Technical Document

AU: Manring ND

Date: 04/23/2001

Job Director: Litherland JW

SUBJ: Estimating Fluid Bulk-Modulus within Hydraulic Systems

Abs: In this report, the fluid bulk modulus is investigated from the theoretical,
computational and experimental respects. Based upon the definition of
secant bulk modulus and the tangent bulk modulus, an effective bulk
modulus is presented which considers the inclusion of entrained air within
the fluid and the expansion of the fluid container itself. The investigation
shows that the effective fluid bulk modulus within a hydraulic system may
vary extensively depending upon the amount of entrained air and the
elasticity of the devices that are being used to contain the fluid. The
techniques for measuring the fluid bulk modulus within the lab and within
a working hydraulic system, along with the estimations of uncertainty for
each measurement, is discussed in detail in this report. By increasing the
accuracy of the instrumentation, the measurement uncertainties may be
reduced significantly.

Keys: Fluid Mass Density

Fluid Bulk Modulus
Secant Bulk Modulus
Tangent Bulk Modulus
Effective Fluid Bulk Modulus
Measuring Fluid Bulk Modulus
Bulk Modulus of Air
Bulk Modulus of Liquid
Bulk Modulus of Container

Retention: Indefinite

Caterpillar Confidential: Yellow 1

 Manring ND
04/23/2001

Date: 04/23/2001

Author: Manring ND

Charge No: 11-27027

Test Period: No test was conducted

Source Div: 394-07, Hydraulics R&D, TC-G

SUBJ: Estimating Fluid Bulk-Modulus within Hydraulic Systems

Introduction:

In this research, the fluid property that describes the compressibility of a hydraulic fluid is
presented. Since imparting velocity to a pressurized fluid is the means for transmitting
power hydraulically, it is important to consider the physical mechanisms that describe the
pressurization of the fluid and the fluid’s resistance to flow. In this report, the mass density
of a liquid is discussed with the bulk coefficients that characterize the equation of state.
One of the bulk coefficients is the fluid bulk modulus. It is shown that two definitions for the
fluid bulk modulus exist: the secant bulk modulus and the tangent bulk modulus. Using
these definitions, an effective bulk modulus is presented which considers the inclusion of
entrained air within the fluid and the expansion of the fluid container itself. It is also shown
that the effective fluid bulk modulus is similar to a consideration of springs that are
arranged in series. Consideration is given to the effort of measuring the fluid bulk modulus
within the lab and within a working hydraulic system. Associated with this discussion are
the estimations of uncertainty that may exist within each measurement. It is shown that
measurements within the lab can be trusted within ±5% while measurements within a
working hydraulic system can be anywhere between ±18 to 36%. Extremely accurate
instrumentation can be used to reduce this uncertainty.

Objectives:

1. Investigate the techniques for analyzing and measuring the fluid bulk modulus
within the lab and a real hydraulic circuit.
2. Provide theoretical guideline for determining the fluid bulk modulus in analyzing,
modeling and designing a hydraulic component and system.

Conclusions:

1. The effective fluid bulk modulus within a hydraulic system may vary extensively
(69%) depending upon the amount of entrained air and the elasticity of the
devices that are being used to contain the fluid.
2. Techniques for measuring the fluid bulk modulus can be used for a fixed volume,
fluid flowing through an orifice, and fluid flowing across a pump or motor.
3. When the bulk modulus of moving fluid is being measured, the measurement
becomes extremely sensitive to the accuracy of the flow meters being used to

Caterpillar Confidential: Yellow 2

 Manring ND
04/23/2001

make the measurement. The more flow meters required in the measurement,
the higher the uncertainty is in the bulk modulus measurement.
4. The effective bulk modulus may vary tremendously based upon construction
(thick or thin walls) and material types of the containing system. Nevertheless, it
is bounded between two extremes that can be calculated for a given material.
These bounds help one to identify the diminishing returns of increasing the wall
thickness of the container.
5. While measuring the bulk modulus may have high uncertainty associated with
the measurement, it may still be better than simply guessing at the amount of
entrained air, etc.

Recommendations:

1. Use high accuracy flow meter in measuring the fluid bulk modulus of a moving
fluid.
2. Eliminate as much entrained air as possible from the hydraulic systems.

Significant Results:

1. The effective fluid bulk modulus is analogous to a set of springs in series.

Therefore, the stiffness of the fluid will never exceed the lowest “spring rate”
associated with the air, the liquid, and the container.
2. The bulk modulus of a pure liquid can be determined based upon reliable test
data shown in Table 1. The bulk modulus of entrained air is a linear function of
the fluid pressure.
3. The fluid bulk modulus for a fluid that is contained within a closed container may
be measured with an estimated uncertainty of ± 5%.
4. For fluid flowing through an orifice, the fluid bulk modulus may be measured with
an estimated uncertainty of ± 18.5%
5. For fluid flowing across a pump or motor, it has been shown that the fluid bulk
modulus may be measured with an estimated uncertainty of ± 36%.

Noah D. Manring
University of Missouri-Columbia
(573) 882-7539

Caterpillar Confidential: Yellow 3

 Manring ND
04/23/2001

Discussion:

Fluid Mass Density

Equation of State. The equation of state for any substance is used to describe the
mass density of that substance as it varies with exposed conditions such as pressure and
temperature. For many substances, the equation of state is extremely complex and
difficult to describe exactly. An example of a relatively simple equation of state is that for
an ideal gas. In this case, the ideal gas law is used to relate the pressure and temperature
of the gas to the mass density using a determined gas constant. This result is given by

P
ρ= , (1)
RT

where ρ is the mass density of the gas, T is the absolute temperature of the gas
(expressed in Kelvin or Rankine), P is the gas pressure, and R is the universal gas
constant that has been determined for the specific gas in question. Unfortunately, for the
case of liquids, the equation of state is not so simple. However, since liquids are fairly
incompressible, it may be assumed that the mass density of a liquid will not change
significantly with the exposed conditions of temperature and pressure. In this case, a first
order Taylor series approximation may be written to describe the small variations in density
that do occur due to changes in pressure and temperature. This Taylor series is given by

∂ρ ∂ρ
ρ = ρo + ( P − Po ) + (T − To ) , (2)
∂P o ∂T o

where ρo , Po , and To represent a reference density, pressure, and temperature

respectively. The equation of state may be more meaningfully expressed by making the
following definitions,

1 1 ∂ρ 1 ∂ρ
≡ , α ≡ − , (3)
β ρ ∂P ρ ∂T

where β is the isothermal fluid bulk modulus and α is the isobar fluid coefficient of
thermal expansion. Using Equation (3) with Equation (2) yields the following equation of
state for a liquid:

 1 
ρ = ρo  1 + ( P − Po ) − αo (T − To )  . (4)
 βo 

If the nearly incompressible assumption of this equation is correct, one may infer that the
fluid bulk modulus, β , is large and the thermal coefficient of expansion, α , is small.
Indeed this is the case for hydraulic fluids as will be shown for the fluid bulk modulus in
subsequent sections of this report.

Caterpillar Confidential: Yellow 4

 Manring ND
04/23/2001

Density-Volume Relationship. To evaluate the relationship between fluid mass

density and fluid volume, consider a fluid element of mass, M . This mass may be
described as

M = ρV , (5)

where ρ is the fluid mass density and V is the volume of the fluid element. An
infinitesimal change in this mass would be described by

dM = ρ dV + V d ρ . (6)

Since the mass itself cannot be diminished or increased, the left-hand-side of equation (6)
must be zero. Therefore, the following differential relationship between the fluid density
and the fluid volume may be expressed as

1 1
− dρ = dV . (7)
ρ V

Integrating both sides of this equation between say, condition 1 and condition 2, yields the
following relationship between the fluid mass density and the fluid volume:

ρ1 V
= 2 . (8)
ρ2 V1

With Equation (5) it can be seen that this result simply states that the mass at condition 1
must equal the mass at condition 2.

Definitions of the Fluid Bulk Modulus

Stress-Strain Curves. The isothermal fluid bulk modulus describes the elasticity or
“stretchiness” of the fluid at a constant temperature. This property is experimentally
determined using a stress-strain test in which the volume of fluid is decreased while
keeping the mass constant. During this process, the stress of the fluid is measured by
measuring the fluid pressure. A plot of the fluid pressure versus the fluid strain is then
generated and the slope of this plot is used to describe the elasticity of the fluid. This
slope is generally referred to as the fluid bulk modulus. Figure 1 shows a plot of the
stress-strain curve for a liquid. As shown in this figure, the stress strain curve is not linear,
as its slope is shown to vary in magnitude with pressure. In Figure 1, both the secant bulk
modulus, Κ , and the tangent bulk modulus, β , are shown.

Secant Bulk Modulus. Using Figure 1, it can be seen that the secant bulk modulus
is defined as

∆P
Κ= , (9)
∆ε

Caterpillar Confidential: Yellow 5

 Manring ND
04/23/2001

Figure 1. The stress-strain curve for a liquid showing the secant bulk
modulus, Κ , and the tangent bulk modulus, β

where P is the fluid pressure and ε is the fluid strain. For the secant bulk modulus, the
fluid strain is defined by

∆V V −V
ε ≡ = o , (10)
Vo Vo

where Vo is the fluid volume at atmospheric pressure and V is the fluid volume at another
point of interest. From Figure 1 and Equation (10), it may be shown that

V1 − V
∆P = P − P1 and ∆ε = . (11)
Vo

Substituting Equation (11) into Equation (9) yields the following result for the secant bulk
modulus:
V ( P − P1 )
Κ = o . (12)
V1 − V

By convention, the reference pressure and volume for the calculation of the secant bulk
modulus is given by P1 = 0 and V1 = Vo . Notice: with this convention, gauge pressures are
assumed. Substituting these reference values into Equation (12) yields the following
conventional expression for the secant fluid bulk modulus:

Caterpillar Confidential: Yellow 6

 Manring ND
04/23/2001

Vo P ρP
Κ = = . (13)
Vo − V ρ − ρo

In this result, we have used the relationship between the fluid volume, V , and the fluid
density, ρ , as shown in Equation (8).

Tangent Bulk Modulus. The tangent fluid bulk modulus is defined by the slope of
a line that is anywhere tangent to the stress-strain curve shown in Figure 1. This quantity
is expressed mathematically by the following limit:

∆P dP
β = lim = . (14)
∆ε → 0 ∆ε dε

The fluid strain for the calculation of the tangent bulk modulus is defined by

V 
ε ≡ − ln   . (15)
 Vo 

At this point, the reader may note that the fluid strain for the tangent bulk modulus has
been defined differently than it was for the secant bulk modulus. See Equation (10). In
fact, one might argue that the strain definition for the secant bulk modulus is more typical
and expected than the strain definition presented in equation (15). This, of course, would
be true especially from a perspective of solid mechanics where the definition of material
strain is very similar in form to that of the secant bulk modulus. The use of equation (15)
for describing the strain of the tangent bulk modulus is more an issue of tradition rather
than principle; however, it may give the reader comfort to recognize that, for values of
V / Vo ≈ 1 , equation (15) may be linearly approximated as the secant bulk modulus just as it
has been presented in equation (10). Since liquid is fairly incompressible, this
approximation is easily justified and the two strain definitions can be viewed as essentially
the same. Using Equation (15), it may be shown that

1
dε = − dV . (16)
V

Therefore, the tangent fluid bulk modulus may be more explicitly expressed as

dP dP
β = −V = ρ . (17)
dV dρ

Note: in this result, we have used the relationship between the fluid volume, V , and the
fluid density, ρ , as shown in Equation (7).

Caterpillar Confidential: Yellow 7

 Manring ND
04/23/2001

Effective Bulk Modulus

General Equations. The fluid bulk modulus has been used to describe the
elasticity of the fluid as it undergoes a volumetric deformation. This elasticity describes a
spring effect that is often attributed to high frequency resonance within hydraulic systems.
High frequency resonance can create irritating noise problems and premature failures of
some vibrating parts; however, it usually does not present a control problem since the
resonance occurs at a much higher frequency than the lowest natural frequency of the
typical device being controlled. If, on the other hand, the effective spring rate of the
hydraulic system becomes soft due to entrained air within the system, or an overly
compliant fluid container, the resonating frequency of the hydraulic system will be come
much lower and a potential for control difficulties will exist.

Figure 2 shows a schematic of a flexible container filled with a fluid mixture of liquid and
air. A piston is moved to the left to compress the fluid while also expanding the structural
volume of the container. The total volume of the chamber is given by

V = Vo + Vδ − A x = Vl + Va , (18)

where Vo is the initial volume of the container, Vδ is an additional volume that results from
expanding the chamber, A is the cross-sectional area of the piston, x is the piston
displacement, Vl is the volume of liquid within the chamber, and Va is the volume of air
within the chamber. By subtracting the deflection volume, Vδ , from each side of Equation
(18), an effective volume may be calculated as

Ve = Vo − A x = Vl + Va − Vδ . (19)

Figure 2. A pressurized flexible container filled with a fluid mixture of liquid and air

Caterpillar Confidential: Yellow 8

 Manring ND
04/23/2001

This definition for the effective volume is useful since it represents a quantity that may be
easily calculated without knowing the deformation characteristics of the chamber. The
change in the effective volume is differentially expressed as

dVe = dVl + dVa − dVδ . (20)

Using the general form of Equation (17), the effective fluid bulk modulus can be described
as
1 1 dVe
= − , (21)
βe Ve dP

where β e is the effective fluid bulk modulus, Ve is the effective volume that undergoes
deformation, and P is the fluid pressure within the hydraulic system. Substituting equation
(20) into equation (21) produces the following result for the effective fluid bulk modulus:

1 Vl  1 dVl  Va  1 dVa  1 dVδ

= −  + −  + . (22)
βe Ve  Vl dP  Ve  Va dP  Ve dP

By definition, the bulk modulus for liquid and air are given respectively as

1 1 dVl 1 1 dVa
= − , = − . (23)
βl Vl dP βa Va dP

It is also useful to define the bulk modulus of the container with respect to the effective
fluid volume as

1 1 dVδ
= . (24)
βc Ve dP

Notice: this definition is different than that of Equation (23) since the volume and
differential volume terms are based upon different volume quantities. Substituting
Equations (23) and (24) into Equation (22) yields the following result for the effective fluid
bulk modulus of the system shown in Figure 2:

1 Vl 1 V 1 1
= + a + . (25)
βe Ve β l Ve β a βc

The volumetric ratios within this expression describe the factional volume content of liquid
and air. Equation (19) may be used with Equation (15) to show that

1  V  1  β V 1 1
= 1 + δ  +  1− a  a + . (26)
βe  Ve  β l  β l  Ve β a βc

Recognizing that Ve >> Vδ and βl >> β a , Equation (26) may be closely approximated as

Caterpillar Confidential: Yellow 9

 Manring ND
04/23/2001

1 1 Va 1 1
= + + . (27)
βe βl Ve β a βc

Equation (27) is a useful expression for describing the effective fluid bulk modulus within
the flexible container shown in Figure 2; however, there are several unknowns in this
equation that must be discussed further. In particular, useful expressions for the bulk
moduli must be developed and a consideration of the fractional content of air should be
discussed.

Figure 3. An equivalent spring system illustrating the compressibility

effects of the liquid, the air, and the container

An Equivalent Spring System. Figure 3 shows a spring system that is equivalent

to the pressurized chamber shown in Figure 2. The equivalent spring system is useful for
showing the effective spring rate of the hydraulic system and for lending insight into the
reduction of the natural frequencies of oscillation. In the top portion of Figure 3, a cylinder
is shown with internal pistons that are separated by springs. The right hand spring is
intended to model the spring rate of the air within the system. The middle spring is used to
model the spring rate of the liquid, and the left hand spring is used to simulate the spring
rate of the container. The pistons are free to slide within the cylinder and may change their
location depending upon the input force, F .

From an overall system view, the input force may be described as

F = x ke , (28)

Caterpillar Confidential: Yellow 10

 Manring ND
04/23/2001

where x is the displacement of the first piston and k e is the effective spring rate of the
overall system. From a static analysis of the spring system, it can be shown that the input
force is also described by

F = ( x − x1 ) ka = ( x1 − x2 ) kl = x2 k c , (29)

where k a , kl , and k c are the spring rates of the air, the liquid, and the container
respectively. From Equation (29) it may be shown that

F
x2 = ,
kc
F F F
x1 = x2 + = + , (30)
kl kc kl
F F F F
x = x1 + = + + .
ka kc kl ka

From Equation (28) it can be seen that x = F / ke . Substituting this expression into the
bottom result of Equation (30) yields the following result for the effective spring rate of the
system:

1 1 1 1
= + + . (31)
ke kc kl ka

This is the classical expression that is used to describe the effective spring rate for a group
of springs that are placed in series with respect to one another.

The bottom portion of Figure 3 is shown with certain springs removed and, now, the
spaces between pistons have been filled with liquid and air. The spring associated with
the container remains, as this is the best model for the effect of deforming a solid material.
From the bottom portion of Figure 3, it can be seen that the input force is statically
equivalent to the pressure of the air (or liquid) times the cross sectional area of a single
piston. This force is simply expressed by

F = PA . (32)

From the definition of the effective fluid bulk modulus given in Equation (21), it can be seen
that the pressure within the system is described by the differential expression

1
dP = − β e dVe , (33)
Ve

where β e is the effective fluid bulk modulus and Ve is the effective volume of the fluid.
Solving Equation (33) produces the following result for the fluid pressure:

Caterpillar Confidential: Yellow 11

 Manring ND
04/23/2001

V  V 
P = β e ln  o − 1 ≈ βe  o − 1 , (34)
 Ve   Ve 

where Vo is the volume of the fluid when the pressure is zero. Note: the right-hand-side of
Equation (34) assumes that Vo ≈ Ve which means that the changes in the fluid volume are
small. Using Equation (34) with Equation (32) yields the following result for the input force
to the equivalent spring system:

V 
F = β e  o − 1 A . (35)
 Ve 

Setting Equation (35) equal to Equation (28) yields the following result for the effective
spring rate of the system:

1 1 Ve
= . (36)
ke βe A2

In this result, it has been recognized from Equation (19) that Ve = Vo − A x . Since the
definitions of the effective fluid bulk modulus and the bulk modulus for air and liquid are
similar in form (compare Equations (21) and (23), a rerun of the previous analysis may be
done for the columns of air and liquid shown in Figure 3. This analysis produces the
following results for the spring rate of air and the spring rate of liquid:

1 1 Va 1 1 Vl
= , = . (37)
ka β a A2 kl β l A2

From the static analysis of Figure 3, it can be shown that

k c x2 = P A . (38)

Equation (24) presents the definition of the container bulk modulus and can be used to
develop an equivalent expression for the fluid pressure within the system. This expression
is determined by rearranging Equation (24) as follows:

1
dP = β c dVδ . (39)
Ve

Solving this equation yields

1 1
P = βc Vδ = β c A x2 , (40)
Ve Ve

where it has been recognized that Vδ = A x2 . Substituting Equation (40) into Equation (38)
yields the following result for the spring rate of the container:

Caterpillar Confidential: Yellow 12

 Manring ND
04/23/2001

1 1 Ve
= . (41)
kc β c A2

By substituting the results of Equations (36), (37), and (41) into Equation (31), the effective
fluid bulk modulus for the system may be expressed as

1 Vl 1 V 1 1
= + a + , (42)
βe Ve β l Ve β a βc

which is the exact same expression presented in Equation (25). Note: making the
appropriate simplifications this result may be further reduced to the form of Equation (27).

All of this discussion has been used to show that the compressibility effects within the
hydraulic system may be considered as a series of springs, which describe the stiffness of
the liquid, the air, and the container itself. Since these springs are arranged in series, the
overall stiffness of the system will never exceed the stiffness of any one spring. The spring
rate of each substance is dependent upon geometry and the bulk modulus property. In the
sections that follow, the bulk modulus of the liquid, the air, and the container, will be
considered in detail.

Bulk Modulus of a Pure Liquid

From experiments, it has been determined that, over a limited pressure range, the secant
bulk modulus of all liquids increases linearly with pressure. That is to say

Κ = Κo + m P , (43)

where Κ o is the secant bulk modulus of the liquid at zero gauge pressure and m is the
slope of increase. For any one liquid, the value of m is practically the same at all
temperatures; however, the value of Κ o carries temperature dependence with it. Equation
(43) is valid for mineral oils and most other hydraulic fluids up to about 800 bar. This same
equation is valid for water up to about 3,000 bar. Consequently, if the appropriate values
for Κ o and m are known for any liquid, the secant bulk modulus for that liquid can be
easily calculated using Equation (43). By setting Equation (43) equal to Equation (13), it
may be shown that

(Κ + m P)
2
 P  dP
V = Vo  1 −  , = − o . (44)
 Κo + m P  dV Κ o Vo

Substituting this result into Equation (17) produces the following expression that may be
used to evaluate the tangent bulk modulus of a liquid as a function of the liquid parameters
Κ o and m :
 ( m − 1) P   mP 
βl = Κ o 1 +  1 +  . (45)
 Κo  Κo 

Caterpillar Confidential: Yellow 13

 Manring ND
04/23/2001

Table 1 presents typical fluid properties for liquids that are commonly used within hydraulic
systems.

Table 1. Fluid bulk modulus properties, Κ o and m . Values for Κ o are in kbar.
Water in Oil Phosphate
Temperature °C Mineral Oil Water Water Glycol Emulsion Ester*
0 20.7 19.7 32.0 20.8 29.7
10 19.8 20.9 31.8 20.2 28.1
20 19.0 21.8 31.5 19.6 26.5
30 18.1 22.4 31.1 19.0 25.0
40 17.3 22.6 30.5 18.4 23.6
50 16.4 22.7 29.9 17.8 22.3
60 15.6 22.5 29.1 17.2 21.1
70 14.7 22.2 28.2 16.6 19.9
80 13.9 21.6 27.2 16.0 18.8
90 13.0 21.1 26.0 15.4 17.8
100 12.2 20.4 24.8 14.8 16.9
m
5.6 3.4 4.5 5.0 5.5
(for all temperatures)

* Viscosity greater than 50 cSt at 22°C

Bulk Modulus of Air

The bulk modulus of air may be determined by assuming that the expansion and
compression of the air within the system occurs adiabatically. That is, we assume that no
heat transfer occurs between the air and the surrounding liquid or container material.
Using the first law of thermodynamics and the ideal gas law, it can be shown that

γ
P Va = constant , (46)

where P is the air pressure, Va is the air volume, and γ is the ratio of the constant
pressure specific heat to the constant volume specific heat. Note: γ = 1.4 for air. Taking
the derivative of equation (46) it may be shown that

γ ( γ −1)
Va dP + P γ Va dVa = 0 . (47)

γ
Dividing this expression through by Va , and rearranging terms, produces the following
expression for the bulk modulus of entrained air within the hydraulic system.

1 1 dVa 1
= − = . (48)
βa Va dP Pγ

Caterpillar Confidential: Yellow 14

 Manring ND
04/23/2001

Bulk Modulus of the Container

To consider the bulk modulus of a container, it will be instructive to examine the cylindrical
container shown in Figure 2-2. The volume of this container is given by

π
Vc = d2 L , (49)
4

where d is the diameter of the container and L is the container length. If it is assumed
that the container expands only in the radial direction, then the expanded volume of the
container may be expressed as

π π 2 δ δ 
2

Vc = ( do + 2 δ ) L = d o 1 + 4   + 4    L ,
2
(50)
4 4   do   d o  

where d o is the original diameter of the container and δ is the radial deflection as shown
in Figure 2-2. If it is assumed that δ / d o << 1 , then the container volume may be closely
approximated as

Vc = Vo + Vδ , (51)

where the original volume and the deformed volume are given respectively by

π 2
Vo = d o L , Vδ = π d o δ L . (52)
4

Notice: these are more explicit expressions for the volume terms that were used in
Equation (18) for describing the total volume of the container. From a strength of materials
text book, we learn that the inside radial deflection of a thick walled cylinder (without
capped ends) is given by

d o P  Do + d o 
2 2

δ =  2 + ν , (53)
2 E  Do − d o 2


where P is the internal pressure, E is the tensile modulus of elasticity for the cylinder
material, ν is Poisson’s ratio, Do is the original outside diameter of the cylinder, and d o is
the original inside diameter of the cylinder. Note: the result of Equation (53) is valid for
both thick and thin walled cylinders. Using Equations (52) and (53) together, it may be
shown that the deformed volume of the cylinder is given by

π P  Do + d o 
2 2

Vδ = + νL .
2
do  2 (54)
2 E  Do − d o 2


The derivative of this expression with respect to the pressure P is given by

Caterpillar Confidential: Yellow 15

 Manring ND
04/23/2001

dVδ π d o 2 L  Do 2 + d o 2 
=  2 + ν . (55)
dP 2 E  Do − d o 2


If we assume that the ratio of the displaced volume in the chamber to the original volume is
much less than 1 (i.e., Ax / Vo << 1 ), then Equation (19) may be used to show that effective
volume of the chamber is given by

π
Ve ≈ Vo =
2
do L . (56)
4

Equations (55) and (56) may now be substituted into Equation (24) to express the
container bulk modulus as

2  Do + d o 
2 2
1
=  2 + ν . (57)
βc E  Do − d o 2


Recognizing that the inside diameter of the container is given by d o = Do − 2 t , where t is

the container wall thickness, an equivalent expression for the container bulk modulus may
be written as

1 2  Do  t  
=   1 +  − (1 − ν )  . (58)
βc E 2t  Do − t  

Two special cases of these results are instructive. For very thick walls, it may be assumed
that Do / d o >> 1 . In this case, Equation (57) may be approximated as

1 2 (1 + ν )
= . (59)
βc E

On the other hand, for a thin walled cylinder, it may be assumed that Do / t >> 1 . In this
case, Equation (58) may be used to show that

1 Do / t
= . (60)
βc E

Equations (59) and (60) may be used to describe the range for the bulk modulus of the
container as

2 (1 + ν ) 1 Do / t
< < . (61)
E βc E

Table 2 presents the modulus of elasticity and Poisson’s ratio for common materials that
are used to construct hydraulic containers.

Caterpillar Confidential: Yellow 16

 Manring ND
04/23/2001

Table 2. Material properties for common hydraulic containers. The modulus of

elasticity, E , is reported in kbar.

Ductile High Press.

Property Steel Copper Brass Aluminum TPE**
Cast Iron Hose*
E 2,069 1,655 1,103 1,034 724 59 3.93E-02
ν 0.30 0.28 0.30 0.34 0.33 0.47 0.47

* flexible and reinforced with stainless steel braids (consult manufacturers for
more accurate numbers)
** thermoplastic elastomer (melt-processible rubber)

Typical Calculations

The preceding work has shown that the effective bulk modulus is a result of various
physical considerations. It is instructive to consider a few typical calculations of the
effective bulk modulus for the purposes of illustrating the wide variation that can be
expected for this parameter. Let us consider a petroleum-based fluid (mineral oil) that is
pressurized to 20 MPa (0.20 kbar) at a temperature of 70°C. Furthermore, let us assume
that the fluid is contained within a cylindrical tube with an outside diameter equal to six
times the wall thickness ( Do / t = 6 ). Using the general results of the preceding analysis, it
may be shown that the air content within the fluid and the use of a flexible versus a rigid
container can significantly alter the effective bulk modulus of the system. Table 3 shows
the results of these calculations.

Table 3. Sample calculations of the effective bulk modulus, β e , at a fluid pressure of 0.20
kbar and 70°C. Bulk modulus values are reported in kbar.

Container Material Va / Ve = 0.00 Va / Ve = 0.01

Steel 16.05 10.20
High Pressure Hose 6.11 5.02

In Table 3, it can be seen that using a rigid container, with essentially no air in the fluid,
creates the stiffest system possible. By adding 1% air (by volume) to the fluid, the bulk
modulus may be reduced by 36%. This, however, may not be the greatest impact since
the use of a flexible hose, rather than a steel pipe, may, in effect, reduce the fluid bulk
modulus by 62%. The maximum reduction in the effective bulk modulus occurs when air is
added to the fluid and when a flexible hose is used instead of a steel pipe. As shown in
Table 3, this reduction can be as great as 69%. In summary, it can be noted that a wide
variation in the effective bulk modulus may be expected from one hydraulic system to the
next. This variation is highly dependent upon the air content of the fluid, which may also
be dependent upon the circulation system for the fluid. For instance, one may suspect that
an open circuit system, where fluid is circulated through a ventilated reservoir, may contain

Caterpillar Confidential: Yellow 17

 Manring ND
04/23/2001

more entrained air than a closed circuit system, where most of the fluid is contained within
the working components of the system. Other issues related with reservoir design may
impact the turbulent mixing that may occur within the reservoir. This may also impact the
amount of entrained air that is introduced by each application. Since the uncertainty of the
fluid bulk modulus is significant from one application to the next, it is worth considering
several techniques that may be used to measure the fluid bulk modulus within an actual
system.

Measuring the Fluid Bulk Modulus

Fluid within a Closed Container. Figure 4 shows a closed container with a piston
that is supported by a fluid generally comprised of a mixture of liquid and air. The piston
has a cross-sectional area given by, A . When the fluid pressure is at zero gauge
pressure, the height of the fluid column is given by the dimension, lo . When the force, F ,
is applied to the piston, the fluid column is decreased in height by the dimension, x . The
support the applied force, the fluid pressure in the container increases.

Figure 4. Geometry of a test device used to measure the fluid bulk

modulus within a closed container

From the definition of the fluid bulk modulus, the fluid pressure within the closed container
may be differentially expressed as

1
dP = β dV , (62)
V

where β is the effective fluid bulk modulus and V is the effective fluid volume. By treating
β as a constant, Equation (62) may be solved and rearranged to show that the effective
fluid bulk modulus is given by

P
β= , (63)
ln (Vo / V )

Caterpillar Confidential: Yellow 18

 Manring ND
04/23/2001

where Vo is the fluid volume in the chamber when the pressure is zero. This volume would
be given more explicitly as A lo . From geometry, the effective fluid volume after
compression takes place may be written as

V = Vo − A x = A ( lo − x ) . (64)

Substituting this result into Equation (63) yields the following result for the effective fluid
bulk modulus:

P
β= . (65)
ln ( lo / (lo − x ) )

From this equation it can be seen that the parameters that must be measured for
determining the fluid bulk modulus are the fluid pressure, P , the original height of the fluid
column, lo , and the displaced distance, x . As all experimentalists know, the
measurements of these parameters will always deviate slightly from the actual physical
values. If we denote these measured parameters using primed notation, the measured
value for the fluid bulk modulus may be written as

P′
β′= . (66)
ln ( lo′ / (lo′ − x ′) )

A dimensionless error and a percent error in our measurements may be defined as

β − β′
ε= , %ε = ε × 100 . (67)
β

The dimensionless error may be more explicitly written as

ln ( lo / (lo − x ) )
ε = 1− β′ , (68)
P

where the unprimed values of P , lo , and x are the unknown true values of these
parameters. If it is assumed that the measurement error is small, a first order Taylor series
expansion of Equation (68) may be written as

∂ε ′  ∂ε ′  ∂ε ′ (
ε =  (
 P−P ′ ) + (
 ∂l  o ol − l ′ ) +   x − x ′) . (69)
 ∂P   o  ∂x 

Using Equation (68), and assuming that x ' / l ' << 1 , it can be shown that

′
 ∂ε ′ = 1 ,  ∂ε  ≈ 1 ∂ε ′
,   ≈
1
   ∂l  . (70)
 ∂P  P′  o lo′  ∂x  xo′

Caterpillar Confidential: Yellow 19

 Manring ND
04/23/2001

Substituting these results back into Equation (69) yields the following result for the error
associated with measuring the fluid bulk modulus in the closed container:

P l x
ε = + o − − 1 . (71)
P′ lo′ x′

The reader will recall that the primed variables are the measured values while the
unprimed variables are the true values. The accuracies for instruments that are used for
measuring pressures and displacements are typically specified in terms of some
percentage of the full-scale capabilities of the instrument itself. This means that the true
values of the measured parameters are somewhere within a range that has been identified
by the manufacturer of the instrument. Mathematically this maximum range of uncertainty
for each measurement may be described as

( P − P ′) = ± ξ P Pmax , ( lo − lo′ ) = ± ξl lmax , ( x − x ′) = ± ξ x xmax , (72)

where the values of ξ P , ξ l , and ξ x are supplied by the instrument manufacturers and Pmax ,
lmax , and xmax are the maximum measurement ranges for these instruments. Using the
results of Equations (69), (70), and (72) an expression for the maximum measurement
error may be given as

 P l x 
ε max = ± ξ P max + ξ l max + ξ x max  . (73)
 P′ lo′ x′ 

Equation (73) describes the maximum range of uncertainty associated with the
measurement of the fluid bulk modulus for the mechanical device shown in Figure 4. If a
strain-gauged pressured transducer is used, it is common to be able to measure pressures
accurately within ±1.5% of the full-scale reading. This means that ξ P = 0.015 . For
measuring linear dimensions, such as lo , the method used may vary from a highly
accurate coordinate-measuring machine (CMM) to a simple visual inspection of a
calibrated scale. Therefore, the range of accuracies for making the linear measurement of
the original height of the fluid column can be any where between ± 1 × 10−4 % and ± 1% of
the full-scale reading. This means that 1 × 10−6 < ξ l < 0.01 . In practice, it is quite common to
use a linear variable-differential-transformer (LVDT) for making linear displacement
measurements. These devices are typically accurate within ±0.5% of full-scale readings
and therefore, ξ x = 0.005 . If it is assumed that each instrument is operating at half of its
maximum reading, and using the values ξ P = 0.015 , ξ l = 0.005 , and ξ x = 0.005 , it can be
seen that range of uncertainty in the bulk modulus measurement is nominally ± 5% . This
range of uncertainty can get better or worse depending upon the quality of the instruments,
their operating range, and the operating point of the measurement. In any event, the
range of expected uncertainty should be checked before conducting the experiment to see
whether or not the measurement is worthwhile, or to design an acceptable test device.
Equation (73) provides the tool for checking this uncertainty as it pertains to the device
shown in Figure 4.

Caterpillar Confidential: Yellow 20

 Manring ND
04/23/2001

Fluid Flowing through an Orifice. The bulk modulus tester shown in Figure 4 is
useful for some applications but finds its primary value in an academic laboratory. In other
words, this test device only determines the fluid bulk modulus for a very well controlled
arrangement and does not give an approximation for the actual bulk modulus that is found
within a working hydraulic circuit. The fluid operating within a real hydraulic circuit is
occasionally exposed to mixing conditions that tend to entrain air bubbles within the fluid.
As previously mentioned, air is a very difficult fluid additive to quantify; therefore, it is of
value to consider techniques for measuring the fluid bulk modulus within a working circuit.

Figure 5. Schematic of a test device used to measure the fluid bulk

modulus as flow is passed through an orifice

Figure 5 shows a schematic of an orifice-type flow passage. The flow entering the orifice
is characterized by a pressure and temperature given by P1 and T1 , which results in a fluid
density, ρ1 . On the outlet side of the orifice, the pressure and temperature are given by P2
and T2 , with a corresponding fluid density, ρ 2 . The mass flow rate through the orifice is
given by

m& = Q1 ρ1 = Q2 ρ 2 , (74)

where Q1 and Q2 are the volumetric flow rates at the inlet and outlet of the orifice
respectively. From Equation (74) it can be seen that

Q2 ρ1
= . (75)
Q1 ρ 2

If the temperatures on both sides of the orifice are similar, i.e., T1 ≈ T2 , then the density
difference between the inlet and outlet of the orifice is a result of pressure only and the
isothermal bulk modulus may be calculated from its definition given by

1
dP = β dρ . (76)
ρ

Integrating this result across the orifice produces the following expression

P1 − P2 = β ln ( ρ1 / ρ 2 ) . (77)

Caterpillar Confidential: Yellow 21

 Manring ND
04/23/2001

This expression may be rearranged to show that

ρ1 P −P 
= Exp  1 2  . (78)
ρ2  β 

Using this result with Equation (75), it can be shown that the fluid bulk modulus may be
expressed as

P1 − P2
β = . (79)
ln (Q2 / Q1 )

From this equation, it can be seen that the fluid bulk modulus may be calculated from the
measured data of fluid pressures and volumetric flow rates in and out of the orifice. Again,
it is helpful to denote the difference between measured and true values using primed
notation, where the primed variables signify measured values and the unprimed variables
are true values. Using this convention, the calculation of the bulk modulus based upon
measured parameter values is given by

P1′ − P2′
β′ = . (80)
ln (Q2′ / Q1′ )

The dimensionless error and percent error associated with the measurement may, once
again, be calculated as follows:

β − β′
ε= , %ε = ε × 100 . (81)
β

Using this definition of the error with the result of Equation (79) produces a more explicit
form of the measurement error. This form is given by

ln (Q2 / Q1 )
ε = 1− β ′ , (82)
P1 − P2

where the unprimed values of Q1 , Q2 , P1 , and P2 are the unknown true values of
volumetric flow rate and pressure. If it is assumed that the measurement errors are small,
a first order Taylor series expansion of Equation (82) may be written as

 ∂ε ′  ∂ε ′  ∂ε ′  ∂ε ′
ε =   ( 1 1 )  ∂P  ( 2
P − P ′ + P − P2)
′ +  ∂Q  ( 1
Q − Q1)
′ +  ∂Q  (Q2 − Q2′ ) . (83)
 ∂P1   2  1  2

Using Equation (82), and recognizing that Q2′ / Q1′ ≈ 1 , it can be shown that

Caterpillar Confidential: Yellow 22

 Manring ND
04/23/2001

 ∂ε ′  ∂ε ′ 1  ∂ε ′  ∂ε ′ 1
 ∂P  = −  ∂P  = P ′ − P ′ ,  ∂Q  = −  ∂Q  ≈ Q ′ − Q ′ . (84)
 1  2  ( 1 2)  1  2 ( 2 1)

The partial differential of the error with respect to a measured parameter may be
considered as a sensitivity coefficient. If the sensitivity coefficient is large for a specific
measured parameter, then the error in the bulk modulus calculation will be large for small
errors in the measured parameter. From Equation (84) it can be seen that the sensitivity
with respect to pressure measurements tends to decrease as the pressure drop across the
orifice, P1′ − P2′ , increases. On the other hand, small pressure drops across the orifice
create large sensitivity coefficients for the pressure measurements. From this we can
conclude that our best bulk modulus calculations will be made using data taken for large
pressure drops across the orifice. Similarly, Equation (84) shows that the sensitivity with
respect to flow measurements tends to decrease as the difference in volumetric flow
measurements, Q2′ − Q1′ , increases. For nearly incompressible flows, however, this
disparity between volumetric flow measurements will never be very large. Therefore, we
conclude that this technique for calculating the effective fluid bulk modulus is very sensitive
to flow measurements and for this reason special care must be taken when selecting the
instruments that will be used for measuring flow.

Substituting the results of Equation (84) into Equation (83) produces the following result for
the error associated with measuring the fluid bulk modulus for the device shown in Figure
5:
(P − P ) (Q2 − Q1 )
ε = 1 2 − , (85)
( P1′ − P2′) (Q2′ − Q1′ )
where, again, the primed variables are measured values and the unprimed variables are
the unknown true values. If the accuracies for the instruments used for measuring
pressure and flow can be specified in terms of some percentage of the full-scale
capabilities of the instrument, the maximum range of uncertainty associated with each
measurement can be written as

( P1 − P1′) = ( P2 − P2′) = ± ξ P Pmax , (Q2 − Q2′ ) = (Q1 − Q1′ ) = ± ξQ Qmax , (86)

where the values of ξ P and ξQ are supplied by the instrument manufacturers and Pmax and
Qmax are the maximum measurement ranges for these instruments. Note: Equation (86)
assumes that both sides of the orifice are equipped with similar instrumentation. Using
these results with Equations (83) and (84), it may be shown that the maximum error in
calculating the fluid bulk modulus across the orifice is given by

 Pmax Qmax 
ε max = ±  2 ξ P + 2 ξQ
(Q2′ − Q1′ ) 
. (87)
 ( P1′ − P2′)

Equation (87) describes the maximum range of uncertainty associates with the
measurement of the fluid bulk modulus across the orifice shown in Figure 5. As previously

Caterpillar Confidential: Yellow 23

 Manring ND
04/23/2001

mentioned, for strain-gauged pressure transducers, the measurement accuracy is typically

found to be within ±1.5% of the full-scale reading. This means that ξ P = 0.015 . For flow
measurements, the accuracy depends upon the instrument chosen for the measurement.
If a turbine flow meter is used, the accuracy of the flow measurement can be anywhere
between ±0.5% and ±2.5% of full-scale readings. If a venturi-type flow meter is used, the
accuracy of the flow measurement can be anywhere between ±0.5% and ±1.0% of full-
scale readings for sufficiently high pressure-drops across the orifice. For positive
displacement flow meters, the accuracy can be more tightly controlled and typical
measurements can be trusted within ±0.25% and ±0.5% of full-scale readings. Having
said all this, it can be seen that a large range exists for the accuracy of the flow
measurements and that in practice, values of ξQ can be anywhere between 0.0025 and
0.0250 (a 10:1 difference). For a quick calculation of an expected error associated with
the above measurement of the fluid bulk modulus, let us assume that the pressure
transducer is operating at half of its maximum capacity and that the flow meter is operating
near its maximum capacity. Furthermore, let us assume that P1' >> P2' and that the flow
rate into the orifice is 96% of the output flow rate; i.e., Q1′ = 0.96 Q2′ . Given the parameters,
ξ P = 0.015 and ξQ = 0.0025 , it can be shown that the uncertainty in the bulk modulus
calculation is ± 18.5% with more than 12% of this uncertainty coming from the volumetric
flow measurements. With this amount of uncertainty, one may be questioning the value in
taking this measurement at all. On the other hand, it was previously shown that the
effective fluid bulk modulus might vary by 69% just going from a steel pipe to a flexible
hose and adding 1% of entrained air. Therefore, an 18.5% uncertainty is quite an
improvement over the 69% uncertainty that we previously had to live with. For this reason,
we will charge ahead and consider the measurement of the fluid bulk modulus across a
working hydraulic component. In the following subsection we will present a similar
technique for measuring the fluid bulk modulus across a pump or motor.

Fluid Flowing across a Pump or Motor. Figure 6 shows a schematic of a pump

or motor with three paths of fluid entering or exiting the device itself. Two of the paths are
the primary power transmission paths; the other path is a leak path where fluid is lost due
to internal clearances between machine parts. The conservation of mass for this system
requires

m& 1 = m& 2 + m& l , (88)

where m& 1 is the mass flow rate into the device through transmission line 1, m& 2 is the mass
flow rate out of the device through transmission line 2, and m& l is the mass flow rate out of
the device through leakage path shown in Figure 6. Equation (88) may be written in terms
of volumetric flow rates and fluid densities. This expression is given by

Q1 ρ1 = Q2 ρ 2 + Ql ρ l . (89)

Caterpillar Confidential: Yellow 24

 Manring ND
04/23/2001

Figure 6. Schematic of a test device used to measure the fluid bulk

modulus as flow is passed across a pump or motor

If the temperatures at all three flow locations are approximately equal, i.e., T1 ≈ T2 ≈ Tl , then
the fluid densities may be considered to change only with pressure. Using the definition of
the isothermal fluid bulk modulus, the following density relationships may be derived for
the system:

 P1  P 
ρ1 = ρo Exp   , ρ 2 = ρ o Exp  2  , ρ l = ρ o , (90)
β β 

where it has been assumed that the pressure in the leakage path is zero, ρo is the fluid
density in the leakage path, and β is the fluid bulk modulus. Since P1 / β << 1 and
P2 / β << 1 , the density relationships of Equation (90) may be approximated as

 P1   P 
ρ1 = ρo  1 +  , ρ 2 = ρo 1 + 2  , ρ l = ρo . (91)
 β  β

Substituting the results of Equation (91) into Equation (89), and rearranging terms, yields
the following expression for the fluid bulk modulus in the pump or motor device shown in
Figure 6:
Q P − Q2 P2
β = 1 1 . (92)
Ql − (Q1 − Q2 )

From this equation it can be seen that the fluid bulk modulus may be calculated by
measuring the volumetric flow rates of the fluid entering and leaving the pump or motor,

Caterpillar Confidential: Yellow 25

 Manring ND
04/23/2001

Q1 , Q2 and Ql , and by measuring the fluid pressures in the power transmission lines, P1
and P2 . Equation (92) assumes that the true values for these parameters are known. In
reality, the measured values are somewhat different than the true values; therefore, using
primed notation to identify measured values, the measure fluid bulk modulus may be
written as

Q1′ P1′ − Q2′ P2′

β′ = . (93)
Ql′ − ( Q1′ − Q2′ )

The dimensionless error and percent error associated with the measurement may, once
again, be calculated as follows:

β − β′
ε= , %ε = ε × 100 . (94)
β

Using this definition with the result of Equation (92) produces a more explicit form of the
measurement error. This form is given by

Ql − (Q1 − Q2 )
ε = 1 − β′ , (95)
Q1 P1 − Q2 P2

where the unprimed values of Q1 , Q2 , Ql , P1 , and P2 are the unknown true values of
volumetric flow rate and pressure. If the measurement errors are small, a first order Taylor
series expansion of Equation (95) may be written as

 ∂ε ′  ∂ε ′  ∂ε ′
ε =   ( 1 1 )  ∂P  ( 2
P − P ′ + P − P2)
′ +  ∂Q  (Q1 − Q1′ )
 ∂P1   2  1
(96)
 ∂ε ′  ∂ε ′
+  (Q2 − Q2′ ) +   (Ql − Ql′ ) .
 ∂Q2   ∂Ql 

Using Equation (95) it can be shown that

 ∂ε ′ Q1′  ∂ε ′ Q2′
 ∂P  = ,   = − ,
 1 Q1′ P1′ − Q2′ P2′  ∂P2  Q1′ P1′ − Q2′ P2′

 ∂ε ′ P1′  ∂ε ′  ∂ε ′ − P2′  ∂ε ′
 ∂Q  = −  ∂Q  ,  ∂Q  = + , (97)
 1 Q1′ P1′ − Q2′ P2′  l  2 Q1′ P1′ − Q2′ P2′  ∂Ql 

 ∂ε ′ 1
 ∂Q  = − Q ′ − Q ′ − Q ′ .
 l l ( 1 2)
Again, the partial derivatives of the error with respect to the measured parameters may be
considered as sensitivity coefficients. If the evaluated partial derivative is large, then the

Caterpillar Confidential: Yellow 26

 Manring ND
04/23/2001

bulk modulus error is very sensitive to small measurement errors associated with that
particular parameter. For large pressure drops across the pump or motor, the partials with
respect to pressure are reasonably small and therefore errors in the bulk modulus
calculation are not overly sensitive to the errors in the pressure measurements. On the
other hand, since Ql′ ≈ (Q1′ − Q2′ ) , it can be seen that the error calculation is very sensitive to
small errors in the flow measurements. This situation was typical of the orifice
measurements described in the previous subsection. Substituting this result into Equation
(96) produces the following expression for the measurement error:

Q1′ P1 − Q2′ P2 + Q1 P1′ − Q2 P2′ Q − (Q1 − Q2 )

ε = − l −1 , (98)
Q1′ P1′ − Q2′ P2′ Ql′ − (Q1′ − Q2′ )

where the primed variables are measured values and the unprimed variables are true
values. If the accuracies of the instruments are specified in terms of a percentage of the
full scale measurement capabilities, the maximum range of uncertainty for each
measurement can be expressed as

( P1 − P1′) = ( P2 − P2′) = ± ξ P Pmax , (Q2 − Q2′ ) = (Q1 − Q1′ ) = ± ξQ Qmax ,

(99)
(Ql − Ql′) = ± ξ q qmax ,

where the values of ξ P , ξQ , and ξ q are supplied by the instrument manufacturers and Pmax ,
Qmax , and qmax are the maximum measurement ranges for each instrument. Note:
Equation (99) assumes that the two power transmission lines of the pump or motor are
equipped with similar instrumentation. Using these results with Equations (96) and (97), it
may be shown that the maximum error in calculating the fluid bulk modulus across the
pump or motor is given by

 (Q1′ + Q2′ )   ( P1′ + P2′) 2 

ε max = ±   ξ P Pmax +  ′ ′ +  ξQ Qmax
 Q1′P1′ − Q2′ P2′   Q1 P1 − Q2′ P2′ Ql′ − (Q1′ − Q2′ ) 
(100)
 1  
+   ξ q qmax  .
 Ql′ − (Q1′ − Q2′ )  

For a quick calculation, let us assume that the device in Figure 6 is operating as a motor
and that P1' >> P2' . Furthermore, let us assume that Q2′ / Q1′ = 0.95 and that Ql′ / Q1′ = 0.07 .
Let us also assume that the pressure transducer on the high pressure side is operating at
half of its maximum capacity and that the flow meters in the power transmission lines are
operating near their full capacity. Also, the flow meter used to measure leakage is
operating at half its maximum capacity. Given the parameters, ξ P = 0.015 , ξQ = 0.0025 ,
and ξ q = 0.005 , it can be shown that the uncertainty in the bulk modulus calculation is
± 36% with more than 25% of this uncertainty coming from the volumetric flow
measurements in the power transmission lines. Again, one may question the value in

Caterpillar Confidential: Yellow 27

 Manring ND
04/23/2001

taking a measurement with this much uncertainty; however, this level of uncertainty may
still be better than the uncertainty associated with estimation related to amounts of
entrained air and estimations of the flexibility associated with the complex geometry
associated with certain container designs.

Conclusion

The fluid bulk modulus within a hydraulic system may vary extensively depending upon the
amount of entrained air and the elasticity of the devices that are being used to contain the
fluid. Typical calculations have shown that uncertainties in the fluid bulk modulus may be
as high as 69%. For this reason, it is of interest to devise techniques for measuring the
actual fluid bulk modulus and to estimate a range of uncertainty within these
measurements. In this research, it has been shown that the fluid bulk modulus for a fluid
that is contained within a closed container may be measured with an estimated uncertainty
of ± 5%. For fluid flowing through an orifice, it has been shown that the fluid bulk modulus
may be measured with an estimated uncertainty of ± 18.5%. For fluid flowing across a
pump or motor, it has been shown that the fluid bulk modulus may be measured with an
estimated uncertainty of ± 36%. While these ranges of measurement uncertainty are
noted to be large, it is worth noting that they are still less than the uncertainty of specifying
the fluid bulk modulus without making a measurement at all. By increasing the accuracy of
the instrumentation, the measurement uncertainties may be reduced. This comment is
especially important as it relates to the accuracies of the flow meters since they are used
to make some of the most sensitive measurements within the experiments.

Caterpillar Confidential: Yellow 28