DK1915 CH10
DK1915 CH10
DK1915 CH10
Hydrostatic Bearings
10.1 INTRODUCTION
The design concept for hydrostatic bearings is the generation of a high-pressure
fluid film by using an external pump. The hydrostatic system of a journal bearing
is shown in Fig. 1-4. The fluid is fed from a pump into several recesses around the
bore of the bearing. From the recesses, the fluid flows out through a thin
clearance, h0 , between the journal and bearing surfaces, at the lands outside
the recesses. Previous literature on hydrostatic bearings includes Opitz (1967),
Rowe (1989), Bassani and Picicigallo (1992), and Decker and Shapiro (1968).
The fluid film in the clearance separates the two surfaces of the journal and
the bearing and thus reduces significantly the friction and wear. At the same time,
the thin clearance at the land forms a resistance to the outlet flow from each
recess. This flow resistance is essential for maintaining high pressure in the
recess. The hydrodynamic load capacity that carries the external load is the
resultant force of the pressure around the bearing.
There is also a fluid film in a hydrodynamic bearing. However, unlike the
hydrodynamic bearing, where the pressure wave is generated by the hydrody-
namic action of the rotation of the journal, hydrostatic bearing pressure is
generated by an external pump.
There are certain designs of hydrodynamic bearings where the oil is also
supplied under pressure from an external oil pump. However, the difference is
that in hydrostatic bearings the design entails recesses, and the operation does not
depend on the rotation of the journal for generating the pressure wave that
Solution
In Sec. 4.9, Example Problem 4-3 is solved for the pressure gradient in a parallel
flow between two stationary parallel plates. The pressure gradient for a constant
clearance, h0 , and rate of flow Q is
dp 12m
¼ 3Q ð10-1Þ
dx bh0
The flow is in the direction of x, and b is the width (perpendicular to the flow
direction). In a circular pad, there are radial flow lines, having radial symmetry.
Pressure Distribution
In order to solve the present problem, the flow is considered in a thin ring, of
radial thickness dr, as shown in Fig. 10-1. Since dr is small in comparison to the
radius, r, the curvature can be disregarded. In that case, the flow along dr is
assumed to be equal to a unidirectional flow between parallel plates of width
b ¼ 2pr. The pressure gradient dp is derived from Eq. (10-1):
dp 12mQ 12mQ dr
¼ or dp ¼ ð10-2Þ
dr 2prh30 2ph30 r
Integrating of Eq. (10-2) yields
6mQ
p¼ ln r þ C ð10-3Þ
ph30
The constant of integration, C, is determined by the following boundary
condition:
p¼0 at r¼R ð10-4Þ
After solving for C, the expression for the radial pressure distribution in the radial
clearance as a function of r is
6mQ R
p¼ ln ð10-5Þ
ph30 r
The expression for the pressure at the recess, pr , at r ¼ R0 is
6mQ R
pr ¼ ln ð10-6Þ
ph30 R0
Load Capacity
The load capacity is the integration of the pressure over the complete area
according to the following equation:
ð
W ¼ pdA ð10-7Þ
ðAÞ
For the total load capacity, the pressure is integrated in the recess and in the thin
clearance (land) according to the following equation:
ð R0 ðR
W ¼ pr ð2pr drÞ þ pð2prdrÞ ð10-10Þ
0 R0
After substituting the pressure equation into Eq. (10-10) and integrating, the
following expression for the load capacity of a circular hydrostatic pad is
obtained:
p R2 R20
W ¼ p ð10-11
2 lnðR=R0 Þ r
Equation (10-11) can be rearranged as a function of the recess ratio, R0 =R, and
the expression for load capacity of a hydrostatic pad is
p R4 R40
Tf ¼ m 1 4 o ð10-14Þ
2 h0 R
E_ h Q pr ð10-16Þ
Here, Q is the flow rate through the pad and pr is the recess pressure. The total
power consumption E_ t is the sum of the power of the drive for turning the bearing
and the power of the pump for circulating the fluid through the bearing resistance.
The following equation is for net power consumption. In fact, the pump and
motor have power losses, and their efficiency should be considered for the
calculation of the actual power consumption:
4
4
_EtðnetÞ ¼ Q pr þ p m R 1 R0 o2 ð10-17Þ
2 h0 R4
Substituting the value of Q and dividing by the efficiency of the motor and drive
Z1 and the efficiency of the pump Z2 , the following equation is obtained for the
total power consumption (in the form of electricity consumed by the hydrostatic
system) for the operation of a bearing:
1 1 ph30 1 p R4 R4
E_ t ¼ p2r þ m 1 04 o2 ð10-18Þ
Z2 6 m lnðR=R0 Þ Z1 2 h0 R
The efficiency Z2 of a gear pump is typically low, about 0.6–0.7. The motor-drive
system has a higher efficiency Z1 of about 0.8–0.9.
pr ¼ 58;824 N=m2
1 1 p
C1 ¼ ð58;824Þ2 ¼ 4:35 1011 N =s-m2
0:6 6 0:01 lnð1=0:5Þ
In a similar way, the second constant, C2 , which is associated with the motor, is
calculated from Eq. (10-20):
1 p
C2 ¼ ð0:01Þð0:14 Þð1 0:54 Þ523:62 ¼ 0:448 N-m2 =s
0:9 2
Substituting these values of C1 and C2 into Eq. (10-19), the total power as a
function of the clearance becomes
0:448
E_ t ¼ ð4:35 1011 Þh30 þ
h0
In this equation, the power of the pump for circulating the fluid is the first term,
which is proportional to h30 , while the second term, which is proportional to h1 0 ,
is the power of the motor for rotating the disk.
The powers of the pump and of the drive are the two power components
required to maintain the operation of the hydrostatic bearing. In Fig. 10-2, the
curves of the hydraulic power and mechanical power are plotted as a function of
the clearance, h0 . The curve of the mechanical power, E_ f , that is provided by the
motor points down with increasing clearance, h0 . The power, E_ f , is for rotating
one disk relative to the other and overcoming the viscous friction in the clearance
between the two disks. The second curve is of the hydraulic power, E_ h , which is
provided by the pump to maintain hydrostatic pressure in the recess. This power
is rising with increasing clearance, h0 . The hydraulic power, E_ h , is for overcoming
the flow resistance in the thin clearance at the outlet from the recess. The total
power, E_ t , is the sum of these curves. For the hydrostatic pad in this problem, the
optimal point (minimum power) is for a clearance of about 0.75 mm, and the total
power consumption of the bearing in this problem is below 0.8 kW.
The result of 0.8 kW is too high for a hydrostatic bearing. It is possible to
reduce the power consumption by using lower-viscosity oil.
Here, B b is the clearance (land) dimension along the direction of flow and l is
the dimension of clearance (land) normal to the flow direction.
3
h0
Q/ p ð10-22Þ
m r
A vertical displacement, Dh, of the slide will increase and decrease the clearance
h0 at the lands of the opposing hydrostatic pads. For constant flow rate Q and
viscosity, Eq. (10-22) indicates that increase and decrease in the clearance h0
would result in decrease and increase, respectively, of the recess pressure (the
recess pressure is inversely proportional to h30 ). High stiffness means that only a
very small vertical displacement of the slide is sufficient to generate a large
difference of pressure between opposing recesses. The force resulting from these
pressure differences acts in the direction opposite to any occasional additional
load on the thrust bearing.
Theoretically, the bearing stiffness can be very high for a hydrostatic pad
with a constant flow rate to each recess; but in practice, the stiffness is limited by
the hydraulic power of the motor and its maximum flow rate and pressure. This
theoretical explanation is limited in practice because there is a maximum limit to
the recess pressure, pr . The hydraulic power of the pump and the strength of the
complete system limit the recess pressure. A safety relief valve is installed to
protect the system from exceeding its allowable maximum pressure. In addition,
the fluid viscosity, m, is not completely constant. When the clearance, h0 , reduces,
dW
k¼ ð10-23Þ
dh0
The meaning of the negative sign is that the load increases with a reduction of the
clearance. High stiffness is particularly important in machine tools where any
displacement of the slide or spindle during machining would result in machining
W
Ae pr ¼ W or Ae ¼ ð10-25Þ
pr
For a constant flow rate, the load capacity, in terms of the effective area and
bearing resistance, is
W ¼ Ae Rc Q ð10-26Þ
Comparison with the equations for the circular pad indicates that the resistance is
proportional to h3
0 or
1
Rc ¼ kh3
0 ; and W ¼ kAe Q ð10-27Þ
h30
Here, k is a constant that depends on bearing geometry, flow rate, and fluid
viscosity
dW 1
Stiffness k ¼ ¼K 4 where K ¼ 3kAe Q ð10-28aÞ
dh0 h0
dW 1
Stiffness k ¼ ¼ 3kAe Q 4 ð10-28bÞ
dh0 h0
Equation (10-28b) indicates that stiffness increases very fast with reduction in the
bearing clearance. This equation can be applied as long as the flow rate Q to the
recess is constant. As discussed earlier, deviation from this can occur in practice if
the pressure limit is reached and the relief valve of the hydraulic system is
opened. In that case, the flow rate is no longer constant.
Equation (10-28a) can be used for any hydrostatic bearing, after the value
of K is determined. For a circular pad:
9mQðR2 R20 Þ
k¼ ð10-30Þ
h40
Solution
Given:
W ¼ 5000 N
R ¼ 0:1 m
R0 ¼ 0:05 m
m ¼ 0:005 N-s=m2
h0 ¼ 120 mm
a. Recess Pressure
In order to solve for the flow rate, the first step is to determine the recess pressure.
The recess pressure is calculated from Eq. (10-12) for the load capacity:
2 p 1 R20 =R2
W ¼R pr
2 lnðR=R0 Þ
b. Flow Rate
The flow rate Q can now be determined from the recess pressure. It is derived
from Eq. (10-13):
!
p h30 1 pð120 106 Þ3
Q¼ p; Q¼ 294;120
6m lnðR=R0 Þ r 6 0:005 lnð0:1=0:05Þ
d. Bearing Stiffness
Finally, the stiffness of the circular pad fed by a constant flow rate can be
determined from Eq. (10-30):
9mQðR2 R20 Þ
k¼
h40
Substituting the values in this stiffness equation yields
9 0:005 76:8 106 ð0:12 0:052 Þ
k¼
ð120 106 Þ4
6
k ¼ 125 10 N=m
This result indicates that the stiffness of a constant-flow-rate pad is quite
high. This stiffness is high in comparison to other bearings, such as hydro-
Solution
a. Flow rates Q1 and Q2
Given that h1 ¼ h2 ¼ 0:2 mm, the flow rate Q can be obtained via Eq. (10-13):
1 ph30
Q¼ p
6 m lnðR=R0 Þ r
The first step is to find pr by using the load capacity equation (for a top pad).
Substituting the known values, the recess pressure is the only unknown:
p 1 ð0:0252 =0:052 Þ
5000 ¼ 0:052 pr1
2 lnð0:05=0:025Þ
Substituting this recess pressure in Eq. (10-13), the following flow rate, Q1 , is
obtained:
1 p 0:00023
Q1 ¼ 1:176 106 ¼ 7:1 104 m3 =s
6 0:01 lnð0:05=0:025Þ
The second step is to find pr2 by using the load capacity equation (for the bottom
pad), substituting the known values; the following equation is obtained, with Pr2
as unknown:
p 1 ð0:0252 =0:052 Þ
10;000 ¼ 0:052 pr2
2 lnð0:05=0:025Þ
Substituting the known values, in Eq. (10-13) the flow rate Q2 is:
1 pð0:00023 Þ
Q2 ¼ 2:352 106 ¼ 14:2 104 m3 =s
6 0:01 lnð0:05=0:025Þ
p h30
Q¼ p
6m lnðR=R0 Þ r
p h31 p h32
Q¼ pr1 ¼ p
6m ln R=R0 Þ 6m lnðR=R0 Þ r2
h1 þ h2 ¼ 0:0004 m
Substituting yields
1 ph31 1 pð0:0004 h1 Þ3
1:176 106 ¼ 2:352 106
6 0:01 ln 2 6 0:01 ln 2
9mQðR2 R20 Þ
k¼
h40
9 ð0:01Þð0:00142ðð0:052 0:0252 Þ
k ðlower padÞ ¼ ¼ 149:76 106 N=m
0:0002
The total bidirectional bearing stiffness is obtained by adding the top and bottom
stiffnesses, as follows:
DW
k¼ where DW ¼ 30 N
Dh
DW
Dh ¼
k
30 N
Dh ¼ ¼ 1:33 107 m ¼ 0:133 mm
224:32 106
This example shows that under extra force, the displacement is very small.
Dp
Rf ¼ ð10-31Þ
Q
For a given resistance, the flow rate is determined by the pressure difference:
Dp
Q¼ ð10-32Þ
Rf
The resistance of the inlet flow restrictor is Rin , and the resistance to outlet flow
through the bearing clearance is Rc (resistance at the clearance). The pressure at
the recess, pr , is bounded between the inlet and outlet resistances; see a schematic
representation in Fig. 10-6. The supply pressure, ps , is constant; therefore, any
change in the inlet or outlet resistance would affect the recess pressure.
From Eq. (10-32), the flow rate into the recess is
ps pr
Q¼ ð10-33Þ
Rin
The fluid is incompressible, and the flow rate Q is equal through the inlet
resistance and clearance resistance (continuity). Equating the preceding two flow
rate expressions yields
pr ps pr
¼ ð10-35Þ
Rc Rin
The recess pressure is solved for as a function of the supply pressure and
resistances:
1
pr ¼ p ð10-36Þ
ð1 þ Rin =Rc Þ s
In terms of the supply pressure and the effective pad area, the load capacity is
1
W ¼ Ae p ð10-38Þ
ð1 þ Rin =Rc Þ s
The inlet resistance of laminar flow through a capillary tube is constant, and the
pressure drop is proportional to the flow rate. However, the flow resistance
Rc ¼ kh3
0 ð10-39Þ
Here, k is a constant that depends on the pad geometry and fluid viscosity.
Equation (10-38) can be written in the form,
1
W ¼ Ae p ð10-40Þ
ð1 þ K1 h30 Þ s
where K is defined as
Rin
K1 ¼ ð10-41Þ
k
In Eq. (10-40) for the load capacity, all the terms are constant except the
clearance thickness. Let us recall that the expression for the stiffness is
dW
Stiffness k ¼ ð10-42Þ
dh0
dW 3K1 h20
Stiffness k ¼ ¼ Ae ps ð10-43Þ
dh0 ð1 þ K1 h30 Þ2
64mlc
Dp ¼ Q ð10-44Þ
pdi4
Here, di is the inside diameter of the tube and lc is the tube length. The inlet
resistance by a capillary tube is
64mlc
Rin ¼ ð10-45Þ
pdi4
For calculating the pad stiffness in Eq. (10-43), the inlet resistance is calculated
from Eq. (10-45), and the value of k is determined from the pad equations.
Equations (10-43) and (10-46) yield a simplified expression for the stiffness as a
function of b:
dW 3
k¼ ¼ Ae ðb b2 Þps ð10-47Þ
dh0 h0
For maximum stiffness, the supply pressure should be twice the recess pressure.
This can be obtained if the inlet resistance were equal to the recess resistance.
This requirement will double the power of the pump that is required to overcome
viscous friction losses. The conclusion is that the requirement for high stiffness in
constant-supply-pressure systems would considerably increase the friction losses
and the cost of power for operating the hydrostatic bearings.
b. Flow Rate
The flow rate Q can now be determined. It is derived from the following
expression [see Eq. (10-13)] for Q as a function of the clearance pressure:
p h30
Q¼ p
6m lnðR=R0 Þ r
Similar to Example Problem 10-3, after substituting the values, the flow rate is
Q ¼ 76:8 106 m3 =s
d. Pad Stiffness
Supply Pressure: Now the supply pressure can be solved for as well as the
e. Stiffness Comparison
In comparison, for a constant flow rate (see Example Problem 10-3) the stiffness
is
k ¼ 125 106 N=m ðfor constant flow rate
For the bearing with a constant pressure supply in this problem, the stiffness is
about half of the constant-flow-rate pad in Example Problem 10-3.
f. Hydraulic Power
The power for circulating the oil through the bearing for constant pressure supply
is twice of that for constant flow rate. Neglecting the friction losses in the pipes,
the equation for the net hydraulic power for circulating the oil through the bearing
in a constant-flow-rate pad is
E_ h Q pr ðfor constant-flow-rate padÞ
¼ 76:8 106 294:18 103 ¼ 22:6 W ðconstant-flow-rate padÞ:
In comparison, the equation for the net hydraulic power for constant pressure
supply is
E_ h Q ps ðFor a constant pressure supply padÞ:
Since ps ¼ 2pr , the hydraulic power is double for constant pressure supply:
E_ h Q pr ¼ 76:8 106 588:36 103 ¼ 45:20 W
ðfor a constant-pressure-supply padÞ
Solution
a. Recess Pressure and Supply Pressure
The recess pressure is derived from the equation of the load capacity:
p 1 ðR0 =RÞ2
W ¼ R2 pr
2 lnðR=R0 Þ
After substitution, the recess pressure is the only unknown in the following
equation:
p 1 ð0:035=0:07Þ2
21;000 N ¼ 0:072 p
2 lnð0:07=0:035Þ r
pr ¼ 2:52 MPa
10.12 APPLICATIONS
An interesting application is the hydrostatic pad in machine tool screw drives
(Rumberger and Wertwijn, 1968). For high-precision applications, it is important
to prevent direct metal contact, which results in stick-slip friction and limits the
machining precision. Figure 10-7 shows a noncontact design that includes
hydrostatic pads for complete separation of the sliding surfaces of screw drive.
Another important application is in a friction testing machine, which will be
described in Chapter 14.
1. It is a simple and compact pump, and does not need inlet and outlet
valves, such as in the piston pump. However, gear pumps require close
running clearances.
2. It involves continuous flow (unlike positive-displacement reciprocating
pumps).
Hp ¼ Hd Hs ð10-50Þ
Pump head units are of length (m, ft). Head is calculated from the Bernoulli
equation. The expression for discharge and suction heads are:
pd Vd2
Hd ¼ þ þ Zd ð10-51Þ
g 2g
ps Vs2
Hs ¼ þ þ Zs ð10-52Þ
g 2g
where
The pump head, Hp , is the difference between the discharge head and
suction head. In a closed loop, Hp is equal to the head loss in the loop. The
expression for the pump head is
pd ps Vd2 Vs2
Hp ¼ þ þ ðZd Zs Þ ð10-53Þ
g 2g
The velocity of the fluid in the discharge and suction can be determined
from the rate of flow and the inside diameter of the pipes. In most gear pumps, the
pipe inside the diameters on the discharge and suction sides are equal. In turn, the
discharge velocity is equal to that of the suction. Also, there is no significant
difference in height between the discharge and suction. In such cases, the last two
E_ h ¼ QgHp ð10-55Þ
E_ h
Z¼ ð10-56Þ
E_ b
about 12–20 RPM. In most cases, the low speed of rotation is not sufficient for
adequate fluid film thickness in hydrodynamic bearings.
There has been continual trend to increase the diameter, D, of rotary mills,
because milling output is proportional approximately to D2:7 and only linearly
proportional to the length. In general, large rotary mills are more economical for
the large-scale production of ores. Therefore, the outside diameter of a rotary
mill, D, is quite large; many designs are of about 5-m diameter, and some rotary
mills are as large as 10 m in diameter.
Two bearings on the two sides support the rotary mill, which is rotating
slowly in these bearings. Although each bearing diameter is much less than the
rotary mill diameter, it is still very large in comparison to common bearings in
machinery. The trunnion on each side of the rotary mill is a hollow shaft (large-
diameter sleeve) that is turning in the bearings; at the same time, it is used for
feeding the raw material and as an outlet for the reduced-size processed material.
The internal diameter of the trunnion must be large enough to accommodate the
high feed rate of ores. The trunnion outside diameter is usually more than 1.2 m.
In the past, as long as the trunnion outside diameter was below 1.2 m, large
rolling-element bearings were used to support the trunnion. However, rolling
bearings are not practical any more for the larger trunnion diameters currently in
use in rotary mills. Several designs of self-aligning hydrodynamic bearings are
still in use in many rotary mills. These designs include a hydrostatic lift, of high
hydrostatic pressure from an external pump, only during start-up. This hydrostatic
lift prevents the wear and high friction torque during start-up. Due to the low
speed of rotation, these hydrodynamic bearings operate with very low minimum
film thickness. Nevertheless, these hydrodynamic bearings have been operating
successfully for many years in various rotary mills. The hydrodynamic bearings
are designed with a thick layer of white metal bearing material (babbitt), and a
cooling arrangement is included in the bearing. However, ever-increasing
trunnion size makes the use of continuous hydrostatic bearings the preferred
choice.
Large-diameter bearings require special design considerations. A major
problem is the lack of manufacturing precision in large bearings. A large-
diameter trunnion is less accurate in comparison to a regular, small-size journal,
for the following reasons.
1. Machining errors of round parts, in the form of out-of-roundness, are
usually proportional to the diameter.
2. The trunnion supports the heavy load of the mill, and elastic deforma-
tion of the hollow trunnion causes it to deviate from its ideal round
geometry.
3. Many processes require continuous flow of hot air into the rotary mill
through the trunnion, to dry the ores. This would result in thermal
distortion of the trunnion; in turn, it would cause additional out-of-
roundness errors.
In addition, the large trunnion becomes slightly oval under the heavy load
of the mill. For a large trunnion, out-of-roundness errors due to elastic deforma-
tion are of the order of 6 mm (one-quarter inch). In addition, mounting errors,
deflection of the mill axis, and out-of-roundness errors in the machining of the
trunnion surface all add up to quite significant errors that require continuous
clearance adjustment by means of radial motion of the pads. Also, it is impossible
to construct the hydrostatic system precisely so that all pads will have equal
clearance, h0 , for equally sharing the load among the pads. Therefore, the
hydrostatic pads must be designed to be self-adjusting; namely they must move
automatically in the radial direction of the trunnion until the load is equalized
among all the pads.
case of oil pump failure. The trunnion bearings are continually monitored, to
prevent early failure due to unexpected conditions. The bearing temperature and
the hydrostatic pressure are continually recorded, and a warning system is set off
whenever these values exceed an acceptable limit. In addition, the operation of
the mill is automatically cut off when the hydrostatic bearing loses its pressure.
The hydraulic supply system is shown in Fig. 10-15. Each shoe is fed at a
constant flow rate by four flow dividers. One positive-displacement pump is used
to pump the oil from the oil tank. The oil is fed into the flow dividers through an
oil filter, relief valve, and check valve. An accumulator is used to reduce the
pressure fluctuations involved in positive-displacement pumps. Four pumps can
be used as well. The oil returns to the sump and passes to the oil tank. Additional
safety devices are pressure sensors provided to ensure that if the supply pressure