding the pump. By measuring the
temperature in the area of the can (PT 100), or at another appropriate
place; by monitoring the motor (overload, underload or break-off) in the
electrical circuit; gas or liquid detector, e.g. also in the control chamber
adjoining the can.
268
Monitoring of the process. By measuring the pressure, temperature and

product flow rate in that part of the system where the chemical process
takes place.
Bibliography
1. Segebrecht, U.: Vakuumpumpen und kompressoren erfiillen hohe Sicherheitsanforderungen; Pumpen - Vakuumpumpen - kompressoren (1987).
2. Behrend, C., Pinkofsky und wiechmann: Ist das f6rdern explosionsfiihiger
gemische mit wasserringkompressoren gefiihrlich? Die BG (August 1986).
3. Ripper, L.: Anforderungen an verdichter zur F6rderung chemisch instabiler
gase und explosionsfiihiger gas - und dampfgemische. Die BG (August 1986).
4. Ufer, W.: Grubenabsaugung mit Fliissigkeitsring-Gaspumpen. Gliickauf 122,
20 (1976).
CHAPTER
THIRTEEN
Leak-proof Roots
vacuum pumps
K. H. Ronthaler
Introduction
Roots pumps have been known in the areas of technology for a long time.
In 1848 the Englishman Isaiah Davies developed a pump whose concept
was later used by the Americans, Francis M. and Philander H. Roots and
became known as the 'Roots Blower'. Since then Roots pumps have been
employed as compression blowers and gas meters by reversing the
rotation of action. In about 1954 the Roots pump was introduced into the
field of vacuum technology. According to ISO 3529 Part 2 (DIN 28400
Part 2) edition 10/80 they are assigned as positive displacement pumps
(Fig. 13.1).
Vekuum ~--mpp~
I
~vvGa's transfer
ecuum pump
I
I
Lvskuum pump J
.... I
.
I : _
Iv ..4
[-
t.
I -
Recip;-ocatlng
~..,v.d,.-~
I ~',,~cemen, I
~I R~ p"mp I
l[ Drag pump iI
I
r
~,~~
Imentpump /
IPump
~
.
T,
pump
_ ,]
- : j
Diaphragm
I [ Plat . . . . . .
" 21
Gaseous
Liquid ring pump
9 pump
'-I
vec u um pump~
I iSoWUve disi P lecement
[,~,,t. ,pro...
ring pump
I ~
_ Ejector pump
_
"lldlno vane
_~
.o,.~..,.,oo
I I I: ~,.0~
... 1 I ~,.~176
drag pump
eJektor pump
-1.'~ ",,'urn,' I ~"l-l~~__J

pump
Rotary plinger
pump
Adlorptioh- i
I r Diffusions
l-l__""""
'~ Cry~
_f .o,,..o... 1
Figure 13.1 Pump assignments according to ISO 3529 Part 2 (DIN 28400 Part 2).
270
Operating principles of a Roots pump

Roots pumps are rotary pumps operating on the positive displacement
principle. The pumping mechanism consists of two counter-rotating
impellers of figure-of-eight design. The impellers sweep all internal surfaces of the pumping chamber for positive displacement of the gases. The
gas drawn in by the pump is compressed by the rotating impellers and
transported to the fore-vacuum side. Both impellers are synchronised by
a gear drive (ratio I'I) which enables them to rotate with little play both
to each other and the housing. The gaps resulting between the impeller
and the housing and between the impellers themselves are kept as small
as possible. The gap width is determined both by the size of pump and its
field of application. The gap width and the rotor geometry influence the
pumping efficiency (1/).
Figure 13.2 shows the operational schematic of a single-stage Roots
pump with vertical pumping action. In impeller positions I and II the
volume of the intake is increased. When the impeller reaches position III
part of this volume is cut off from the intake side. In position IV this
volume is opened to the exhaust side and gas under fore-vacuum pressure
(higher than the intake pressure) flows in. This gas compresses the gas
coming from the intake. As the impeller continues this compressed gas is
forced out through the exhaust flange. This process repeats for each
impeller twice per revolution.
Figure 13.2 Operational schematic of a single-stage Roots pump with vertical

pumping action.
271
Leak-proof Roots vacuum pumps
Normal Roots pumps cannot pump atmospheric pressure and therefore

require a backing pump such as a sliding vane, a rotary piston or some
other type of backing pump. In combination with an oil sealed backing
pump the Roots pump can extend its range of operation into the medium
vacuum range (fine vacuum).
When evacuating large volumes, it is advisable to reduce the pumping
pause time by additionally switching on the Roots pump at atmospheric
pressure. Due to thermal problems additional devices such as a gas inlet
cooling system, a hydrodynamic coupling or a pressure by-pass line for
overload protection must be added.
Figure 13.3 shows the pumping speed characteristics of Roots pump
systems with and without pressure by-pass lines.
Characteristics of the Roots vacuum pump

The effective quantity of gas (Qe[f) transferred by a Roots pump is
calculated from the theoretical transported quantity Qth and the backstreaming quantity (QiR)
Qeff = Qth - QiR.
(1)
The theoretical quantity is given by
1000
IIiii
IIIII
000
IIII!
IIIII
000
illil/
II11t'
.e. ,oo
' Iii[!
I~
100
lte-O$
-
IIII]
11111
Illlll
.~,l'lTlt
1
II!1
i'lli'l"t
,, i
i11111-"-,~ dl 1I!
!11111
I~_
llllil
II!!1
Jill
!111
~"
Iii!11
!!i111 IIIIII
!11111
"i-I IIII
..I. IIIII
, I'11111
Iii111
!11111
11o.-Ol
lg+OO
11e+01
I IIiill
I I!!111
~ ~,i
i11111
I!!1 !il/'11111!1111ilil
lle-o2
I illlfl
I IIIIII
iae+oIR
I I
lse+08
Intake p r e n ~ r e [mbar]
Figure 13.3. Pumping speed curve of a Roots pump system. (1) Pumping speed
of the backing pump. (2) Pumping speed of the Roots pump system without a bypass valve. (3) Pumping speed of the Roots pump system with a by-pass valve.
(4) Gain in pumping speed due to the pressure by-pass line.
272
Leak-free Pumps
and Compressors
(2)
Qth = Pa" Sth"
where Pa is the intake pressure and Sth the theoretical pumping speed.
Sth is the product of the swept volume Vs and the number of revolutions
n
Sth --- n . Vs.
(3)
Similarly the inner back streaming QiRis given by

QiR --
Pv. SiR-
(4)
where Pv is the pressure at the exhaust or fore-vacuum side and SiR the
reverse suction value
SiR "- n . ViR.
(5)
where ViR corresponds to the volume of gas transferred in the reverse

direction and n the number of revolutions (speed). The volume efficiency
1/is defined by
(6)
O -" Q e f f / Q t h .
When equations (1)-(4) are combined the following is obtained:

rl = 1
Pv/Pa
Substituting K for the compression ratio

71= 1 - K .
SiR/Sth.
PJPa
SiR/ath.
(7)
the following is obtained:

(7a)
The ratio Sth/SiR is a characteristic value for the internal back-streaming at

zero throughput and is designated as Ko. This characteristic value for the
Roots vacuum pump is usually given in respect to the fore-vacuum pressure
(see Fig. 13.4). The value of Ko depends on the type of gas being pumped.
The efficiency I/has the following relationship:
rl = 1 - K / K o .
(8)
When the Roots vacuum pump is operated in combination with a roughing vacuum pump with the pumping speed S~, the continuity equation
yields the following relationship:
273

60
6O
5~ ~
~0
l~O
10
0,001
iilli
Iill
LIll.,
~l
Liil LlL.I,IIILI
ll
Lli~,
Iill. .'~" llilll l ilLi
Illl Lllll84 Ill
I!!i llli l[llll IlL
0,01
0,1
Fore -
10
il(llL llll
Llill Ill
llilll
L!
!
[llllL
100
1000
vaouum premmre [mbar]
Figure 13.4 Shows the curve for the compression Ko of a Roots vacuum pump,
type RUVAC WS 1200, for air.
Sv. Pv = /7. Sth. Pa.
(9)
In this relationship only the values for the Roots pump - backing pump
combination come into consideration. This enables the calculation of the
pumping speed curve for a given combination (see Fig. 13.3).
The required power is determined by the compression power and the
friction loss (e.g. in the bearings). The total power required is given by the
relationship:
Ptot -
Sth.
(Pv -
e a ) "l- Y.
ev.
(10)
Technical design
The design of a Roots vacuum pump is shown in Fig. 13.5 (longitudinal
section).
In contrast to the rotary vane and rotary piston pumps the pump
chamber of a Roots pumps is free of any sealing material or lubricant. Oil
is only found in a separate side chamber used to lubricate the gear drive
and the rotor bearings. Leak-proof valves are advisable for the four rotor
shaft feed-throughs needed from the pump chamber to both side
chambers, however, they are not technically available due to the high
demands on:
274
Figure 13.5 Longitudinal section drawing of a Roots vacuum pump type

RUVAC.
9
9
9
9
service life/working life time (more than 20000 working hours)

peripheral speed (significantly higher than 10 m s-1)
high temperature resistance (more than 150~
simultaneously highest reliability as well as low cost.
This resulted in both side chambers being separated from the pump
chamber by a piston ring labyrinth seal. If an absolutely particle-free
and/or CHx-free vacuum is required then purge gas or evacuation of the
gear box is available as an option. The pressure in the side chambers is
usually the mathematical average of the pressure between that of the
intake and the exhaust, i.e. a pressure lower than atmospheric. Thus the
sealing problem is transferred to the drive shaft between the chamber and
the outer atmospheric pressure.
The normal direct drive for these rotary pumps is provided by a 3-phase
asynchronous motor. The aim is to provide a leak-proof drive system to
the impellers over a long period with a high reliability.
275
Types of drives
Demands on a direct drive of a Roots vacuum p u m p
The following demands are made on a leak-proof drive of a Roots pump:
9
9
9
9
9
9
9
9
9
9
9
integral He leak rate _< 1 x 10 -4 mbar l/s

operational life > 20000 h
operating reliability up to 100 Hz (6000 rain -1 driving speed)
high efficiency
low manufacturing costs
compact design
preferably air cooling, however, liquid cooling should be possible as an
option
resistant to environmental influences such as dirt, shocks etc.
simple maintenance
suitable for all voltages and frequencies worldwide
the use of an explosion proof standard motor must be possible.
Types of drives
In accordance with the demands discussed above, the following alternative drives are defined:
(1)
(2)
(3)
(4)
Standard motor with flange.

Magnetic coupling in combination with a flange motor.
Built-in motor.
Canned motor.
When sealing aspects are taken into account the above drive alternatives
can be divided into two groups:
9 Shaft lead through to atmosphere and simultaneous sealing with shaft
ring and/or slip ring seal (applies to drive alternative (1))
9 Leak-proof sealing of the piston driving shaft, i.e. no dynamic shaft
seals, only static seals (applies to drive alternatives (2--4)).
In the latter case the pumps are defined as hermetically sealed pumps, i.e.
the speed needed for the rotor is initiated via a rigid housing (canned
motor tube) by permanent or electromagnetic means. In the following
sections these alternative drives are described.
Drive with standard flange motor. The flanged standard motor is the
most simple type of drive (Fig. 13.6). Two oil-dipped radial shaft sealing
276
Figure 13.6 Roots vacuum pump with flange-mounted standard motor, type
RUVAC WA.
rings are used to seal the shaft lead-through of the drive piston between
atmosphere and vacuum. By a simple coupling a direct connection is
made to the flange motor.
The advantages of such a drive are:
9
9
9
9
9
low price
compact design
simple connection to the application specified motor (special edition)
high efficiency
simple operating conditions.
These properties have made the flanged motor a widely accepted drive for
the Roots vacuum pumps.
277
Drive with magnetic coupling and standard flange motor. This drive alternative excellently meets the requirements of leak-proofness, operational
life, and flexibility in the choice of motor to be selected. A further advantage is that the specifications laid down by the Physikalisch-Technische
Bundesan stalt (PTB Braunschweig/Germany) concerning the danger of
explosion in workshops are avoided.
Magnetic couplings can be of a permanent magnet coupling or of a
hysteresis coupling construction.
In principle, the direct drive has to be oversized because of the specific
spring rigidity curve, otherwise the coupling may break off during run up.
Due to the existing centrifugal forces as well as the required stability of
the material, the magnets must be enclosed in a capsule.
The result of the considerable expenditure to achieve a solution is a
relatively costly and bulky machine. In comparison to other solutions the
magnetic coupling is seldom used in Roots pumps.
Drive with a built-in motor. This alternative drive, comprising the stator,
the rotor with the drive shaft and the bearings, is at the same pressure as
the side chambers of the Roots pump. The electric motor is practically
installed in the vacuum system (see Fig. 13.7).
At first glance, this kind of drive seems amazingly simple, if it were not
for the physical problems arising from the dielectric strength of the motor
windings.
Figure 13.7 Built-in motor (with water cooling).
278
1400
1200
1000
8OO
E
800
-..d
4OO
2OO
tO
20
30
40
84)
00
p * 9 [tuber * mini
Figure 13.8 Paschen curve for air.

For electrical motor windings the 'Paschen' curve must be considered
(see Fig. 13.8). This shows the relationship between the break down
voltage Ud and the product pressure multiplied by the distance apart of
two electrodes (P * S).
Roots pumps with built-in motors operating under the normal supply
voltage range are in the region of critical voltage break down. The danger
of arcing is high, especially in areas of weak insulation, which considerably
reduces the working life of the motor.
In principle, this problem can be solved by better insulation of the
windings (e.g. embedding the windings in an insulating substance) and/or
reducing the operation voltage (additional transformer). Both solutions
are normally not considered because of the extra work and cost involved.
Therefore this type of drive is rarely used (with reduced voltage and special
precautions concerning the insulation of the windings).
Drive with an air-cooled canned motor. The canned motor is a 2-pole,
3-phase motor designed to operate under harsh conditions.
The characteristic of the canned motor is the separation of the stator
from the rotor by a vacuum-tight canned motor tube. The rotor operates
on the drive shaft under the vacuum conditions of the side chambers while
the stator remains exposed to atmospheric pressure. Thus neither the
shaft lead through to atmosphere nor the shaft seal, which would be
subject to wear, are needed. The above mentioned concerns about the
stability of the built-in motor are no longer a problem. High reliability and
a working life of well above 20000 hours prove that this type of drive is
well suitable for application in Roots pumps.
279
Figure 13.9 Roots vacuum pump with canned motor drive and air cooling.
Canned motors with either floating rotors or rotors supported on both
sides by bearings are commonly used with Roots pumps.
The floating type has proved more successful as it is the only principle
which offers a compact design (see Fig. 13.9)
The motor is cooled by an external fan (axial fan). Due to the increased
power loss the drive must be larger in design than standard motors with
comparable power (see Fig. 13.10).
Nevertheless, this very compact design is lower in cost than drives with
magnetic couplings and flange motors.
The motor illustrated in Fig. 13.9 has to be additionally modified when
explosive gases are used.
Design criteria and loss minimising of canned motors with air cooling. As
already mentioned the drive has to be stronger compared to a standard
motor due to power loss. By simple measures, however, the efficiency
can be increased by 10-20%.
-
Optimising the cooling system of the windings. The conductivity of the

windings decreases considerably with the rise in temperature. The motor
280
12
mmmmmmm
mmmmm
m
mmmmsmmmm
mmm
Immmmm
e
mmnmmmm
~ir
'
j
kon~.seov
so,,.
t
VA..~.n
8 ~~'-
o
0
Motor s h a f t power [Klr]
Figure 13.10 Power input as a function of the shaft power for a canned motor
with P = 5.5 kW.
must be designed in such a way that the windings be efficiently cooled and
that the temperature gradient of the windings be kept low. This can be
achieved by optimizing both the motor fan and the heat emission at the
coil ends (constant air flow around the winding).
Figures 13.11 and 13.12 show the increase in efficiency at a temperature
of 105 K for ISO-class F according to VDE 530 at the expense of a slight
increase in slip.
It is also common to use a cooling fluid in the canned motor tube to
cool the rotor. The rotor operating in the fluid requires additional power
to overcome the friction caused by the fluid. These friction losses are
calculated with a rotor diameter tolerance raised to the power of 10-5 and
a rotational speed accuracy raised to the power of 10-3. To keep this loss
of power as low as possible the relation of length/diameter of the motor
rotor must be considerably greater than one.
This prevents the use of a floating rotor which in turn affects the size of
the pump. Because of these factors Roots pumps normally dispense with
rotor cooling.
Reducing eddy current losses. As already mentioned it is necessary to

design a larger motor than normally required by a similar standard motor
due to the increased loss in power. Primarily this is due to the losses caused
by eddy currents. The reason being the electrical conductivity of the metallic can. The rotating field induces an electrical current which increases the
temperature of the can resulting in a negative influence on the efficiency

t5o
--
__i
140 ~
~'
130
""
120
,,
[ ......
i , t , , ~ . , < l fin wlth .,,,~. t
. . . . . . . . . . . . .
mL'iaoe ooll end =ntilatiom I

Al/-3?em'; Aottt-illl, ilom"
110
.~
_._.__2
100
90
.
.
.
.
.
.
..
.
.
.
.
.
.
.
281
I
.
........
. ,. .,. . . .
,,d,,,,l,,
.
\ /,'
'' , li t ~, ' f
,,,"i
.
~AU]IaUOD
.o
aO
eo
10
.
.
.
.
.
2
.
.
.
.
@
.
.
4
.
5
Figure 13.11 Measurements at the canned motor drive with PN = 5.5 kW and
VA-can with U = const. + 380 V and 50 Hz. 'Excess temperature of the winding
as function of the shaft power'.
of the motor. Therefore canned motors with metallic cans reach only
5 5 - 6 0 % of the normal output of a standard motor of the same size.
From the electrical power
P=V*I
(11)
V = R *I
(12)
and Ohm's law
3000
2950
"~ 2900
2850
1~ 2aoo
2?50
2700
Motor sh~t, po~rer [ K i l l
Figure 13.12 Measurements at the canned motor drive with PN = 5.5 kW and VAcan with U = const. = 380 V and 50 Hz. 'Motor revs as function of the shaft power'.
282
as well as the electrical resistance of a conductor
R= p , I/A [f~]
results in eddy current losses:
P = V2 * A-'* p-'.
(13)
The formula indicates that the lower the specific electrical resistance of
the can, the higher are the losses caused by eddy currents. This suggests
that the ideal can would be made of a non-conducting material such as
plastic (see Figs 13.13 and 13.14).
Figure 13.14 comparison between a canned motor BG 160 with a plastic
can and a standard motor BG 160/11 kW of identical design.
Plastic cans, especially those designed for higher temperature and
mechanical stress, are increasingly being used.
Intensive research has made it possible to design a canned motor with
reduced losses equivalent to those of the classic standard motor.
Summary
The demands made by environmental laws preventing emission of products that may be dangerous to life or health make the use of leak-proof
drives imminent.
120
.......
~.,100
/
.
/r
,/
,/
j
_l
10
12
Motor ohaft Dower [ X ~
Figure 13.13 Measurements of a canned motor drive with P~ = 6.25 kW and

U = const. = 380 V 'Excess temperature of winding as function of the shaft power
with VA- and plastic can'.
283
120
~100
,,__,,
, , ,
I-
j l
n
,,.
. . . . . . .
ml
II
v - v
10
12
Motor shaft power [lOfJ
Figure 13.14 Comparison between a canned motor BG 160 with a plastic can
and a standard motor BG 160/11 kW of identical design.
A drive that has reached these demands is the canned motor type drive
because of its:
9
9
9
9
very high reliability

very high operational time
minimal leak rate
rugged construction, low cost and compact design.
With the elimination of parts subject to wear, such as the shaft seal, the
operational life and the reliability of Roots pumps as well as the leak rate
are improved to well within the required specifications.
intensive research led to the development of a canned motor with a
power consumption reduced to that of the standard motor with similar
compact design, rugged construction and a continuous reduction in price
difference.
The conclusion is that the canned motor is the best alternative as a leakproof drive for modern Roots pumps.
Bibliography
1. ISO 3529 Part 2 (DIN 28400 Part 2) 10, Beuth-verlag, Berlin (1980).
2. Firmenschrift, Grundlagen der vakuumtechnik, 1, Fa. Leybold Heraeus
GmbH, K61n (1986).
3. Niirnberg, W.: Die asynchronmaschine, Springer-Verlag, Berlin (1979).
4. Ronthaler, K.H.: Untersuchung zur leistungserh6hung yon spaltrohrmotoren,
untersuchungsberichte EU 015 (1986), EU 043 (1989) (unver6ffentlicht), Fa.
Leybold AG K61n.
284
5. Schilling, R., Siegle, H., Stoffel, B.: Str6mung und verluste in drei wichtigen
elementen radialer kreiselpumpen-eine literaturfibersicht str6mungsmechanik
und str6mungsmaschinen, Mitteilungen des Institutes ffir Str6mungslehre und
Str6mungsmaschinen, Universit~it (TH) Karlsruhe (1974).
6. Steffens, R.: Leckfreie wfilzkolben-vakuunpumpen, vortrag im rahmen der
fachveranstaltung Leckfreie pumpen und verdichter, ed. Vetter, G., Haus der
Technik (1992).
7. Wutz, M., Adam. H., Walcher, W.: Theory and Practice of Vacuum Technolgy.
Buyer's guide
Centrifugal pumps
Magnetic drive
9 Ansimag
9 Cleghorn Waring Pumps
9 Hermetic Pumpen
9 H M D Seal/Less Pumps Ltd
9 IT'F Richter Chemie-Technik GmbH
9 Sihi Halberg
9 Stork Pompen
Canned motor
9 Hermetic Pumpen
9 Sihi Halberg
Reciprocating displacement pumps

Diaphragm, oil-free
9 Lewa Herbert
Diaphragm, double or single acting
9 Lewa Herbert
Metering pumps
9 Lewa Herbert
Dosing and proportioning pumps
9 Lewa Herbert
Solenoid pumps
9 Lewa Herbert
Magnetic drive rotary positive displacement pumps

Liquid ring vacuum pumps
9 Sihi Halberg
Internal gear
9 Stork Pompen
286
Sliding vane
9 Cleghom Waring Pumps
Orbital lobe
Other pumps
Non-metallic/plastic pumps
9 Hermetic Pumpen
9 I T r Richter Chemie-Technik GmbH
Motors, pump ancillaries and services

Electric motors
Magnetic drives/couplings
9 Ansimag
9 Sihi Halberg
9 Stork Pompen
Overload protection devices
9 Stork Pompen
Coatings and linings
9 Ansimag
Control and measurement
9 Lewa Herbert
Condition monitoring
9 Ansimag
9 Stork Pompen
Computer aided pump selection software (CAPSS)
9 Stork Pompen
Pump applications
Boiler circulating
9 Stork Pompen
Boiler feed (industrial)
9 Stork Pompen
Boiler feed (power station)
9 Stork Pompen
Brewery stuff
Cargo oil
9 Stork Pompen
Buyer's guide
Cargo stripping
9 Stork Pompen
Chemical process
9 Ansimag
9 Hermetic Pumpen
9 ITI" Richter Chemie-Technik GmbH
9 Lewa Herbert
9 Stork Pompen
Chemical abrasive
9 ITT Richter Chemie-Technik GmbH
9 Lewa Herbert
Condensate extraction
9 Hermetic Pumpen
Cooling water
9 Ansimag
9 Hermetic Pumpen
9 Sihi Halberg
9 Stork Pompen
Cryogenic (liquid gases)
9 Ansimag
9 Hermetic Pumpen
Descaling
9 Lewa Herbert
Foodstuffs and drink
9 Lewa Herbert
9 Sihi Halberg
Fuel Oil (heavy)
9 Hermetic Pumpen
9 Stork Pompen
Fuel Oil (light)
9 Ansimag
9 Hermetic Pumpen
9 Sihi Halberg
9 Stork Pompen
287
288
Garden fountain
9 Lewa Herbert
Glue
9 Stork Pompen
Grease/lubricating oil
9 Stork Pompen
Heating
9 Sihi Halberg
9 Stork Pompen
High pressure pumps
9 Ansimag
9 Lewa Herbert
9 Hermetic Pumpen
9 Sihi Halberg
9 Stork Pompen
Laboratory
9 Ansimag
9 Hermetic Pumpen
9 Lewa Herbert
Liquor pumps
9 Ansimag
Machine tool lubricating (suds)
9 Ansimag
Molasses
9 Stork Pompen
Oil burner/fuel injection
9 Stork Pompen
Oil extraction
9 Ansimag
9 Stork Pompen
Oil pipeline
9 Ansimag
9 Stork Pompen
Oil transfer
9 Ansimag
9 Hermetic Pumpen
9 Stork Pompen
Buyer's guide
Petrol/light fuel/solvents
9 Ansimag
9 ITY Richter Chemie-Technik GmbH (Solvents)
9 Sihi Halberg
9 Stork Pompen
Portable pumps
Pulp and paper
9 Ansimag
9 Lewa Herbert
9 ITF Richter Chemie-Technik GmbH
Printers' ink
9 Ansimag
9 Stork Pompen
Radioactive liquid
9 Lewa Herbert
9 Sihi Halberg
Shower booster
9 Ansimag
Sinking/dewatering wellpoint
9 Ansimag
Tar and liquor
9 Stork Pompen
Viscous
9 Lewa Herbert
9 Stork Pompen
Water
9 Ansimag
9 Hermetic Pumpen
9 Sihi Halberg
9 Stork Pompen
289
Trade n a m e s index
A N S I M A G - heavy duty, mag-drive, seal-less, non-metallic centrifugal
pumps (ANSI and I S O ) - Ansimag Inc.
A L - non-metallic ANSI/ISO to 25 k W - HMD Seal/Less Pumps Ltd
CASTER - Seal-less magnetic drive pump - Cleghorn Waring Pumps
CS - from -80 to + 150~ - HMD Seal/Less Pumps Ltd
GS - ANSI/ISO range up to 50 k W - HMD Seal/Less Pumps Ltd
GSL - from -100 to + 1150~
GSS - low flow, high h e a d - HMD Seal/Less Pumps Ltd
G T - Small (0.5-3.5 kW) close coupled- HMD Seal/Less Pumps Ltd
HD - API and similar duties- HMD Seal/Less Pumps Ltd
H E R M E T I C - canned motor pumps, pumps with permanent magnetic
p u m p s - Hermetic Pumpen GmbH
INSIGHT II - patented electronic pump condition m o n i t o r - HMD
Seal/Less Pumps Ltd
LEWA m o d u l a r - metering pumps for all applications- LEWA Herbert
LEWA ecodos - standard diaphragm metering pumps - LEWA Herbert
LEWA t r i p l e x - high pressure process diaphragm p u m p s - LEWA
Herbert
LEWA lab - diaphragm metering pumps for laboratories - LEWA
Herbert
LEWA metering systems - fluid metering with closed-loop controls LEWA Herbert
LEWA p a c k a g e s - custom tailored metering pump p a c k a g e s - LEWA
Herbert
LEWA t r o n i c - instrumentation for metering pump packages - LEWA
Herbert
Q-MAX - seal-less magnetic drive pump - Cleghorn Waring Pumps
S A F E G L I D E - dry-run optimized SIC-plain b e a r i n g s - ITI' Richter
Chemie-Technik GmbH
SP - tanker off loading speciality - HMD Seal/Less Pumps Ltd
STORK
- mag drive p u m p s - Stork Pompen Nederland BV
Editorial index
A
Agitators, magnetic drives 25, 63--64
Antifriction bearings 127
Automated processes, metering of fluids
181, 189-211
B
Bearings 17
chemical pumps 127
Safeglide 81-83
service life 127
silicon carbide 55, 64, 71, 78-83
Bellows pumps 164, 165, 176, 185
Blowers, buyer's guide 286
Buyer's guide 285-289
C
Canned motor pumps 17, 18-19, 89-125,
127-147
see also Liquid ring pumps
advantages 89
axial thrust 91, 116-119
can thickness 102
cell scavenging system 107-109
chemical pumps 127-147
component materials 131
construction 23
cooling 94-96, 138
costs 156-157
design 90-92, 128-131
differential heads 99-101
disadvantages 67
double wall security system (DWS)
122-124, 125
efficiency 20, 140, 154
explosion protection 18, 109-111, 124,
131-138, 151
externally cooled 94-95
flameproof enclosure 133-135
gas, liquefied 93-94, 103-105, 151
heat transport 94-96
high-pressure systems 102-105, 154
increased safety 133-134
inductive load monitoring 120-122, 125
installation 155
interchangeability 91-92, 143-145
level measurement 111-113, 124
magnetic drive pump comparison 149-158

modular construction 143-144
monitoring 111-125, 142-143
multistage design 90--91, 93, 100
noise levels 115, 140-141, 155-156
opto-electronic monitoring 113
pressure monitoring 119-120, 125
protective devices 142-143
Roots pumps see Roots pumps
rotor position measuring device 116-119,
124-125
safety 98, 109-111, 149-151
sandwich can 21
seals, static 98
self-cooled 95-96
self-priming 105-109, 145
service life 140
serviceability 152
side channel design 106, 145
single-stage design 91-92, 93-94
specialized applications 93-105, 145-147
standardization 143-145
starting behaviour 155
stator chamber 119-120
suspensions pumping 97
tandem 100-101
temperature measurement 111-113, 124,
135-138
temperature ranges 95, 152-153
terminal box 102
test certificate 115
testing 136-138
thermal oils 94-95, 151
thermistors 135-136, 143
vibration measurement 115-116, 124
Cell scavenging system 107-109
Centrifugal pumps 17-22
see also Canned motor pumps; Magnetic
drive pumps
buyer's guide 285
configurations 18-20
efficiency 20, 225-226
limiting factors 2
plastic 19, 69--87
safety 20-22
semi-hermetic 105
shaft seals 7--8
294
Ceramics
isolation shrouds 35
magnetic drive pumps 47, 52
plastic pumps 71
Chemical industry
canned motor pumps 91, 127-147
conventional pumps 127-128, 140
diaphragm pumps 213--214
magnetic drive pumps 51
plastic pumps 69
safety 45
Combustible liquids, pump comparison 151
Compressors
see also Liquid ring pumps; Roots pumps
buyer's guide 286
diaphragm type 231-249
applications 232, 243-244
capacity regulation 246
design 232-235, 238-240
drive mechanism 244
efficiency 241
flow diagram 248
hydraulic drive 235-237
monitoring 242
multi-stage 238
oil overflow valves 236
operation 233, 247
safety 247
sandwich diaphragm 233, 242
service life 242
types 243
types 231
Control circuits, metering pumps 202-207
Conventional pumps, disadvantages
127-128
Cooling, canned motor pumps 94-96,
138-139
Costs, pumps comparison 156-157
D
Diaphragm compressors see Compressors

Diaphragm pumps
see also Metering pumps; Process
diaphragm pumps
advantages 213-214
properties 16-17
safety 20-21
types 9-12
Diaphragms
see also Compressors; Metering pumps;
design 12-15, 175-179
elastomer 14
PTFE 13-14
sandwich type 20-21
service life 14-15
Dosing pumps 9-10

see also Metering pumps
Double wall security system (DWS)
122-124, 125
Dry-running, protection 78--83
E
Efficiency, pumps comparison 154

canned motor pumps 18, 109-111,
131-138, 151
pump comparison 151
F
Fire hazard
flameproof enclosure 133-135
pump comparison 151
Flow meters 203--207
Fluid conveying pumps 10-12
Flush systems, magnetic drive pumps
55-57, 64-65
G
Gases
see also Compressors
liquid, canned motor pumps 93-94,

103-105, 151
self-priming pumps 105-109
Glass technology, bellows-type pumps 164,
185
H
Hastelloy
canned motor pumps 128
isolation shrouds 33, 54
Hazardous locations, canned motor pumps
109-111
Hazardous substances
canned motor pumps 98, 150-151
liquid ring pumps 254-255
pump comparison 150-151
Hermetic pumps, definition 1
High performance liquid chromatography
171
High-pressures
canned motor pumps 102-105
pump comparison 154
I
Impellers, flouroplastic 74
Inductive load monitoring 120--122,
125
J
Journal bearings, design 55
Editorial index
L
Leak-free pumps, definition 1
Leakage rates 214
Liquid ring pumps 251-268
applications 254-255, 263-267
canned motor 257-259
design 251-254
hazardous substances 254-255
magnetic drive 259-263
operating behaviour 255-256
operation 251-254
safety 254-255, 267-268
M
Magnetic drive pumps 17, 19-20, 23-49,

51-67
see also Liquid ring pumps
advantages 67
air gap moments 35-37, 39
aluminium nickel cobalt 24, 53
buyer's guide 285-286
can materials 74, 77
canned motor system comparison
149-158
ceramics 47, 52
characteristics 25-35
corrosion resistance 35
costs 156--157
design criteria 23-25
design examples 62-66
disadvantages 127-128
double wall system 61--62, 124, 149
eddy currents 19, 30-31, 54
efficiency 20, 29-32, 54, 154
explosion protection 151
flush system 55-57
separate 64-65
future developments 47
heat loss 34-35
high pressure systems 154
installation 155
isolation shrouds 30, 32-35, 54, 61-62
journal bearings 55-57
magnet materials 52-53
monitoring 57--60
noise levels 155-156
non-metallic cans 19, 74-78
plastic 19, 69--87
can damage monitoring 78
can units 74--78
clamshell design 71
construction 71-74
double can 74-78
dry-running 78-83
efficiency 78
flushing 85-87
295
materials 69-71
solids in medium 74, 83-87
pump types 25
reaction moment 32
safety 21, 45-47, 60-62, 149-151
samarium cobalt 24, 25, 45-46, 53
screw pump 66
seal-less 53
Sealex CDS system 61-62
seals 45
self-priming side channel 65
serviceability 152
shock factor 40
shock impulse measurement 59-60
starting behaviour 37, 40-42, 52, 53, 155
static moment 31
synchronous coupling 51-53
temperature ranges 45-47, 53, 152-153
torque 25-29, 37-43, 51-52, 155
vibration measurement 59
zero-leakage 51-67
Metering pumps 159-187, 189-211
accuracy 165, 184-185, 190-195
applications 160-161
automated processes 181, 189-211
bellows 164, 165, 176, 185
control circuits 202-207
diaphragm pumps 9
accuracy 185
design 195-197
efficiency 182, 185
high pressure 180
hydraulically actuated 9, 166--174,
179-181,183, 185-186, 199-200
mechanically actuated 9, 161-166, 182,
184-185, 197-198
plunger pump comparison 184-186
remote actuation 179
diaphragms
design 175-179
elastomer 164, 165, 175-176
metal 170-171, 178-179, 180
position control 167-171
PTFE 164-166, 176, 180, 197
safety 176-179
sandwich 164, 166, 177-178, 198, 200
service life 165-166, 175-176
efficiency 192-193
flow monitoring 206-210
gas bubbles 168-170, 185
magnetic linear drive 161, 171-172
measuring time cycle 207-210
micro-dosing 171-172, 180--181
plunger pumps, diaphragm pump
comparison 184-186
296
L e a k - f r e e P u m p s a n d Compressors
Metering pumps (contd)

reliability 190-195, 200-202
rotary 184, 193-194
spring-loaded cam drive 161-162,
171
stroke adjustment 172-174
suspensions 97, 176
venting valves 168
Monitoring
canned motor pumps 111-125,
142-143
magnetic drive pumps 57-60
Motors, buyer's guide 286
N
Niveaustat level monitoring device 113

Noise levels, pump comparison 155-156
Non-metallic/plastic pumps, buyer's guide
286
Nuclear power stations, liquid ring pumps
257
P
Peristaltic pumps 3
Pharmaceutical industry, plastic pumps
69
Plastic see Magnetic drive pumps
Plunger pumps
diaphragm pump comparison 184-186
disadvantages 213-214
microdosing 171
Plunger seals 5
Positive displacement pumps see Liquid
ring pumps; Roots pumps
Pressure switch, canned motor pumps
119-120
Process diaphragm pumps 213-229
applications 225
chemical industry 213-214
designs 214-220
diaphragms 220-221
ecological benefits 227
efficiency 225-227
modular design 214-216
start-up arrangement 223
torque 222-223
triplex design 214, 217-218
Process medium, solids content 83-87
Pump ancilleries, buyer's guide 286
Pump applications, buyer's guide
287-289
Pump comparisons 149-158, 184-186,
213-214, 225-227
Pump/magnetic drive/motor drive system
29, 35-44
torque 37-44
Radioactive substances
canned motor pumps 98
liquid ring pumps 255
metering pumps 179
Reciprocating displacement pumps 2,
9-17
see also Diaphragm pumps; Metering
pumps; Process diaphragm pumps
buyer's guide 286
efficiency 192-193
Roots pumps 269-284
built-in motor 277-278
canned motor 278-283
cooling 278-280
eddy current losses 280-282
efficiency 280
plastic 282
design 273-274
drive types 275-282
flange motor 275-277
magnetic coupling 276-277
operating principles 270-271
service life 274
Rotary displacement pumps 2
see also Liquid ring pumps;
Roots pumps
configurations 18-20
efficiency 20, 193, 225-226
metering use 159, 184, 193-194
Rotor position measuring device (RPM)
116-119, 124-125
s
Safeglide bearings 81-83

Safety
magnetic drives 45-47
pump comparison 149-151
zero-leakage pumps 60-62
Samarium cobalt see Magnetic drive pumps
Screw pump, magnetic drive 66
Sealex CDS system 61-62
Seals 4-9
dynamic 214
flushing 5
high pressure 8-9
mechanical 127
plunger 5
Self-priming pumps 105-109
side channel 65, 106, 145
Serviceability, pump comparison 152
Shafts, problems 127
Shock impulse measurement (SPM) 59-60
Side channel pumps, self-priming 65, 106,
145
Editorial index
Silicon carbide
bearings 55, 64, 71, 78-83
seals 8-9
Slurries, diaphragm pumps 12, 14
Solids
canned motor pumps 97, 107
plastic pumps 74, 83-85
Spare parts 91-92, 218
Suspensions, canned motor pumps 97
Symbols 49
T
Temperature ranges, pump comparison
152-153
Thermal oils, canned motor pumps 94--95,
151
Thermistors 135-136, 143

Torque, magnetic drive pumps 25-29,
37--43, 51-52
Trade Names index 291
q3~e test 135
V
Vacuum pumps see Liquid ring pumps;
Roots pumps
Vibration measurement
canned motor pumps 115-116
Z
Zero-leakage pumps 51--67
297
Index to advertisers
Ansimag, Inc., 1090 Pratt Boulevard, Elk Grove Village, Illinois 60007,
USA
Tel: + 1 708 290 0482 Fax: + 1 708 290 0481 . . . . . . . . . . . . . . .
Facing 156
APV Industrial Pumps & Mixers, 41-43 Glenburn Road, College Milton
North, East Kilbride, Glasgow G74 5BJ, United Kingdom
Tel: +44 (0)13552 25461 Fax: +44 (0)13552 63496 . . . . . . . . . Facing 149
Cleghorn Waring Pumps, Icknield Way, Letchworth, Hertfordshire
SG6 1EZ, United Kingdom
Tel: +44 (0)1462 480380 Fax: +44 (0)1462 482422 . . . . . . . . . . . Facing 6
Dickow Pumpen KG, Siemenstrasse 22, 8264 Waldkraiburg, Germany
Tel" +49 8638 602232 Fax: +49 8638 5520 . . . . . . . . . . . . . . . . .
Facing 54
Hermetic Pumpen GmbH, Gewerbestr 51, D-79194 Gundelfingen,
Germany
Tel" +49 761 583 0220 Fax: +49 761 583 0280 . . . . . . Facing 55, 125, 155
HMD Seal/Less Pumps Ltd, Hampden Park Industrial Estate, Eastbourne, East Sussex BN22 9AN, United Kingdom
Tel: +44 (0)1323 501241 Fax: +44 (0)1323 503369 . . . . . . . . . . . Facing v
ITT Richter, PO Box 100609, D-47883 Kempen, Germany
Tel" +49 2152 146 190 Fax: +49 2152 146 190 . . . . . . . . . . . . . .
Facing 12
Klaus Union, Postfach 101349, D-44713 Bochum, Germany

Tel" +49 2344 595 203 Fax: +49 2344 595 204 . . . . . . . . . . . . . .
Facing 62
Lewar Herbert Ott GmbH & Co., Ulmer Str. 10, 71229 Leonberg,
Germany
Tel: +49 7152 140 Fax: +49 7152 14303 . . . . . . . . . . . . . . . . . .
Facing 202
SIHI Halberg, Lindenstrabe 170, D-25524 Itzehoe, Germany
Fax: +49 4821 771274 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Facing 154
300
Stork Pompen BV, PO Box 9, 9400 AA Assen, The Netherlands

Tel: +31 5920 40541 Fax: +31 5920 41942 . . . . . . . . . . . . . . . . . .
Facing 7
Wanner International Ltd, Grange Court, Grange Road, Tongham, Surrey
GU10 1DW, United Kingdom
Tel: +44 (0)1252 781234 Fax: +44 (0)1252 781235 . . . . . . . . . . Facing 13

G. Vetter-Leak-Free Pumps and Compressors Handbook-Elsevier Science (1996)

Uploaded by

Copyright:

Available Formats

G. Vetter-Leak-Free Pumps and Compressors Handbook-Elsevier Science (1996)

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

G. Vetter-Leak-Free Pumps and Compressors Handbook-Elsevier Science (1996)

Uploaded by

Copyright:

Available Formats

Preface

The increasingly strict regulations to avoid dangerous and hazardous

Leak-free Pumps and Compressors

The first edition of these papers published as a book in 1989 by Vulkan

Terms and their significance

*Hermes Trismegistos, a person from ancient Greek history, was considered to

Leak-free Pumps and Compressors

Figure 1.1 Leak-freereciprocating displacement pumps.

Leak-free displacement pumps

Figure 1.2 Leak-freerotary centrifugal and displacement pumps.

Peristaltic pumps with mechanical drive mechanisms for the rollers

Leak-free Pumps and Compressors

Rotary peristaltic pumps.

For fluids possessing special properties - presence of abrasive particles,

Leak-free displacement pumps

Figure 1.4 High pressure plunger seal.

Leak-free Pumps and Compressors

Figure 1.6 Sterile interfaces with flushing provisions.

Leak-free displacement pumps

Figure 1.7 ~;haftseals for centrifugal pumps.

Leak-free Pumps and Compressors

Figure 1.8 Rotating mechanical seals for operation at high pressure.

Leak-free displacement pumps

Leak-free reciprocating displacement pumps

Metering pumps with mechanically actuated diaphragms are applied for

Leak-free Pumps and Compressors

Fluid conveying pumps with hydraulically actuated diaphragms and

Figure 1.9 Leak-free reciprocating displacement pumps, application limits,

Leak-free displacement pumps

Leak-free Pumps and Compressors

Figure 1.12 Diaphragm metering pump with hydraulic diaphragm position

Leak-free displacement pumps

Leak-free Pumps and Compressors

Figure 1.14 Diaphragm pump for slurries with a tube-shaped diaphragm

Leak-free displacement pumps

Figure 1.15 Reciprocatingdisplacement pumps for slurries (WIRTH, GEHO).

Figure 1.16 Elastomer diaphragms: (a) plane-parallel diaphragm, (b) undulated

Leak-free Pumps and Compressors

Properties and characteristics

(f) If attention is paid towards suitable suction conditions (unobstructed

Leak-free centrifugal and rotary displacement

Leak-free Pumps and Compressors

Principal configurations and properties

Figure 1.18 Routing of the partial flow of fluid in a canned motor

Leak-free Pumps and Compressors

In case of can rupture, the outer housing normally serves as a temporary

Leak-free displacement pumps

Leak-free Pumps and Compressors

Figure 1.21 Canned-motor pump with sandwich-can and rupture monitoring

Properties and design

The state of current technology

Leak-free Pumps and Compressors

Figure 2.1 Principlesof construction of a magnetic drive system.

Magnetic drives on pumps

Characteristics of a magnetic drive

Magnetic drives on pumps

Figure 2.4 Ball valve with magnetic drive.

Leak-free Pumps and Compressors

Figure 2.6 Characteristic torque of a magnetic drive.