Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Fluid Flow

Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

-4

FLUID FLOW

4.1 INTRODUCTION The characteristics and complexity of flow pattern are such
that most flows are described by a set of empirical or semi-empirical
Transportation of fluids is important in the design of chemical equations. These relate the pressure drop in the flow system as

--
process plants. In the chemical process industries (CPI), pipework a function of flow rate, pipe geometry, and physical properties
and its accessories such as fitting make up 20-30% of the total of the fluids. The aim in the design of fluid flow is to choose
design costs and 10-20% of the total plant investment. Mainte- a line size and piping arrangement that achieve minimum capital
nance requirements and energy usage in the form of pressure drop and pumping costs. In addition, constraints on pressure drop and
(AP) in the fluids being pumped add to the cost. Also, these items maximum allowable velocity in the process pipe should be main-
escalate each year in line with inflation. As a result, sound pipe- tained. These objectives require many trial-and-error computations
sizing practices can have a substantial influence on overall plant which are well suited with the aid of a computer.
economics. It is the designer’s responsibility to optimize the pres-
sure drops in piping and equipment and to assess the most economic
conditions of operations. Figure 4-1 illustrates piping layouts in a 4.2 FLOW OF FLUIDS IN PIPES
chemical plant.
Pressure drop or head loss in a piping system is caused by fluid
rising in elevation, friction, shaft work (e.g., from a turbine), and
turbulence due to sudden changes in direction or cross-sectional
area. Figure 4-2 shows the distribution of energy between two
points in a pipeline. The mechanical energy balance equation
expresses the conservation of the sum of pressure, kinetic, and
potential energies, the net heat transfer q. the work done by the
system w , and the frictional energy e,. The e, term is usually posi-
tive and represents the rate of irreversible conversion of mechanical
energy into thermal energy or heat, and is sometimes called head
loss, friction loss, or frictional pressure drop. Ignoring this factor
would imply no energy usage in piping.

L i

q -w = 1/ d P + a / udu + g / dz + ef (4- 1)
1 I

Total head (energy grade line)

-
144 x P, -
144 x P,

P,
Pressure head, P 02

1 Velocity head

-.+Flow- 2 2

Elevation head

Arbitrary horizontal datum line Pipe length (ft)


7L’44p,+v:,z *144p,+y,2+h
‘ P, 29 P2 29

Figure 4-1 Chemical plant piping layout. (Source: IChemE safer piping
training package (Courtesy of the I.ChemE., UK)) Figure 4-2 Distribution of fluid energy in a pipeline

133
134 FLUID FLOW
Integrating Eq. (4-1) gives In general, pressure loss due to flow is the same whether the pipe
is horizontal, vertical, or inclined. The change in pressure due to
the difference in head must be considered in the pressure drop
calculation.
or Considerable research has been done on the flow of compress-
ible and non-compressible liquids, gases, vapors, suspensions, slur-
ries, and many other fluid systems to allow definite evaluation
of conditions for a variety of process situations for Newtonian
+ef+w-q (4-3) fluids. For non-Newtonian fluids, considerable data are available.
However, its correlation is not as broad in application, due to the
where significant influence of physical and rheological properties. This
presentation reviews Newtonian systems and to some extent the
P = pressure (force/area) non-Newtonian systems.
p = fluid density (mass/volume) Primary emphasis is given to flow through circular pipes or
g = acceleration due to gravity (lengthhime’) tubes since this is the usual means of movement of gases and liquids
z = vertical height from some datum (length) in process plants. Flow through duct systems is treated with the fan
u = fluid velocity (lengthhime) section of compression in Volume 3.
e, = irreversible energy dissipated between points 1 and 2
(length2/time2)
w = work done per unit mass of fluid [net work done 4.3 SCOPE
by the system is (+), work done on the system is (-)
The scope of this chapter emphasizes applied design techniques
(mechanical energykime)]
for 85% & of the usual situations occurring in the design and
q = net amount of heat transferred into the system (mechanical
evaluation of chemical and petrochemical plants for pressure and
energykime)
vacuum systems (Figure 4-3). Whereas computer methods have
A = difference between final and initial points
been developed to handle many of the methods described here, it
a = kinetic energy correction factor ( a % 1 for turbulent flow,
is the intent of this chapter to present only design methods that
a = 1/2 for laminar flow).
may also be applied to computer programming. First, however, a
The first three terms on the right side in Eq. (4-2) are point func- thorough understanding of design methods and their fundamental
tions - they depend only on conditions at the inlet and outlet of variations and limitations is critical. There is a real danger in losing
the system, whereas the w and e, terms are path functions, which sight of the required results of a calculation when the computer
depend on what is happening to the system between the inlet and program is “hidden” from the user and the user becomes too enam-
the outlet points. These are rate-dependent and can be determined ored with the fact that the calculations were made on a computer. A
from an appropriate rate or transport model. good designer must know the design details built into the computer
The frictional loss term e, in Eq. (4-1) represents the loss of program before using its results. There are many programs for
mechanical energy due to friction and includes losses due to flow process design that actually give incorrect results because the
through lengths of pipe; fittings such as elbows, valves, orifices; programmer was not sufficiently familiar with the design proce-
and pipe entrances and exits. For each frictional device a loss term dures and end 1imitsAimitations of the method. Then, when such
of the following form is used programs are purchased by others, or used in-house by others, some
serious and erroneous design results can be generated. On the other
U2 hand, many design procedures that are complicated and require
ef = K,- 2 (4-4) successive approximation (such as distillation) but are properly
where K , is the excess head loss (or loss coefficient) due to the programmed can be extremely valuable to the design engineers.
pipe or pipe fitting (dimensionless) and u is the fluid velocity In this book, reference to computer programs is emphasized
(lengthhime). where necessary, and important mechanical details are given to
For fluids flowing through pipes, the excess head loss term K, emphasize the mechanical application of the process requirements
is given by (Figure 4-4). Many of these details are essential to the proper
functioning of the process in the hardware. For two fundamental
L aspects of fluid flow, see Figures 4-3 and 4-4 (Moody diagram).
K , = 4fF, (4-5) In the laminar region or the viscous flow (e.g., Re c 2000), the
where roughness and thus the relative roughness parameter do not affect
the friction factor which is proportional only to the reciprocal of
fF= Fanning friction factor the Reynolds number. The slope of the relationship is such that
L = flow path length (length) fi,, = 64/Re, which is plotted in Figure 4-5. In this region, the only
D = flow path diameter (length). fluid property that influences friction loss is the viscosity (because
Note: the density cancels out). The “critical zone” is the transition from

g, = 32.174 (s)(f&)
Conversion factor: g, [ML/ft2], F = ma/g,
= 9.806 kg m = 980.6 (*)cm
g
laminar to turbulent flow, which corresponds to values of Re from
about 2000 to 4000. Data are not reproducible in this range, and
correlations are unreliable. The transition region in Figure 4-5 is
the region where the friction factor depends strongly on both the
Reynolds number and relative roughness. The region in the upper
lb, ft g,cm
kg, m right of the diagram where the lines of constant roughness are
= (poundals:) = (x) = (dyns’) horizontal is known as “complete turbulence”, “rough pipes”, or
“fully turbulent”. In this region the friction factor is independent of
=l(%) Reynolds number (e.g., independent of viscosity) and is a function
only of the relative roughness.
4.3 SCOPE 135

p= P + Pbr Any pressure Level Above Atmospheric [gauge or absoiute=(gauge)+(barometer)]

-N
Sea Level Standard
P' !I
iA
----%-
760 mmHg abs, or
14.696 psia
0 psig
Atmospheric Pressure (pa,),vades with GeographlcalAltitude Location,called

'
Absolute Absolute Pressure

-{
vacuum] is Above Reference
Measurement of Absolute Zero
I Dsia I Absolute Zero Pressure
(Perfect or Absolute Vacuum)
Also, Absolute Reference Level

Notes: 1. At sea level, barometric pressure= 14.696 Ib/in2. absolute, or 760 mm of mercury,
referred to as "standard". This is also 0 Ib/in'. gauge for that location.
2. Absolute zero pressure is absolute vacuum. This is 0 psia, also known as 29.92 in. of
mercury below atmospheric pressure, or 33.931 fi of water below atmospheric, all
referenced at sea level.
3. Important equivalents: 1 atm pressure at sea level =
(a) 14.696 psia
(b) 33.931 fl of water (at 60" F)
(c) 29.921 in.Hg (at 32" F)
(d) 760 mmHg (at 32" F)
(e) 1.0332 kg/cmz
(f) 10,332.27 kg/rn2
4. Barometric pressure for altitudes above "standard" sea level are given in the appendix.
These correct values must be used whenever the need for the local absolute barometric
pressure is involved in pressure calculations.
5. Vacuum is expressed as either
(a) Inches (or millimeters) vacuum below atmospheric or local barometric, or
(b) Inches vacuum absolute, above absolute zero pressure or perfect vacuum.
(c) For example, at sea level of 29.92 in.Hg abs barometer; ( 1 ) 10 in. vacuum is a gauge term,
indicating l o i n. Hg below local barometric pressure; (2) 10 in. vacuum (gauge) is
equivalent to 29.921 in.Hg abs- 10 in. = 19.921 in.Hg abs vacuum.

Figure 4-3 Pressure level references. (Adapted by permission from Crane Co., Engineering Div., Technical Paper No. 410, 1957.)

stop Safety

Spira-tec trap
leak indicator
J 1

Figure 4-4 Portion of a plant piping system. (By permission from Spiral-Sarco, Inc. 1991.)
Figure 4-5 Moody or “regular” Fanning friction factors for any kind and size of pipe. Note: T h e friction factor read from this chart is four times the value of the .f’ factor rcad from Perry’s
Handbook, 6th ed. [ I ] (Reprinted by permission from pipe Friction Manual, 1954 by the Hydraulic institute. Also see Engineering Dutubook, 1st ed., The Hydraulic Institute, 1979 121. Data
form L.F. Moody, “Friction Factors for Pipe Flow” by ASME [ 3 ] . )
4.6 COMPRESSIBLE FLOW: VAPORS AND GASES 137
4.4 BASIS The single-phase friction loss (pressure drop) for these situa-
tions in chemical and petrochemical plants is still represented by
The basis for single-phase and some two-phase friction loss (pres-
sure drop) for fluid flow follows the Darcy and Fanning concepts the Darcy equation with specific limitations [4]:
(e.g., the irreversible dissipation of energy). Pipe loss can be char-
1. For larger pressure drops in long lines of a mile or greater
acterized by either the Darcy or Fanning friction factors. The exact
in length than noted above, use methods presented with
transition from laminar or viscous flow to the turbulent condition
the Weymouth, Panhandle Gas formulas, or the simplified
is variously identified as between a Reynolds number of 2000 compressible flow equation. (Can integrate the compressible
and 4000. form of the Bernoulli equation directly for ideal gas. Must be
For an illustration of a portion of a plant piping system, see careful to recognize when the flow is choked).
Figure 4-4. 2. For isothermal non-choked conditions [4]:

4.5 INCOMPRESSIBLE FLOW


The friction loss for laminar or turbulent flow in a pipe [4]

APf = p f v 2 L Ib,/in.’ (4-7)


144D(2g) ’
(4-10)
In SI units,
where

AP, = - (4-8) A= cross-sectional area of pipe or orifice, in.’


g,= dimensional constant = 32.174(lb,/lbf)(ft/s2)
In terms of feet (meters) of fluid h, is given by f= Moody friction factor
L= length of pipe, ft
D = internal diameter of pipe, ft
hf = -
f L ’’
ft (m) of fluid flowing
D (2g) ’
(4-9) P’= Pressure, Ib/in.’ abs
p = fluid density, lb/ft3
g = acceleration of gravity = 32.2ft/s2 (9.81m/s2) Subscripts
See nomenclature for definition of symbols and units. The
units presented are English engineering units and Metric units. The 1 = inlet upstream condition
friction factor is the only experimental variable in (that must be 2 = outlet downstream condition.
determined by reference to) the above equations and it is repre-
sented by Figure 4-5. Note that this may sometimes be (referred In SI units.
to as) expressed in terms of the Fanning formula which is (and
may be modified to yield a friction factor) one-fourth that of the
Darcy factor ( e g , f o = 4fF). Also, it is important to note that
the Figure 4-5 presented here is the Moody friction chart in terms
of the Darcy friction factor, recommended and consistent with the
engineering data of the Hydraulic Institute [2] (It is used mostly (4- 11)
by Mechanical and Civil engineers, but Chemical engineers use the where
Fanning, f F ) .
Note: There is confusion in the notation for friction factors. A = cross-sectional area of pipe or orifice, m2
The Darcy friction factor (used mainly by MEs and CEs) is four f = Moody friction factor
times the Fanning friction factor (used mainly by ChEs). Either L = length of pipe, m
one can be represented on a Moody diagram of f vs. Reynolds D = internal diameter of pipe, m
and & I D . P‘ = Pressure, N/m2 abs
p = fluid density, kg/m3
4.6 COMPRESSIBLE FLOW: VAPORS AND GASES [4]
Subscripts
Compressible fluid flow occurs between the two extremes of
isothermal and adiabatic conditions. For adiabatic flow the temper- 1 = inlet upstream condition
ature decreases (normally) for decreases in pressure, and the 2 = outlet downstream condition.
condition is represented by P‘Vk = constant, which is usually an
isentropic condition. Adiabatic flow is often assumed in short and In terms of pipe size (diameter in in./mm),
well-insulated pipe, supporting the assumption that no heat is trans-
ferred to or from the pipe contents, except for the small heat
generated by friction during flow. For isothermal conditions P’V =
constant temperature, and is the mechanism usually (not always)
assumed for most process piping design. This is in reality close to
actual conditions for many process and utility service applications.
Note: Adiabatic is best for short pipes, and approaches the (4- 12)
isothermal equations for long pipes, so adiabatic conditions are
most often assumed. where d = internal pipe diameter, in.
138 FLUID FLOW
In SI units, may lead to unacceptable conditions causing erroneous decisions
and serious effects on the sizing of automatic control valves
internal trim.
As a general guide, factors of safety of 20-30% on the friction
w,= 0.0002484 factor will accommodate the change in roughness conditions for
(fi+ 2 log, 3) steel pipe with average service of 5-10 years, but will not neces-
sarily compensate for severe corrosive conditions. Corrosion condi-
(4-13) tions should dictate the selection of the materials of construction for
the system as a part of establishing design criteria. Beyond this, the
where d = internal pipe diameter, mm. condition often remains static, but could deteriorate further. This
still does not allow for increased pressure drop due to increased flow
The derived equations are believed to apply to good plant rates. Such factors are about 10-20% additional (e.g.. increasing
design procedures with good engineering accuracy. As a matter of flow rate by 2 0 8 will increase friction loss by 40%). Therefore for
good practice with the exercise of proper judgment, the designer many applications the conservative Cameron Tables [ 7 ] give good
should familiarize himself/herself with the background of the direct-reading results for long-term service (see Table 4-46).
methods presented in order to better select the conditions associated
with a specific problem.
Note: These equations can be applied to piping including 4.9 PIPE, FITTINGS, AND VALVES
fittings. if the substitution f L / D = KRttlngs is made, in terms of
the sum of all loss coefficients for pipe plus fittings. They apply To ensure proper understanding of terminology, a brief discussion
only if the flow is choked). of the “piping” components of most process systems is appropriate.
Design conditions include: The fluids considered in this chapter consist primarily of
liquids, vapors, gases, and slurries. These are transported usually
1. Flow rate and pressure drop allowable (net driving force) estab- under pressure through circular ducts, tubes, or pipes (except for
lished, determine pipe size for a fixed length. low pressure air), and these lengths of pipe are connected by fittings
2. Flow rate, diameter, and length known, determine (net driving (screwed or threaded, butt-welded, socket-welded, or flanged) and
force) pressure drop. the flow is controlled (stopped, started, or throttled) by means of
valves fixed in these line systems. The components of these systems
Usually either of these conditions requires a trial approach based will be briefly identified in this chapter, because the calculation
upon assumed pipe sizes to meet the stated conditions. Some design methods presented are for flows through these components in a
problems may require determination of maximum flow for a fixed system. These flows always create some degree of friction loss
line size and length. (pressure drop) (or loss of pressure head) which then dictates the
Optimum economic line size is seldom realized in the average power required to move the fluids through the piping components
process plant. Unknown factors such as future flow rate allowances, (Figure 4-4). (Pump power may be required for other purposes than
actual pressure drops through certain process equipment, and so just overcoming friction.)
on, can easily overbalance any design predicated on selecting the
optimum. Certain guides as to order of magnitude of costs and
sizes can be established either by one of several correlations or by 4.10 PIPE
conventional cost-estimating methods. The latter is usually more
realistic for a given set of conditions, since generalized equations Process plants use round pipe of varying diameters (see pipe
often do not fit a plant system (Darby [ 5 ] ) . dimensions in Tables D-1, D-2, D-3 in Appendix D). Connec-
There are many computer programs for sizing fluid flow tions for smaller pipe below about 1‘h-2 in. (Figures 4-6a and b)
through pipe lines [6]. However, the designer should examine the are threaded or socket-welded, while nominal pipe sizes 2 in. and
bases and sources of such programs; otherwise, significant errors larger are generally butt- or socket-welded (Figure 4-6c) with the
could destroy the validity of the program for its intended purpose. valves and other connections flanged into the line. Steam power
plants are a notable exception. This chapter, however, does not
deal with power plant design, although steam lines are included
4.7 IMPORTANT PRESSURE LEVEL REFERENCES in the sizing techniques. Pipe is generally designated by nominal
Figure 4-3 presents a diagrammatic analysis of the important rela- size, whereas calculations for flow considerations must use the
tionships between absolute pressure, gauge pressures, and vacuum. actual standard inside diameter (ID) of the pipe. For example:
These are essential to the proper solution of fluid flow, fluid Note that OD refers to outside diameter of pipe in the table
pumping, and compression problems. Most formulas use absolute below.
pressures in calculations; however, there are a few isolated situ-
ations where gauge pressures are used. Care must be exercised
in following the proper terminology as well as in interpreting the Nominal Pipe Size in OD Inches ID Inches
meaning of data and results.
Inches Sch. 40 Sch. 80 Sch. 40 Sch. 80
4.8 FACTORS OF ”SAFETY” FOR DESIGN BASIS 1.050 1.050 0.824 0.742
3i4
Unless noted otherwise the methods suggested here do not contain 1 1.315 1.315 1.049 0.957
1’I2 1.900 1.900 1.610 1.500
any built-in design factors. These should be included, but only 2 2.374 2.375 2.067 1.939
to the extent justified by the problem at hand. Although most 3 3.500 3.500 3.068 2.900
designers place this factor on the flow rate, care must be given 4 4.500 4.500 4.026 3.826
in analyzing the actual conditions at operating rates below this
value. In some situations a large factor imposed at this point See Appendix for other sizes.
4.11 USUAL INDUSTRY PIPE SIZESAND CLASSES PRACTICE 139

-f

Coupling Reducing Half


4 4
coupling coupling

-r
R
L
Square head Hex head Round head Hex head Flush
Plug Plug Plug bushing bushing

Figure 4-6a Forged steel threaded pipe fittings, WOG (water, oil, or gas service). Note: The working pressures are always well above actual plant
operating levels. Pressure classes 3000psi and 6000psi, sizes 1/8 in. through 4in. nominal. (By permission from Ladish Co., Inc.)

90” Elbows 45” Elbows Tees Crosses

T T
4 L
Laterals Couplings Caps

Figure 4-6b Forged steel socket weld fittings, WOG (Water, oil, or gas service). Note: the working pressures are always well above actual plant
operating levels and are heavy to allow for welding. Pressure classes 3000psi and 6000psi, sizes 1/8 in. through 4in. nominal. Do not weld on malleable
iron or cast iron fittings. (By permission from Ladish Co. Inc.)

American Standards Association piping pressure Classes are 4.11 USUAL INDUSTRY PIPE SIZES AND CLASSES
given in the table below. PRACTICE
Certain nominal process and utility pipe sizes are not in common
~~

ASA Pressure Class Schedule Number of Pipe


use and hence their availability is limited. Those not usually used
5250 Ib/in.’ 40 are l/8, 11/4, 2lI2, 3112, 5 , 22, 26, 32, and 34 (in inches).
300-600 80 Some of the larger sizes, 22 in. and higher are used for special
900 120 situations. Also, some of the non-standard process sizes such as
1500 160 2’12, 3’/2, and 5 in. are used by “packaged” equipment suppliers to
2500 (I/2 x 6in.) XX (double extra strong)
connect components in their system for use in processes such as
2500 (8in. and larger) 160
refrigeration, drying, or contacting.
140 FLUID FLOW

Figure 4-6c Forged steel Welded-end fittings. (By permission form Tube Turn Technologies, Inc.)

The most common Schedule in use is 40, and it is useful for Not all Schedules are in common use, because after Sch. 40,
a wide range of pressures defined by ANSI Std. B 36.1 (American the Sch. 80 is usually sufficient to handle most pressure situa-
National Standards). Lighter wall thickness pipe would be desig- tions. The process engineer must check this Schedule for both
nated Schedules 10, 20, or 30; whereas, heavier wall pipe would pressure and corrosion to be certain there is sufficient metal wall
be Schedules 60, 80, 100, 120, 140, or 160 (see Appendix Table). thickness.
4.12 BACKGROUND INFORMATION 141
When using alloy pipe with greater tensile strength than carbon Gas or vapor specific gravity referred to air:
steel, the Schedule numbers still apply, but may vary, because it is
unnecessary to install thicker-walled alloy pipe than is necessary for
the strength and corrosion considerations. Schedules 10 and 20 are sg= MW of gas - MW of gas
- (4-15)
MW of air 29
rather common for stainless steel pipe in low pressure applications.
For example, for 3 in. nominal carbon steel pipe, the Sch. 40 (Assumes gas and air are ideal. Specific gravity for gases is usually
wall thickness is 0.216in. If the pressure required in the system referred to the density of air at standard conditions.)
needs 0.200in. wall and the corrosion rate over a 5-year life Conversion between fluid head loss in feet (meter) and (pres-
required 0.125 in. (l/xin.), then the 0.200in. +0.125 in. = 0.325in. sure drop) friction loss in psi (bar), any fluid (with specified units
and the Sch. 40 pipe would not be strong enough at the end of for p and standard value of g ) :
5 years. Often the corrosion is calculated for 10 or 15 years’ life
before replacement. Currently Sch. 80, 3in. pipe has a 0.300in.
wall thickness, even this is not good enough in carbon steel. Rather Friction loss (pressure drop), lb/in.*, AP - h, P (4- 16)
than use the much heavier Sch. 160, the designer should reconsider
‘-144
the materials of construction as well as re-examine the corrosion Friction loss (pressure drop, bar), APf = -hLP (4- 17)
data to be certain there is not unreasonable conservatism. Perhaps 10,200
stainless steel pipe or a “lined’ pipe would give adequate strength
and corrosion resistance. For a bad corrosion condition, lined pipe For water, psi, APf = h, (ft) SP gr (4- 18)
2.31 (ft/psi)
using linings of PVC (polyvinyl chloride), Teflon@,or Saran@typi-
cally as shown in Figures 4-7a-d can be helpful.
While threaded pipe is joined by threaded fittings Equivalent diameter and hydraulic radius for non-circular flow
(Figure 4-6a), the joints of welded pipe are connected to each other ducts or pipes
by butt or socket welding (Figure 4-6b) and to valves by socket
welds or flanges of several types (Figure 4-8) using a gasket of R, =hydraulic radius, ft
composition material, rubber, or metal at the joint to seal against cross-sectional flow area for fluid flow
leaks. The joint is pulled tight by bolts (Figure 4-9). R, = (4- 19)
wetted perimeter for fluid flow
For lower pressure systems of approximately 15Opsig at
400” F or 225 psig at 100”F, and where sanitary precautions (food D, =hydraulic diameter (equivalent diameter), ft
products or chemicals used in food products) or some corrosion
resistance is necessary, tubing is used. It is joined together by
(4-20)
butt welds (Figure 4-10) or special compression or hub-type end
connectors. This style of “piping” is not too common in the chem-
icallpetrochemical industries, except for instrument lines (sensing, d, = hydraulic diameter (equivalent diameter), in.
signal transmission) or high pressures above 2000 psig.
Figure 4-1 1 compares the measurement differences for tubes d, = 48RH,in (4-21)
(outside diameter) and iron or steel pipe size (IPS), nominal inside 4 (cross-sectional area for fluid flow)
diameter. One example of dimensional comparison for IPS pipe d, = (4-22)
for Schs 5 and 10 are compared to one standard scale of tubing in (wetted perimeter for fluid flow)
Table 4-1. The tubing conforms to ANSUASTM A-403-78 Class
CR (stainless) or MSS Manufacturers Standard Society SP-43, For the narrow shapes with small depth relative to width (length),
Sch. 5 s . the hydraulic radius is approximately [4]:

R, = (depth of passage) (4-23)


TOTAL LINE PRESSURE DROP
The total piping system friction loss (pressure drop) for a particular For those non-standard or full circular configurations of flow, use
pipe installation is the sum of the friction loss (drop) in pipe, valves, d equivalent to actual flow area diameter, and D equivalent to 4RH.
and fittings, plus other pressure losses (drops) through control
valves, plus drop through equipment in the system, plus static drop cross-sectional area available for fluid flow, of duct
due to elevation or pressure level. For example, see Figure 4-4.
d=4( wetted perimeter of duct
This total pressure loss is not necessarily required in determining
the frictional losses in the system (Note: Pressure loss is not the (4-24)
same as pressure change). It is necessary when establishing gravity
flow or the pumping head requirements for a complete system. This also applies to circular pipes or ducts and oval and rectan-
Design practice breaks the overall problem into small compo- gular ducts not flowing full. The equivalent diameter is used in
nent parts which allow for simple analysis and solution. This is the determining the Reynolds number for these cases.
recommended approach for selection and sizing of process piping. Minimum size of pipe is sometimes dictated by structural
considerations, that is, ll/~-in.Sch. 40 steel pipe is considered the
smallest size to span a 15 ft-20 ft pipe rack without intermediate
4.12 BACKGROUND INFORMATION (ALSO S E E
support.
CHAPTER 5)
Gravity flow lines are often set at 11/4-2in. minimum, disre-
Gas or vapor density following perfect gas law: garding any smaller calculated size as a potential source of trouble.
Pump suction lines are designed for about a one footlsecond
p = 144 P’ ( T )
(K>
- ,w t 3 (4-14)
velocity, unless a higher velocity is necessary to keep small solids
or precipitates in suspension. Suction line sizes should be larger
142 FLUID FLOW

S E C T I O N H-H

Figure 4-7a Lined-steel pipe and fittings for corrosive service. (By permission from Performance Plastics Products.)

1 " . 6 " Sch 80


8''. Sch 40 FLANGE 3

FLANGE 2 /

Figure 4-7b Lined-steel pipe flanged sparger for corrosive service. (By permission from Performance Plastics Products.)

Connection of reinhmd flared


face to gasketedplasibllined pipe Wlh taper reducing spacer2
2 Only the following size
reductions should be made by
this technique when connecting
pipe with molded raised laces:
1Mx1, 2x1, 2xlM, 2hxlM,
2'/x2, 3x2, 3x2%,4x2M,
4x3, 6x4, 8x6. All other
reductions require use of
reducing liller flanges or
concentric reducers.

~ COLLAR
REDUCING SPACER

Figure 4-7c Flanged line-steel pipe fittings for corrosive service. (By permission from Dow Plastic-Lined Products, Bay City, Mich. 48707, 1-800-233-
7577.)
4.13 REYNOLDS NUMBER, Re (SOMETIMES USED N,) 143
1 to 5psi per 100 equivalent feet of pipe. The Appendix presents
useful carbon steel and stainless steel pipe data.

4.13 REYNOLDS NUMBER, Re (SOMETIMES


USED NRe)
This is the basis for establishing the condition or type of fluid
flow (flow regime) in a pipe. Reynolds numbers below 2000-2100
correspond to (are usually considered to define) laminar or viscous
flow; numbers from 2000 to 3000-4000 correspond to (define) a
transition region of peculiar flow, and numbers above 4000 corre-
spond to (define a state of) turbulent flow. Reference to Figure 4-5
and Figure 4-13 will identify these regions, and the friction factors
associated with them [2].

DvP = 123.9-d W = 6.31- W


Re = -
(4-26)
Pe P dP
where
d = pipe internal diameter, in.
D = pipe internal diameter, ft
u = mean fluid velocity, ftls
W = fluid flow rate, lbih
Figure 4-7d Lined plug valve for corrosive service. (By permission from P = fluid density, 1b/ft3
Dow Plastic-Lined Products, Bay City, Mich. 48707, 1-800-233-7577.) P L=
~ absolute viscosity, Ibm/ft
p = absolute viscosity, cP.

4; s g
than discharge sizes to minimize the friction loss to reach a realistic R e = 22,700- 9P = 50.6-QP = 0.482 - (4-27)
dP dP dP
value of available Net Positive Suction Head (NPSH,) for the
given system. where
As a general guide, for pipe sizes use: q = fluid flow rate, ft3/s
threaded pipe - up to and including 1'/I in. or 2 in. nominal q;1 = fluid flow rate, ft3/h
welded or flanged pipe - 2 in. and larger. Q = fluid flow rate, gpm
S, = specific gravity of a gas relative to air.
Situations may dictate deviations. Never use cast iron fittings
or pipe in process situations unless there is only gravity pres- In SI units,
sure head (or not over 1Opsig) or the fluid is non-hazardous. dvp
Dvp = - W
Re = - = 354- (4-28)
One exception is in some concentrated sulfuric acid applications, P' P dP
where extreme caution must be exercised in the design of the
safety of the system area. Never use in pulsing or shock service. d = pipe internal dismeter, mm
Never use malleable iron fittings or pipe unless the fluid is non- D = pipe internal diameter, m
= mean fluid vel city, m / s
hazardous and the pressure not greater than 25psig. Always use
a pressure rating at least four times that of the maximum system
u
W = fluid flow rate, kgih
q'
p = fluid density, k /m3
pressure. Also, never use cast iron or malleable iron fittings or
valves in pressure-pulsating systems or systems subject to physical p' = absolute viscosity, Pasf
p = absolute viscos'lty, CP
shock.
Use forged steel fittings for process applications as long as ICP = I O - ~ P ~ S 'I
1 Pas = 1 Ns/m2 = 1 kgJms = lo3cP.
the fluid does not create a serious corrosion problem. These fittings
are attached to steel pipe andor each other by threading, socket 4Pl QP 4; s,
welding, or direct welding to steel pipe. For couplings attached R e = 1,273,000 - = 21.22- = 432- (4-29)
by welding to pipe, Figure 4-6b, use either 2000psi or 6000psi
rating to give adequate area for welding without distortion, even where
dPl dP dP

though the process system may be significantly lower (even atmo- q = fluid flow rate, m31s
spheric). Branch connections are often attached to steel pipe using q; = fluid flow rate, m3/h
Weldolets@ or Threadole@ (Figure 4-12). Q = fluid flow rate, I/mFn
Note: @ = Registered Bonney Forge, Allentown, PA.
S, = specific gravity of h gas relative to air.
Mean pressure in a gas line [58].
The mean velocity of flowing liquid can be determined by the
following:
P (mean or average) = In English Engineering units,

This applies particularly to long flow lines. The usual economic v = 183.3 q/d' 70.408 Q/d' = 0.0509 W / d 2 p ,ft/s
range for pressure loss due to liquid flow: (a) Suction piping
(4-30)
to 11/4psi per 100 equivalent feet of pipe; (b) Discharge piping
144 FLUID FLOW

Welding neck flanges are distinguished from other types by their long tapered hub and gen-
tle transition of thickness in the region of the butt weld joining them to the pipe. Thus this type
of flange is preferred for every severe service condition, whether this results from high pres-
sure or from sub-zero or elevated temperature, and whether loading conditions are substan-
tially constant or fluctuate between wide limits.

Slip-on flanges continue to be preferred to welding neck flanges by many users on account
of their initially lower cost, the reduced accuracy required in cutting the pipe to length, and the
somewhat greater ease of alignment of the assembly; however, their final installed cost is
probably not much, if any, less than that of welding neck flanges. Their calculated strength
under internal pressure is of the order of two-thirds that of welding neck flanges, and their life
under fatigue is about one-third that of the latter.

Lap joint flanges are primarily employed with lap joint stubs, the combined initial cost of the
two items being approximately one-third higher than that of comparable welding neck flanges.
Their pressure-holding ability is little, if any, better than that of slip-on flanges and the fatigue
life of the assembly is only one-tenth that of welding neck flanges. The chief use of lap joint
flanges in carbon or low alloy steel piping systems is in services necessitating frequent dis-
mantling for inspection and cleaning and where the ability to swivel flanges and to align bolt
holes materially simplifies the erection of large diameter or unusually stiff piping. Their use at
points where severe bending stress occurs should be avoided.

Threaded flanges made of steel, are confined to special applications. Their chief merit lies in
the fact that they can be assembled without welding; this explains their use in extremely high
pressure services, particularly at or near atmospheric temperature, where alloy steel is essen-
tial for strength and where the necessary post-weld heat treatment is impractical. Threaded
flanges are unsuited for conditions involving temperature or bending stresses of any magni-
tude, particularly under cyclic conditions, where leakage through the threads may occur in rel-
atively few cycles of heating or stress; seal welding is sometimes employed to overcome this,
but cannot be considered as entirely satisfactory.

Socket welding flanges were initially developed for use on small-size high pressure piping.
Their initial cost is about 10% greater than that of slip-on flanges; when provided with an inter-
nal weld as illustrated, their static strength is equal to, but their fatigue strength 50% greater
than double-welded slip-on flanges. Smooth, pocketless bore conditions can readily be
attained (by grinding the internal weld) without having to bevel the flange face and, after weld-
ing, to reface the flange as would be required with slip-on flanges.

Figure 4-8 Forged steel companion flanges to be attached to steel pipe by the methods indicated. (By permission from Tube Turn Technologies, Inc.)
4.13 REYNOLDS NUMBER, Re (SOMETIMES USED NRJ 145

Orifice flanges are widely used in conjunction with orifice meters for measuring the rate of
flow of liquids and gases. They are basically the same as standard welding neck, slip-on and
screwed flanges except for the provision of radial, tapped holes in the flange ring for meter
connections and additional bolts to act as jack screws to facilitate separating the flanges for
inspection or replacement of the orifice plate.

Blind flanges are used to blank off the ends of piping, valves and pressure vessel openings.
From the standpoint of internal pressure and bolt loading, blind flanges, particularly in the larg-
er sizes, are the most highly stressed of all American Standard flange types; however, since
the maximum stresses in a blind flange are bending stresses at the center, they can safely be
permitted to be higher than in other types of flanges.

1. In Tube Turns tests of all types of flanged assemblies, fatigue failure invariable occurrred in the pipe or in an unusually weak weld, never in the flange proper.
The type of flange, however, and particularly the method of attachment, greatly influence the number of cycles required to cause fracture.
2. ANSI B16.5-1961-Steel Pipe Flanges and Flanged Fittings.
3. ASME Boiler and Pressure Vessel Code 1966, Section I, Par. P-300.

Figure 4-8-(continued)

x=
Alternates

Male to Mole Male to Female


Flanged Joint Flanged Joint

Raiaad Face Tongue B Gruve Joint A88omblrd Ring Joint


(Rat gasket) (uses Rat gasket)

Figure 4-9 Most common flange connection joints. Cross section of a pair of flanges with bolts to draw joint tight.

In SI units, in the order of their appearance are D =inside diameter of pipe,


ft; u = liquid velocity, ft/s; p =liquid density, lb/ft3, p =absolute
v = 1273.2 x lo3 q / d 2 =21.22 Q/d' viscosity of liquid, Ib,/fts; d =inside diameter of pipe, in.; k = z/S
= 353.1 W / d 2 p ,m/s (4-31) liquid flow rate, lb/h; B =liquid flow rate, bbl/h; k =kinematic
viscosity of the liquid, Cst; q =liquid flow rate, ft3/s; Q =liquid
Table 4-2 gives a quick summary of various ways in which the flow rate, ft3/Illjn. Use Table 4-2 to find the Reynolds number of
Reynolds number can be expressed. The symbols in Table 4-2, any liquid flowing through a pipe.
146 FLUID FLOW

Figure 4-10 Light weight stainless steel butt-weld fittingshbing for low pressure applications. (By permission from Tri-Clover, Inc.)

4.14 PIPE RELATIVE ROUGHNESS such as 10, 15, or 20 years in service. Usually a IO-15-year
life period is a reasonable expectation. It is not wise to expect
Pipe internal roughness reflects the results of Pipe mmufacture Or smooth internal conditions over an extended life, even for water.
Process corrosion, or both. In designing a flow system, recogni- air, or oil flow because some actual changes can occur in the
tion must be given to (a) the initial internal pipe condition as well internal surface condition. Some fluids are much worse in this
as (b) the expected condition after some reasonable life period, regard than others. New, clean steel pipe can be adjusted from
4.15 DARCY FRICTION FACTOR, F 147

TUBE O D IPS
--- -
I/ f
,

INDICATES SIZE INDICATES


NOMINAL
DIAMETER DIAMETER

How Tube OD Differs from IPS


Dependable performance; fast, easy installation In Tube OD the size specified indicates i t s outside diameter
Uniformity of wall thickness and geometric accuracy of . . . whereas in Iron Pipe Size (IPS), the size has reference
ends permit precise alignment of joints. to a nominal diameter. See Table 2-1.

Figure 4-11 Dimension comparison of tubing and IPS (iron pipe size) steel piping. (By permission from Tri-Clover, Inc.)

TABLE 4-1 Comparison of Dimensions and Flow Area for Tubing and Iron Pipe Size (IPS) Steel Pipe

OD Tubing IPS Pipe


-___
Schedule 5s Schedule 10s
Od Tubing Outside Inside Flow Area IPS Pipe Outside Inside Flow Area Inside Flow Area
Size Diameter Diameter (in.z) Size Diameter Diameter (in.2) Diameter (in.2)

314 0.750 0.625 0.307 313 1.050 0.920 0.665 0.884 0.614
1 1.000 0.870 0.595 1 1.315 1.185 1.10 1.097 0.945
1’h 1.500 1.370 1.47 1’I2 1.goo 1.770 2.46 1.682 2.22
2 2.000 1.870 2.75 2 2.375 2.245 3.96 2.157 3.65
2 ‘I2 2.500 2.370 4.41 2 ’I2 2.875 2.709 5.76 2.635 5.45
3 3.000 2.843 6.31 3 3.500 3.334 8.73 3.260 8.35
3’12 - - - 3’12 4.000 3.834 11.55 3.760 11.10
4 4.000 3.834 11.55 4 4.500 4.334 14.75 4.260 14.25
6 6.000 5.782 26.26 6 6.625 6.407 32.24 6.357 31.75
8 8.000 7.782 47.56 8 8.625 8.407 55.5 8.329 54.5
10 10.000 9.732 74.4 10 10.750 10.482 86.3 10.420 85.3
12 12.000 11.732 108 12 12.750 12.438 121.0 12.390 120.0

(Source: By permission from Tric-Clover, Inc.)

the initial clean condition to some situation allowing for the addi- 4.15 DARCY FRICTION FACTOR, F
tional roughness. The design-roughened condition can be interpo-
For laminar or viscous flow,
lated from Figure 4-13 to achieve a somewhat more roughened
condition, with the corresponding relative roughness E / D value.
Table 4-3 shows the wall roughness of some clean, new pipe f = -64 For Re i 2000 (4-32)
Re
materials.
Note that the E / D factor from Figure 4-13 is used directly in For transition and turbulent flow, use Figure 4-13 (the f in this
Figure 4-5. As an example that is only applicable in the range of figure only applies for fully turbulent flow corresponding to the
the charts used, a 10% increase in E / D to account for increased flat portions of the curves in Figure 4-5) with Figure 4-5, and
roughness yields from Figure 4-5 an f of only 1.2% greater than a Figures 4-14a and 4-14b as appropriate. Friction factor in long
clean, new commercial condition pipe (Note: This number depends steel pipes handling wet (saturated with water vapor) gases such
on where you are on the Moody diagram. Figure 4-13 applies as hydrogen, carbon monoxide, carbon dioxide, nitrogen, oxygen,
only for fully turbulent flow to very high Reynolds numbers). and similar materials should be considered carefully, and often
Generally the accuracy of reading the charts does not account for increased by a factor of 1.2-2.0 to account for corrosion.
large fluctuations in f values. Of course, f can be calculated as Important Note: The Moody [3] friction factors (fD) repro-
discussed earlier, and a more precise number can be achieved, duced in this text (Figure 4-5) are consistent with the published
but this may not mean a significantly greater accuracy of the values of references [2-4], and cannot be used with the values
calculated pressure drop. Generally, for industrial process design, presented in Perry’s Handbook [ 11 (e.&., the Fanning friction
experience should be used where available in adjusting the rough- factor, f F ) , as Perry’s values for fF are one-fourth times the
ness and effects on the friction factor. Some designers increase values cited in this chapter (e.g., f F = ifD).
It is essential to
the friction factor by 10-15% over standard commercial pipe use f values with the corresponding formulas offered in the
values. appropriate text.
Next Page

148 FLUID FLOW

Figure 4-12 Branch connections for welding openings into steel pipe. See Figure 4-6c for alternate welding fittings. (By permission from Bonney Forge
Crop., Allentown. PA.)

The Colebrook equation [9] is considered a reliable approach Note: The turbulent portion of the published Moody diagram is
to determining the friction factor, f D (Moody factor). actually a plot of this equation, which is derived to fit “sand rough-
ness” data in pipes [IO].

&
1
--
3.70
”)
+ Re& For Re > 4000 (4-33) Re=
VD
-
V‘

You might also like