1

TNO-TPD

TNO Institute of Applied Physics

Department of Flow and Structural Dynamics
Stieltjesweg 1
P.O. Box 155
2600 AD DELFT
The Netherlands
Phone +31 15 269 2000
Fax +31 15 269 2111

PULSIM, a powerful tool for

the control of vibrations in pipe
systems
By
Bert Egas

1.

Phone: +31 15 269 2134

Email: egas@tpd.tno.nl

Introduction

Pulsations or pressure waves are flow dynamic

phenomena, which can cause various kinds of
problems in installations in the process industry, the
power industry and the oil and gas industry. For
example pulsations can cause:
- Vibration and fatigue problems in piping and
fluid machinery (reciprocating compressors and
pumps, screw compressors, rootsblowers and
centrifugal compressors and pumps).
- A decrease of the efficiency of running
equipment (increase of power consumption or
reduction of capacity).
- Damage to fluid machinery (e.g. compressor or
pump valves or impellers).
- Cavitation problems.
- Errors in flow meter readings and process
control systems.
- Increased noise levels.
To avoid these problems a simulation program,
called PULSIM (PULsation SIMulator) has been
developed within the TNO-TPD. Simulation of flow
and piping dynamics in the design stage of a system
can reduce the number of problems during
operation.
2.

History Pulsim

In 1965 the so-called PULSIM project has been

started. An electrical analogue simulator
(PULsation SIMulator) has been built within the
department of Instrumentation of the TPD. Several
Industrial partners: Shell, Unilever, AKZO,
Thomassen and Gasunie financially contributed to
the project.
With this analogue computer rather simple pipe
systems of reciprocating compressors have been
investigated during approximately 15 years.

Since 1980 a digital simulation program has been

developed within the department of Flow and
Structural Dynamics of the TNO-TPD (PULSIM).
With this flexible program all kinds of dynamic
flow phenomena can be investigated in very
complex and long pipe systems. New modules for
the modeling of fluid machinery with their pipe
system have been added to PULSIM over the years,
i.e. detailed compressor valve dynamics, liquid
pumps, centrifugal compressors, burners, flow
induced pulsations and start-up and blow-down
conditions.
Moreover, the PULSIM program has been coupled
to the finite element program ANSYS for the
mechanical response analysis. In this way the
calculated pulsation induced forces (amplitude and
phase) are used directly for the calculation of
vibrations and cyclic stress levels of the piping. The
results of the analyses are continuously being
checked in practice for instance during trouble
shooting of existing systems.
Nowadays the Flow and Structural Dynamics
Department carries out approximately 30 project
per year for clients all over the world [7]. These
clients are mainly manufacturers of compressors
and pumps, engineering companies, oil and gas
industries, chemical and petrochemical industries.
The analyses are carried out according to the API
Standard 618, fourth edition of June 1995, design
approaches 1, 2 or 3 for reciprocating compressors
[1] and according to the API Standard 674 of June
1995, design approaches 1 and 2 [8] for plunger
pumps. For a full design 3 approach analysis two
steps are taken:
1. The acoustical analysis with PULSIM
2. The mechanical analysis with ANSYS
The results of the analyses, preferably performed
during the design stage of a system, are
recommendations to reduce pulsations, vibrations
and cyclic stress levels in the system to acceptable
levels.
The PULSIM program, combined with ANSYS, has
been proven to be a powerful tool for the design of
safe systems and optimal processes [2]. Moreover,
the PULSIM team acts as a partner during the
complete course of the analysis and even when the
system has been put into operation.

3.

A short description of the PULSIM

program

3.1
Basic equations
The following equations describe the flow through a
pipe, assuming that the flow velocity is constant in
the pipe cross section:
-

The continuity equation.

Eq(1)
The momentum equation.
Eq(2)
The effect of gravity and friction is taken into
account by the source term. Eq(3)
The energy equation.
Eq(4)

To complete the set of equations, a relation between

pressure, density & temperature, the ideal gas law is
used Eq(5).
The basic equations for the flow in pipe systems can
be simplified considerably, if the following
conditions are fulfilled:
- The fluctuations of pressure and density are
small, compared to their mean values.
- The (total) flow velocity is small compared to
the wave propagation speed (velocity of
sound).
- There is no heat exchange with the pipe wall.
- The fluidum can be described as an ideal gas.
- The acoustic wavelength is much larger than
the pipe diameter.
The simplified equations are called the acoustic
approximation. For this approximation 2 equations
remain:
- The mass balance.
Eq(6)
- The momentum balance.
Eq(7)
3.2

The analog simulation method

For the solution of the acoustic approximation of he

flow equations, the space variable x has to be made
discrete, This is equivalent to dividing a long pipe
into small sections. In this way the two partial
differential equations are converted into two
ordinary first order differential Eq(8) and Eq(9). In
these equations W represents the pressure loss due
friction and proportional to u|u|. For the analogue
solution this term is linearized, so that friction can
be modelled by a linear resistance.
Equation Eq(8) can be considered as a mass balance
for the pipe element, for which the electric analog is
a capacitor. Equation Eq(9) represents the impulse
balance for a pipe element, which can be
represented electrically by an inductor. In this way a
sequence of pipe elements can be represented by an
inductor capacitor network.

3.3

The digital simulation method

A method suitable for the solution of the one

dimensional (hyperbolic) flow equations on a digital
computer, is the method of characteristics. This
method, which is very popular in engineering, leads
to minimal computation time and thus is very
efficient. In this paper the method of characteristics
will not be given in detail. Instead, a qualitative
description will be given of the numerical solver
chosen by TNO-TPD. In the method of
characteristics both time and space variable x are
made discrete. The time step and the length of a
pipe element are coupled by the equation Eq(10).
This relation represents the so-called characteristic
directions in the x-t plane in case of the acoustic
approximation. A fixed grid of characteristics can
be drawn in the x-t plane. In case the flow velocity
cannot be neglected with respect to the wave
propagation speed, the characteristic directions
depend on the local flow velocity and the grid of
characteristics is not fixed. It is obvious that in that
case the solution will be far more complex. By
introducing so called characteristic variables
Eq(11) and Eq(12), the differential equations are
simplified to ordinary equations Eq(13), Eq(14) and
Eq(15), which are coupled by the friction term W.
Starting with a known initial condition, the values
of the characteristic variables for the next time step
can be calculated by solving equations Eq(13) and
Eq(14) for each point on the x-axis. Apart from the
calculation of the friction term W, the solution is
reduced to shifting the characteristic variables along
the characteristic directions, along the positive
and along the negative characteristic direction.
Besides, the value of at the right-hand boundary
and the value of at the left-hand boundary have to
be known for the next time step. These values are
called boundary values.
Apart from the method of solving the partial
differential equations the so called boundary
conditions, which are terminating elements at both
ends of a pipe section are equally important. The
boundary conditions describe for instance
components like a compressor cylinder, a Tee
connection, a reducer, a volume, a closed valve
(zero flow pulsations), a very large volume (zero
pressure pulsations) etc. In the PULSIM program, a
flexible data structure is available with which a flow
simulation model of any pipe system can be
defined. A complex pipe system consists of various
interconnected pipe sections. The model data
structure consists of tables with pipe sections, start
and end nodes of each pipe section and the
boundary condition at each node. The model size is
basically unlimited. A simple example is given in
chapter 4.2.1.

Equation 8:

Basic equations:
Equation 1:

u
+
=0
t
x
Equation 2:

u
u
p
+ u
= f
x
t
x

Equation 9:

d ( u )i pi pi +1
=
Wi
dt
x
Equation 10:

c = +

Equation 3:

f =g+

d pi
2 ( u)i ( u)i 1
= c
x
dt

2 fa
uu
D

Equation 4:

h h p u p
+u
=
+
q
t
x t
t

x
t

Equation 11:

= p + cu
Equation 12:

= p cu
Equation 13:

Equation 5:

i ( t + t ) = i 1( t ) Wi 1

p = R T

Equation 14:

Equation 6:

i (t + t ) = i +1 ( t ) Wi +1

p
2 u
+c
=0
t
x

Equation 15:

Equation 7:

Wi = W ( i , i )

u p 2 fa
+

u u g = 0
D
t x
Notation:
c
D
f
fa
h
p
q
t

Wave propagation speed

Pipe diameter
Volume force
Fanning friction factor
Specific enthalpy
Pressure
Heat flow
Time

T
u
x

Temperature
Flow velocity
Distance along pipe axes
Characteristic variables
Density
Isentropic constant
Friction Term

3.4

Presentation of results.

To understand the acoustical behaviour of the

system and to present the simulation results the
PULSIM program offers various options:
1.
The frequency response curves of the
system.
2.
Plots of the calculated peak to peak level of
pressure, flow and force at a selected node
or pipe section as a function of the velocity
of sound or compressor speed.
3.
Plots of the peak to peak pulsation level of a
selected compressor harmonic at a certain
condition as a function of the location in the
pipe system.
4.
Plots of the time function of pressure, flow
or force during the revolution time of the
compressor.
5.
Tables with the calculated pulsation level at
selected points for various operating
conditions.
Examples of the plots and tables are given in
chapter 4.
4.

Steps in a pulsation analysis

4.1

Collecting data for the analysis (Kick

off meeting)

4.1.1

Scope of the analysis.

The scope of the analysis has to be defined very

well. In case more than one pulsation source is
present in the piping, it has to be decided for which
source or combination of sources the pulsations
have to be calculated. To find the maximum
possible worst-case pulsations and pulsationinduced forces, all possible combinations have to be
analyzed. For each source the operating conditions
and the type of capacity control play an important
role for the resulting pulsations. For instance for a
reciprocating compressor the following parameters
are very important:
- Compressor speed (fixed or variable).
- Number of stages and cylinders per stage.
- Single or parallel running compressors.
- Type of capacity control (suction valve
unloaders, reversed flow control, clearance
pockets)
- Pressure ratio (fixed or variable pressures in the
system).
- Gas composition. Fixed or variable.
In case many of the above mentioned parameters
vary significantly in practice, a selection of the
cases to be investigated will be made to limit the
calculations. To guarantee a safe operation for all

possible operating conditions much experience is

required to make an optimal selection.
4.1.2

In search for good boundary conditions.

As described in 3.3 the boundary conditions play a

very important role in the accuracy of the
simulation results. Good boundary conditions to
start or stop the simulation are:
- Large volumes. A volume can be considered to
be large in case the volume is 5 to 10 times a
well designed pulsation damper in the system
to be investigated. For instance the suction
K.O. drum should be 5 to 10 times the suction
pulsation damper.
- A closed valve. For instance the normally
closed relief or bypass valves. Also valves with
sufficient pressure drop (more than 5% of the
static pressure in the system) can be considered
as being acoustically closed.
- A large diameter change in the piping. For
instance in case the compressor piping is
connected to a header with considerable larger
diameter. A diameter ratio of 4 can be
considered to be sufficient. However, the
length of the piping with large diameter should
be sufficiently long.
- A very (infinitely) long pipe system. In an
infinitely long pipe system the pulsations,
generated by the compressor will not be
reflected and therefor reflections will not
increase the pulsations near the compressor. A
pipe system can be considered to be infinitely
long in case the length is more than 10 times
the wavelength of the lowest compressor
harmonic. For instance for an air (velocity of
sound 340 m/s) compressor with a compressor
speed of 300-rpm (lowest harmonic 5 Hz) the
wavelength is 68m. So the length should be
more than approximately 680 meters before the
length can be considered as infinitely long.
During the Kick off meeting Process Flow and
Instrumentation Diagrams are very helpful to define
the points to start and stop the simulation. Also
detailed data of process equipment like heat
exchangers or separators need to be collected.
Moreover, good insight in the process conditions
(pressures, temperatures, and densities) and
pressure losses of equipment and valves should be
gathered.

4.2
4.2.1

figures 8 through 10 the corresponding calculated

peak to peak levels as a function of the velocity of
sound are given.

The check of the pulsation dampers

The original pulsation damper

Instead of the formula given in the API Standard

618, paragraph 3.9.2.2.2, for the design of pulsation
dampers, PULSIM applies a different approach.
For the design of the damper, the cylinder internals
with a simple compressor valve model and a suction
and discharge damper are modelled. The suction
and discharge piping is replaced by infinitely long
lines. A sketch of such a model is given in figure 1.
The structure of a part of the PULSIM model is
shown in the pipe table of figure 2.
480
470

420

500

240
210

430
270

510

40

340

10

114001 *** PULSIM 1.11

***
Customer : Cetim
System
: example
Duty
: Natural Gas
Part
: Cylinder with pulsation dampers
Condition: 100% load

14-MAY-2001 11:28:32
SYSNR:
MODNR:
SPEED:
VOS :
PMEAN:

1
0
756.0 rpm
402.9 m/s
1600.0 kPa

Listing of pipes
---------------------------------------------------------------------------Pipe From-To
bound.
L
D
v
r
f
q
P
[m]
[m] [m/s] [kg/m3]
[-] [kg/hr]
[kPa]
---------------------------------------------------------------------------1
10- 420 RE-TE
0.470
0.070
403
13.2 0.004
-3165 1493.9
410
KW
0.270
419
KW
0.190
420
TE
0.010
2 420- 430 TE-Q0
0.350
0.330
403
13.2 0.004
0 1497.3
3 420- 440 TE-TE
1.750
0.330
403
13.2 0.004
-3164 1497.3
4 210- 470 RE-TE
0.470
0.070
403
13.2 0.004
-3057 1494.0
460
KW
0.270
469
KW
0.190
470
TE
0.010
5 470- 480 TE-Q0
0.350
0.330
403
13.2 0.004
0 1497.3
6 470- 440 TE-TE
1.750
0.330
403
13.2 0.004
-3057 1497.3
7 440- 510 TE-Z0
1.210
0.107
403
13.2 0.004
-6221 1522.6
490
OR
0.010
500
KW
0.200
510
Z0
1.000
8
10- 20 RE-TE
0.050
0.087
403
13.2 0.004
3165 1493.6
9
20- 40 TE-VO
0.250
0.110
403
13.2 0.004
1535 1493.6
30
KW
0.225
40
VO
0.025
10
40- 50 VO-CY
0.050
0.080
403
13.2 0.004
1535 1493.3
11
20- 70 TE-VO
0.250
0.110
403
13.2 0.004
1630 1493.6
60
KW
0.225
70
VO
0.025
12
70- 80 VO-CY
0.050
0.080
403
13.2 0.004
1630 1493.2
13
50- 140 CY-VO
0.050
0.080
465
28.7 0.004
1535 4885.1
14 140- 120 VO-TE
0.250
0.110
465
28.7 0.004
1535 4884.8
130
KW
0.025
120
TE
0.225
15
80- 170 CY-VO
0.050
0.080
465
28.7 0.004
1630 4885.3
16 170- 120 VO-TE
0.250
0.110
465
28.7 0.004
1630 4885.3
160
KW
0.225
120
TE
0.025
17 120- 110 TE-RE
0.050
0.087
465
28.7 0.004
3165 4884.7
18 210- 220 RE-TE
0.050
0.087
403
13.2 0.004
3057 1493.6
19 220- 240 TE-VO
0.250
0.110
403
13.2 0.004
1571 1493.6
230
KW
0.225
240
VO
0.025
20 240- 250 VO-CY
0.050
0.080
403
13.2 0.004
1571 1493.3
21 220- 270 TE-VO
0.250
0.110
403
13.2 0.004
1485 1493.6
260
KW
0.225
270
VO
0.025
22 270- 280 VO-CY
0.050
0.080
403
13.2 0.004
1485 1493.3
23 250- 340 CY-VO
0.050
0.080
465
28.7 0.004
1572 4887.9
24 340- 320 VO-TE
0.250
0.110
465
28.7 0.004
1572 4887.5
330
KW
0.025
320
TE
0.225
25 280- 370 CY-VO
0.050
0.080
465
28.7 0.004
1485 4887.8
26 370- 320 VO-TE
0.250
0.110
465
28.7 0.004
1485 4887.8
360
KW
0.225
320
TE
0.025
27 320- 310 TE-RE
0.050
0.087
465
28.7 0.004
3057 4887.5
----------------------------------------------------------------------------

310

70

Figure 2.

360
690

720

time

Discharge flow [kg/s]

spec

740

140

4
1.5

110
730

170
610

640

1
2
0.5
650

660

760

770

90

180

270

0
360

Crank angle [degrees]

Qm = 0.45 [kg/s] = 1637 [kg/hr]

Figure 1.

Figure 3.

In this model the PV diagram and the flow pulses

are calculated based on the actual suction and
discharge pressure and depend on the pulsations in
the cylinder passage.

In case special attention has to be paid on the

performance of the compressor valves, the
acoustical model is extended with a mechanical
model (2nd order system) of the compressor valves.
For each acoustical time step the response (lift) of
the compressor valve is calculated. For this type of
simulation many parameters are necessary.
Research, together with compressor and compressor
valve manufacturers, to optimise this type of
simulation is still necessary. It is clear that a poorly
designed compressor valve can have a strong effect
on the damper analysis and can generate high
frequency pulsations by itself.

The calculated time function of the discharge flow

pulse, generated by one cylinder side, is shown in
figure 3. Moreover the frequency analysis of the
pulse is given from the 1st to the 15th harmonic. It
is obvious that the interaction between cylinder,
cylinder passage and damper effects the shape of
the flow pulses. The PV diagram shows the effect of
the pulsations on the power consumption and
capacity. In figure 4 the calculated PV diagram is
given at a certain velocity of sound.
The generated flow pulses cause pressure pulsations
in the system. In figures 5 and 6 the plots of the
time function of the calculated pulsations near the
discharge compressor valves and at the outlet of the
discharge damper are given. In figure 7 the plot of
the time function of the calculated pulsation
induced force on the discharge damper is shown. In

In case the piping is replaced by an infinitely long

line, there is no contribution of reflections of the
pipe system to the calculated pulsation level and in
most cases the minimum pulsation level is
calculated in this way. Therefor, we take into
account a certain safety margin for the allowable
level of the calculated pulsation level at the line
connection of the damper.

In case there is only one pulsation suppression

device is attached to the pipe system, we design the
damper such that, with an infinitely long line, the
pulsations at the line connection are 80% of the
allowable level. In case more pulsation suppression
devices are attached to the pipe system, 70% of the
allowable level is applied. This approach has been
proven to be very practical. The check or design of
the pulsation damper can be carried out in an early
stage of the project as the piping lay out need not to
be known yet. Only in very exceptional cases it
appeared, during the investigation of the complete
system, that installation of an additional volume was
still necessary.

[bar]

PV diagram

[bar]

40
40

20

20

0.002
Volume [m3]

0.004

Consumed power = 77.67 [kW]

Adiabatic power = 66.26 [kW]

Figure 4 .
[%]

Recommendations

Pressure node 140 (VO)

10
5

[%]

[%]

20

15

10
5

-10

Figure 5 .

[%]

2
1

1.5

-1

0.5

-2

Pressure node 770 (Z0)

[%]
2

[%]

Pressure node 770 (Z0)

[%]
2

0.2

1.5

0.5

1.5

0.1

0
-0.1

0.5

-0.2

Pm = 4635 [kPa]
Harmonic 1-15
Ppp = 70 [kPa] = 1.51 [%]

Pm = 4624 [kPa]
Harmonic 1-15
Ppp = 19 [kPa] = 0.42 [%]

Figure 6 .

Figure 13.

Force section 1: 660- 740

[kN]
30

10

[kN]

Force section 2: 660- 740

[kN]

0.4

0.5

0.2

20

10

-10
-20

Figure 7 .

0.4
0.3

0.2

-0.2

0.1

-0.4

Harmonic 1-15

Fpp = 33.031 [kN]

[%pp]

For the system as shown in figure 1 it was

recommended to:
- Increase the diameter of the piping between the
cylinder.
- Install baffles in both dampers.
- Install orifice plates in the pipe sections
between the damper and the cylinder (close to
the damper).

Pressure node 140 (VO)

Figure 12.

0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8

[kN]

0.004

Pm = 4641 [kPa]
Harmonic 1-15
Ppp = 219 [kPa] = 4.71 [%]

Pm = 4650 [kPa]
Harmonic 1-15
Ppp = 991 [kPa] = 21.32 [%]

[%]

0.002
Volume [m3]

Figure 11.

In case the calculated pulsations or pulsation

induced
forces
exceed
allowable
levels,
modifications are investigated to reduce these
levels.
Simple acoustical design rules for improving
pulsation dampers are:
Cylinder connection:
- As short as possible and maximum diameter.
- Equal length between cylinder connection and
caps (minimum forces).
For dampers with more cylinders:
- Install baffle plate.
- Or extend cylinder connection in the damper to
the axial middle of the damper.
In most systems, the installation of orifice plates in
the pipe section between the cylinder and the
damper is effective. The optimal location is as close
as possible to the damper.

Consumed power = 69.57 [kW]

Adiabatic power = 66.84 [kW]

-5

4.2.2

PV diagram

0
Harmonic 1-15

Fpp = 0.802 [kN]

Figure 14.

Pressure node 140 (VO)

[%pp]
6

20

Pressure node 140 (VO)

10

-10

10

-10

10

Dev. [%]
Peak-to-peak and AutoHarmonic nr 2

Figure 8 .
[%pp]

Dev. [%]
Peak-to-peak and AutoHarmonic nr 5

Figure 15.

Pressure node 770 (Z0)

[%pp]

Pressure node 770 (Z0)

1.5

In figures 11 through 17 the simulation results are

shown for the recommended damper and can be
easily compared with the results of the original
damper (figures 4 through 10). The results show
clearly that the power consumption, the pulsation
levels and the pulsation-induced force on the
discharge damper have been reduced considerably.

1
1
0.5

0.5
0

-10

10

-10

10

Dev. [%]
Peak-to-peak and AutoHarmonic nr 2

Dev. [%]
Peak-to-peak and AutoHarmonic nr 2

Figure 9 .

Figure 16.

[kN pp]

[kN pp]

Force section 1: 660- 740

30

Force section 2: 660- 740

0.8
0.6

20

0.4
10
0

0.2
-10

10

-10

Peak-to-peak and AutoHarmonic nr 5

Figure 10.

10
Dev. [%]

Dev. [%]

Peak-to-peak and AutoHarmonic nr 1

Figure 17.

4.3 Investigation of the pipe system:

1700

Recommended orifice

Recommended orifice
1490

1710

1610

1600

1480

Recommended orifice
1730

1590
1230

1220

1090

1410

1250

1470
1100

1200
1460

1260

Recommended orifice
1120
1380

1270

1370
1280

1360

1290

1310
1340

1350

1320

Figure 18.
After the check or the redesign of the damper has
resulted in allowable pulsation levels and dynamic
forces, the acoustical model of the pipe system is
connected to the model of the damper. An example
of a pipe system is shown in figure 18. For the
analysis of the complete pipe system the results can
be presented in the same way as for the check of the
damper design. Some examples of the calculated
levels (time function of the pulsation levels for one
revolution) are shown in figures 19 and 20.

In the tables the actual peak to peak pulsation levels

can be given (see figure 22) or for each presentation
point a comparison can be made with the allowable
level. In case the calculated pulsation level exceeds
the allowable level, the ratio of the calculated and
allowable level is printed.
[%]

Pressure node 1250 (RE)

1.5

1
0.5

Because it is not useful to present a large number of

plots of the calculated pulsations throughout the
pipe systems, normally the simulation results of
pipe systems are presented in tables. A number of
presentation points in the pipe system are selected
and the calculated pulsations and forces at these
points are given for a range of the velocity of sound
or for the range of the compressor speed.

-0.5

[%]
5

4
3

-2

-1
Pm =
Ppp =

Pressure node 1280 (RE)

Another useful way of presenting the simulation

results, to understand the acoustical behavior of the
pipe system, is a plot as shown in figure 21. In this
plot the calculated peak to peak pulsation level is
given as a function of the location in the pipe
system between both dampers. With this type of
plot acoustical standing waves can be analyzed.

[%] [%]

0 -4

Pm =
Ppp =

420 [kPa]
Harmonic 1-20
10 [kPa] = 2.48 [%]

Figure 19.

0
420 [kPa]
Harmonic 1-20
27 [kPa] = 6.45 [%]

Figure 20.

Pressure pulsation [%pp]

8
6
4
2
0

0
Peak level plus AutoHarmonic 2

Figure 21.

10
Distance [m]

For calculated pulsations below the allowable level

the following symbols are used (see figure 23):
*
Calculated pulsation level between 0.9 and
1.1 times the allowable level
+
Calculated pulsation level between 0.5 and
0.9 times the allowable level
Calculated pulsation level between 0.1 and
0.5 times the allowable level
.
Calculated pulsation level below 0.1 times
the allowable level.
The figure behind the slash sign indicates the
dominant harmonic of the calculated pulsation
level. In this way, the tables give a clear overview at
which location and which velocity of sound (or
compressor speed) the pulsations exceed the
allowable level.
114001 *** PULSIM 1.11
Customer : Cetim
System
: example
Duty
: Design
Part
: Interstage I
Condition: 100% load

STANDARD

***

15-MAY-2001 11:52:48
SYSNR:
MODNR:
SPEED:
VOS :
PMEAN:

29
0
585.0 rpm
638.0 m/s
429.6 kPa

Pressure pulsations [%pp], based on system mean pressure

Variation of velocity of sound
--------------------------------------------------VOS [m/s]
558
590
622
654
686
718
DEV
[%] -12.5
-7.5
-2.5
2.5
7.5
12.5
--------------------------------------------------User defined allowable level of 14.00 [%pp]
1020
13/4
15/5
15/5
12/5
12/6
15/6
1060
13/4
15/5
15/5
12/5
12/6
15/6
1520
12/6 8.5/6 7.4/7 9.8/7 7.5/7 8.1/8
1560
12/6 8.4/6 7.5/7 9.9/7 7.4/7 8.3/8
According to API 618 June 1995, for compressor valves 6.3 [%pp]
1090
13/4
14/5
14/5
11/5
11/6
14/6
1590
According
1250
1260
1280
1330

10/6
to API
3.2/15
4.2/13
4.8/4
2.5/2

7.5/6 6.7/7 8.9/7 6.8/7

618 June 1995, for piping
3.3/13 2.9/5 2.5/2 2.7/15
3.4/13 2.5/15 3.0/15 3.3/15
2.8/13 4.3/5 6.5/5 6.9/2
3.0/2 4.6/2 7.5/2 9.7/2

7.4/8
2.4/6
2.6/17
6.4/2
9.7/2

1370
2.8/2 3.1/2 4.6/2 7.6/2
10/2
10/2
1410
4.2/13 5.1/13 5.7/13 7.7/2
10/2
10/2
1460
8.3/4 7.3/13 8.8/13 13/2
16/2
14/2
1470
5.0/4 3.4/13 4.9/5 6.3/5 6.7/2 6.7/2
1490
2.0/2 2.3/12 2.9/13 3.0/2 3.3/2 3.3/2
---------------------------------------------------

Figure 22.
114001 *** PULSIM 1.11
Customer : Cetim
System
: example
Duty
: Design
Part
: Interstage I
Condition: 100% load

Legend

STANDARD

***

15-MAY-2001 11:52:48
SYSNR:
MODNR:
SPEED:
VOS :
PMEAN:

| Symbol
Range
| Symbol
Range
| Symbol
Range
|
|
.
0.00 .. 0.10 |
0.10 .. 0.50 |
+
0.50 .. 0.90 |
|
*
0.90 .. 1.10 | VAC
vacuum | CAV
cavitation |

Pressure pulsations [comparison], based on system mean pressure

Variation of velocity of sound
--------------------------------------------------VOS [m/s]
558
590
622
654
686
718
DEV
[%] -12.5
-7.5
-2.5
2.5
7.5
12.5
--------------------------------------------------User defined allowable level of 14.00 [%pp]
1020
*
*
*
+
+
*
1060
*
*
*
+
+
*
1520
+
+
+
+
+
+
1560
+
+
+
+
+
+
According to API 618 June 1995, for compressor valves 6.3 [%pp]
1090
2.0/4 2.2/5 2.3/5 1.8/5 1.8/6 2.3/6
1590
According
1250
1260
1280
1330

29
0
585.0 rpm
638.0 m/s
429.6 kPa

1.7/6
to API
+
2.0/13
1.8/4
+

1.2/6
*
1.4/7
*
618 June 1995, for piping
+
+
+
1.6/13 1.7/15 1.8/15 1.4/15
1.1/13 +
1.1/5 1.6/2
+
1.2/2 2.0/2 2.8/2

1.2/8
+
1.2/17
1.6/2
2.8/2

1370
+
+
1.2/2 2.0/2 2.8/2 2.8/2
1410
*
1.4/13 *
*
1.4/2 1.4/2
1460
*
1.2/13 *
1.4/2 1.9/2 1.9/2
1470
1.5/4 1.5/13 1.2/5 1.1/5 1.4/2 1.4/2
1490
*
+
+
+
---------------------------------------------------

Figure 23.

In case more than one compressor is acting on a

pipe system PULSIM applies the following criterion
[1]:
- The pulsations caused by each compressor
separately should be within the allowable level.
- The pulsations caused by each compressor
separately are added at each point of the pipe
system and the sum of the pulsations of n
compressors running parallel should be within
n times the allowable level for each
compressor separately.
In case the pulsations exceed the allowable level,
modifications are investigated to achieve acceptable
levels.
Pulsations in the piping can be reduced in the
following ways:
- To install orifice plates or other damping
devices [3] which dampen the pulsations during
acoustical resonance conditions.
- To shift acoustical resonance conditions out of
the operation range. This can be achieved by
changing pipe length or partly changing the
diameter of the piping.
- To install more volume by increasing the pipe
diameter.
- To install an acoustical filter. This can be
achieved by installing a secondary volume near
the existing damper. With these two volumes
and the pipe section in between an acoustical
filter can be created with a low cut-off
frequency. The filter should be designed such
that the cut-off frequency is below one or two
time compressor speed (for single or double
acting cylinders).
Which modification is preferred depends on which
parameter is most critical in the system or which
modification is less expensive on the short or long
term. For instance orifice plates are not expensive
and can be installed easily in most cases, but the
additional pressure loss can reduce the capacity and
gives energy costs during the lifetime of the
compressor. So these matters need to be discussed
in detail with all parties (compressor manufacturer,
engineering company and end user).
In tables 24 and 25 the calculated pulsation levels
(actual levels and comparison) are given in case in
the system of figure 18 optimal orifice plates are
installed.
From the plots with the remaining pulsation induced
forces as a function of the velocity of sound (see
figures 26 and 27), the acoustical resonance
conditions (the peaks) can be selected and used as
worst-case conditions for the mechanical response
analysis (maximum vibrations and cyclic stress
levels).

114001 *** PULSIM 1.11 STANDARD

*** 15-MAY-2001 14:46:56
Customer : Cetim
SYSNR:
System
: example
MODNR:
Duty
: Design
SPEED:
Part
: Interstage I
VOS :
Condition: 100% load
PMEAN:
Orifices at cyl. connections dis. bore=123mm, suc. bore=111mm
Orifice at line connection dis. bore=90mm, suc. bore=87mm (dP=0.5%)

29
2
585.0 rpm
621.9 m/s
431.7 kPa

Pressure pulsations [%pp], based on system mean pressure

Variation of velocity of sound
--------------------------------------------------VOS [m/s]
544
575
606
637
669
700
DEV
[%] -12.5
-7.5
-2.5
2.5
7.5
12.5
--------------------------------------------------User defined allowable level of 14.00 [%pp]
1020
9.4/4 8.7/4 7.9/4 7.3/4 6.9/6 6.8/6
1060
9.2/4 8.5/4 7.9/4 7.3/5 6.8/6 6.8/6
1520
6.5/27 5.1/6 5.1/29 4.7/7 4.1/7 3.9/7
1560
6.0/27 4.8/6 4.7/29 4.1/7 4.0/7 3.9/8
According to API 618 June 1995, for compressor valves 6.3 [%pp]
1090
8.7/4 8.0/4 7.4/4 6.5/5 6.5/6 6.4/6
1590
According
1250
1260
1280
1330

4.2/6
to API
2.5/15
1.9/13
1.7/13
1.6/2

4.3/6 4.0/7 4.0/7 3.6/7

618 June 1995, for piping
2.3/13 2.0/17 2.0/17 2.0/15
2.0/13 1.5/13 1.7/15 1.7/15
1.6/13 1.4/13 1.8/15 2.0/15
1.7/2 1.8/2 2.1/2 2.2/2

3.4/7
1.9/19
1.3/17
2.1/2
2.3/2

1370
1.8/11 1.7/2 1.8/2 2.2/2 2.1/2 2.2/2
1410
2.9/11 2.1/12 2.0/12 3.0/13 2.2/2 2.3/2
1460
6.2/11 4.0/12 3.6/12 5.2/13 6.9/4 7.6/4
1470
1.8/11 1.3/12 1.2/12 1.7/13 1.3/2 1.4/2
1490
2.6/11 1.9/12 1.9/12 2.4/13 1.7/2 1.7/2
---------------------------------------------------

Figure 24.
114001 *** PULSIM 1.11 STANDARD
*** 15-MAY-2001 14:46:56
Customer : Cetim
SYSNR:
System
: example
MODNR:
Duty
: Design
SPEED:
Part
: Interstage I
VOS :
Condition: 100% load
PMEAN:
Orifices at cyl. connections dis. bore=123mm, suc. bore=111mm
Orifice at line connection dis. bore=90mm, suc. bore=87mm (dP=0.5%)
Legend

1590
+
+
+
According to API 618 June 1995,
1250
+
+
1260
*
*
1280
+
+
1330
-

+
+
for piping
+
+
+
+
+

+
+

1370
+
+
+
1410
+
+
1460
+
+
*
1.2/4
1470
+
+
1490
*
*
---------------------------------------------------

Figure 25.
Force section 4: 1220-1260

0.15
0.1
0.05
0

-10

10

Dev. [%]
Peak-to-peak and AutoHarmonic nr 13

Figure 26.
[kN pp]

Many types of flow measuring devices are sensitive

for flow pulsations and the flow pulsations will
cause a measuring error. For measuring orifice
plates, the (always-positive) measuring error can be
calculated. For other types of measuring devices,
for example turbine meters or vortex flow meters,
sometimes data of the measuring error caused by
pulsations is available at the manufacturer [4].

6.

The mechanical response analysis

6.1
Introduction
One of the objectives of the pulsation analysis is to
reduce the pulsation induced vibration forces to a
minimum. However, unallowable vibrations and
cyclic stresses can occur in case a mechanical
natural frequency is close to or coincides with a
frequency component of the pulsation induced
vibration forces, even in case the pulsation levels
itself are below the allowable level.
The objective of the mechanical response analysis is
therefore to check the design of the pipe system,
including the pipe supports and the construction on
which the supports are mounted, to make sure that
the vibration levels and the cyclic stresses are lower
than the allowable levels.
The mechanical response analysis is carried out
according to the API Standard 618, design approach
3, which also includes an analysis of the compressor
manifold system. However, the analysis of the
compressor manifold system is only included upon
request of the customer [5].

Force section 15: 1380-1470

0.2

0.1

Besides the pressure pulsations, also the flow

pulsations in each pipe section are a result of the
simulation. The level of the flow pulsations play an
important role at the following devices:
- Non return valves. In case of strong acoustical
resonances in a pipe system it is possible that
flow reversal occurs for a part of the revolution
time of the compressor. This will cause
hammering of non return valves.
- Flow measuring equipment.

| Symbol
Range
| Symbol
Range
| Symbol
Range
|
|
.
0.00 .. 0.10 |
0.10 .. 0.50 |
+
0.50 .. 0.90 |
|
*
0.90 .. 1.10 | VAC
vacuum | CAV
cavitation |

Pressure pulsations [comparison], based on system mean pressure

Variation of velocity of sound
--------------------------------------------------VOS [m/s]
544
575
606
637
669
700
DEV
[%] -12.5
-7.5
-2.5
2.5
7.5
12.5
--------------------------------------------------User defined allowable level of 14.00 [%pp]
1020
+
+
+
+
1060
+
+
+
+
1520
1560
According to API 618 June 1995, for compressor valves 6.3 [%pp]
1090
1.4/4 1.3/4 1.2/4
*
*
*

[kN pp]

29
2
585.0 rpm
621.9 m/s
431.7 kPa

Figure 27.
5.
Flow pulsations

-10

10
Dev. [%]

Peak-to-peak and AutoHarmonic nr 11

For the mechanical analysis the general-purpose

finite element program ANSYS is used. The pipe
system is divided into basic parts (called finite
elements) such as straight pipe sections, elbows, Tjoints, flanges, reducers, constructions on which
pipe supports are mounted, etc. The pipe supports
up to the supporting beam are mostly regarded as
infinitely stiff elements.

10

The elements which are used in the calculations are

so called Timoshenko beam elements (shear
deflection is included). Special attention is paid to
the flexibility of the nozzle/shell intersections. The
flexibility of these intersections is calculated with
the finite element program FE-PIPE and the results
are used in the mechanical response analysis with
the ANSYS program. In case the compressor
manifold is taken into account so called
substructures from the shell element models are
made of the crankcase, crosshead guide, distance
piece and cylinder to reduce computer time. An
example of such a model is shown in figure 28.
Figure 29.

Figure 28.
6.2 Step 1: Calculation of the mechanical natural
frequencies and mode shapes of the pipe
system
For these calculations, the pipe system is usually
split up into parts, which have no, or only small
interaction. When the compressor manifold is not
taken into account in the analysis, the cylinders are
taken as rigid points.
The model consists of the pipe system including the
supporting constructions on which the pipe supports
are mounted and models of the equipment
(dampers, coolers, and separators with their
supports). An example of a plot of such a system is
given in figure 29. This systems consists of all the
piping and equipment on the skid and the
supporting beams on the skid of a 4 stage
compressor. In figure 30 the plot of the model is
shown with all the locations where the piping is
restrained and in which direction. In the plot of
figure 31 the node numbers of the model are shown
of which the calculated results can be given.
At first, the lower mechanical natural frequencies
of these models are calculated to get a feeling of the
flexibility of the system and the mode shapes. For
this system the lower mechanical natural
frequencies are shown in figure 32.

Figure 30.

Figure 31.
114001 *** PULSIM 1.11
Customer : Cetim
System
: example
Duty
: Natural Gas
Part
: Skid piping
Condition: 100% load

***

List of the lower 25 natural frequencies

---------------Frequency
Number
[Hz]
---------------1
6.16
16
2
8.52
17
3
13.64
18
4
15.95
19
5
17.42
20
6
7
8
9
10

17.92
18.45
18.65
21.01
22.22

11
12
13
14
15

22.49
23.78
24.30
25.06
27.19

Figure 32.

21
22
23
24
25

22-MAR-2001 13:58:03
SYSNR:
MODNR:
SPEED:
VOS :
PMEAN:

27.28
27.93
28.46
29.06
29.70
30.35
31.56
32.40
32.50
34.01

20
0
585.0 rpm
548.9 m/s
199.0 kPa

11

6.3 Step 2: Calculation of the mechanical response

of the system
The pulsation forces are calculated with the
PULSIM program for the pipe system with the
modifications, which have been accepted by the
client. The mechanical response calculations are
carried out for all important acoustical resonance
conditions and the calculated amplitude and phase
of the pulsation forces for each acoustical resonance
condition are transferred automatically from the
PULSIM program to the ANSYS program. For the
calculation of the natural frequency an inaccuracy
of 20% is taken into account. Therefore, natural
frequencies, which deviate less than 20% from
one of the harmonic components of the pulsation
forces, will be excited by these components in the
response calculation. In figure 33 a plot of a
vibration mode of a natural frequency which is
excited by the pulsation-induced forces is shown.
For this vibration mode the calculated vibration
level exceeds the allowable level. The maximum
calculated vibration level (120mm/s peak to peak) is
indicated in the plot.

120 mm/s

Instead of presenting a lot of plots as shown in

figure 33 we also present the calculated vibration
and cyclic stress levels in tables. These tables have
the same layout and same symbols as the tables
used for the presentation of the results of the
acoustic analysis (See 4.3 and figures 34 and 35).
When the calculated vibration levels and cyclic
stresses exceed the allowable levels, modifications
are investigated to decrease the levels to allowable
levels. This can be achieved in many cases by
shifting the natural frequencies far enough from the
excitation frequency. Shifting the natural frequency
can be achieved by installing extra pipe supports or
by increasing the stiffness of the structures on which
the supports are mounted.

114001 *** PULSIM 1.11

Customer : Cetim
System
: example
Duty
: Natural Gas
Part
: Skid piping
Condition: 100% load

Legend

***

22-MAR-2001 22:27:58
SYSNR:
20
MODNR:
0
SPEED:
VOS :
PMEAN:

585.0 rpm
548.9 m/s
199.0 kPa

| Symbol
Range
| Symbol
Range
| Symbol
Range
|
|
.
0.00 .. 0.10 |
0.10 .. 0.50 |
+
0.50 .. 0.90 |
|
*
0.90 .. 1.10 |

Vibrations [comparison]
Variation of velocity of sound
------------------------------------------------------------------------------VOS [m/s]
467
489
505
525
541
549
562
583
609
631
DEV
[%] -15.0 -10.9
-8.0
-4.4
-1.5
0.0
2.3
6.2
10.9
15.0
------------------------------------------------------------------------------Nodes above 1.2 x constant allowable level of 80.00 [mm/s pp]
5715
+
1.2/2 1.6/2 1.7/2 1.6/2 1.3/2
+
5724
*
1.6/2 2.2/2 2.2/2 2.1/2 1.7/2 1.2/2
+
5732
*
1.6/2 2.3/2 2.3/2 2.2/2 1.8/2 1.2/2
+
5733
5734
5735

*
*
+

1.7/2
1.7/2
*

2.3/2
2.3/2
1.3/2

2.4/2
2.4/2
1.4/2

2.3/2
2.3/2
1.3/2

1.9/2
1.9/2
*

1.3/2
1.2/2
+

+
+
-

5742
*
1.5/2 2.1/2 2.1/2 2.0/2 1.7/2
*
+
5743
*
1.5/2 2.0/2 2.1/2 1.9/2 1.6/2
*
+
5744
+
1.4/2 1.9/2 1.9/2 1.8/2 1.5/2
*
------------------------------------------------------------------------------Stresses [comparison]
Variation of velocity of sound
--------------------------------------------------------------------------------VOS [m/s]
467
489
505
525
541
549
562
583
609
631
DEV
[%]
-15.0 -10.9
-8.0
-4.4
-1.5
0.0
2.3
6.2
10.9
15.0
--------------------------------------------------------------------------------Nodes above 0.26 x constant allowable level of 60.00 [N/mm2 pp]
2480
.
.
.
.
.
.
.
5400
.
+
.
---------------------------------------------------------------------------------

Figure 34.
114001 *** PULSIM 1.11
Customer : Cetim
System
: example
Duty
: Natural Gas
Part
: Skid piping
Condition: 100% load

Legend

***

22-MAR-2001 22:27:58
SYSNR:
20
MODNR:
0
SPEED:
VOS :
PMEAN:

585.0 rpm
548.9 m/s
199.0 kPa

| Symbol
Range
| Symbol
Range
| Symbol
Range
|
|
.
0.00 .. 0.10 |
0.10 .. 0.50 |
+
0.50 .. 0.90 |
|
*
0.90 .. 1.10 |

Vibrations [mm/s pp]

Variation of velocity of sound
------------------------------------------------------------------------------VOS [m/s]
467
489
505
525
541
549
562
583
609
631
DEV
[%] -15.0 -10.9
-8.0
-4.4
-1.5
0.0
2.3
6.2
10.9
15.0
------------------------------------------------------------------------------Nodes above 1.2 x constant allowable level of 80.00 [mm/s pp]
5715
20/2
62/2
96/2 131/2 137/2 129/2 107/2
70/2
35/2
13/2
5724
27/2
81/2 126/2 172/2 179/2 169/2 140/2
92/2
46/2
17/2
5732
28/2
85/2 132/2 180/2 188/2 177/2 146/2
96/2
49/2
18/2
5733
5734
5735

29/2
29/2
17/2

88/2
87/2
50/2

137/2
135/2
78/2

187/2
185/2
106/2

195/2
193/2
111/2

184/2
182/2
104/2

152/2
150/2
86/2

100/2
99/2
57/2

50/2
50/2
29/2

18/2
18/2
10/2

5742
26/2
77/2 120/2 164/2 171/2 161/2 133/2
88/2
44/2
16/2
5743
25/2
75/2 116/2 159/2 166/2 156/2 129/2
85/2
43/2
15/2
5744
23/2
70/2 109/2 149/2 155/2 146/2 121/2
79/2
40/2
14/2
------------------------------------------------------------------------------Stresses [N/mm2 pp]
Variation of velocity of sound
--------------------------------------------------------------------------------VOS [m/s]
467
489
505
525
541
549
562
583
609
631
DEV
[%]
-15.0 -10.9
-8.0
-4.4
-1.5
0.0
2.3
6.2
10.9
15.0
--------------------------------------------------------------------------------Nodes above 0.26 x constant allowable level of 60.00 [N/mm2 pp]
2480
17/3 6.3/3 4.1/3 3.6/2 3.5/3 3.6/3 3.8/3 4.0/4
12/4 4.9/4
5400
4.6/2
14/2
22/2
30/2
31/2
29/2
24/2
16/2 8.0/2 2.8/2
---------------------------------------------------------------------------------

Figure 35.
6.4 Allowable levels
6.4.1 Allowable cyclic stress
The calculated cyclic stresses include the stress
intensification factors for T-joints, flanges, elbows
and reducers and are taken according the piping
code ANSI B31.3. When the peak value of the
cyclic stresses exceeds the endurance limit of the
material, fatigue problems will occur.
Unless the client explicitly specifies another
criteria, an allowable cyclic stress according the
API Standard 618 of 179 N/mm2 pp will be used
for each excited natural frequency. This value has
been based on the endurance limit and is only valid
for carbon steel pipe with an operating temperature
below 371 C. All other stresses must be within the
applicable code limits.

Figure 33.
For welds an extra stress concentration factor of 3 is
used. However, because not all the field weld
locations are known during the mechanical response

12

analysis, the weld stress concentration factors are

not included in the stress calculation. Therefore, the
allowable level of 179 N/mm2 will be divided by
the stress concentration factor of the weld. An
allowable cyclic stress level of 60 N/mm2 will
therefore be used in the tables.
6.4.2 Allowable vibration levels
To make sure that equipment such as temperature
transmitters, pressure gauges, valves, flanges etc.
will not fail due to too high vibrations the vibration
levels are also calculated.
During the analysis the vibration levels are
calculated during coincidence of acoustical and
mechanical resonance conditions (worst case
conditions). Therefore, we apply an allowable level
of 80 mm/s pp during the analysis. In practice,
during measurements, we apply an allowable level
of 42 mm/s pp (15 mm/s RMS).
6.5 How vibration levels and cyclic stresses can be
reduced
When the vibration levels and/or the cyclic stresses
exceed the allowable level there are several possible
modifications to reduce the vibration levels and/or
cyclic stresses. Some possible modifications are
explained below.
Shifting resonances
An effective way to decrease the vibration levels
and/or cyclic stresses is to shift the mechanical
natural frequencies in such a way that they cannot
be excited by the pulsation forces.
This can be achieved by installing additional
supports, relocating of existing supports or by
stiffening the structures on which the supports are
mounted. It is also possible to shift the natural
frequencies by changing the diameter or the layout
of the pipe system. However, this will often have a
considerable effect on the pulsation levels and
pulsation induced forces so that the pulsation
analysis would have to be carried out again for the
modified layout.
Changing pipe properties
When the cyclic stresses only exceed the allowable
level at stress risers in the pipe system such as Tjoints, flanges, reducers, etc. it is possible to
increase the local stiffness at such points. The
increase in stiffness can be achieved by increasing
the pipe diameter and/or wall thickness. When the
stiffness is only increased locally this will have in
most cases a minor influence on the pulsations and
pulsation forces.

When extra supports are installed, or when the

stiffness of the structures on which the supports are
mounted is increased, the expansion stresses will be
probably increased. It is important that the
expansion stresses are always within the allowable
levels according to the applicable piping code.
So the static and dynamic aspects compete in the
design of a pipe system. This problem can be solved
by searching acceptable compromises. At locations
in the pipe system where the static and dynamic
designs compete, e.g. a so-called spring hold down
support can be used. When the friction force
between the spring hold down support and the
structure on which the support is mounted is equal
to the peak value of the pulsation force than the
spring hold down support restrains the dynamic
motion.
The expansion stresses are mostly very large in
comparison with the friction force and therefore the
spring hold down support is free for the thermal
motion.
7.
Other dynamic
investigated with PULSIM

flow

phenomena

Besides the pulsation analysis according the API

Standard 618, design approaches 1, 2 and 3 and
according appendix M of this Standard, points M1
through M 11, the PULSIM program has been
extended for the investigation of the following kind
of dynamic flow phenomena:
- PULLIQ. This is a special version of the
PULSIM program for the investigation of pipe
systems of reciprocating pumps. This analysis
is carried out according the API Standard 674.
In this analysis the pipe elevation (static
pressure) and the static pressure loss in the
piping and equipment play an important role to
investigate if cavitation can occur.
- PULTRAN. This is a special version of the
PULSIM program for the investigation of high
pressure amplitude transients. These high
amplitude transients (shock waves) can for
instance occur in case an emergency valve is
opened within a very short time. During such a
transient the pressure variation is high and the
density cannot longer be considered constant
and in PULTRAN the density variation is taken
into account.
- The investigation of so called FIPS (Flow
induced pulsations) [6]. These pulsations are
generated by unsteady separation of the flow at
sharp edges, for instance at valves, orifice
plates, tee connections and sharp bends. Strong
pulsations can occur when the unsteady flow
separation is coupled with acoustical
resonances in the pipe system. In case the
source (frequency and amplitude) can be

13

predicted with sufficient accuracy and is

implemented in the PULSIM model, the
pulsations thoughout the system can be
calculated.
9.

10.

References

[1]

Eijk, A., Smeulers, J.P.M. (TNO-TPD),

Blodgett, L.E., Smalley, A.J. (SwRI),
Improvements and Extensions to API 618
Related to Pulsation and Mechanical
Response Studies, 1st EFRC Symposium,
Dresden, Germany, 4. 5. November
1999.

[2]

Eijk, A., Egas, G. (TNO-TPD), Effective

Combination of On-Site Measurements
and Simulations for a Reciprocating
Compressor System, 2nd EFRC
Symposium, The Hague, Netherlands, 17.
18. May 2001.

[3]

Peters, M.C.A.M. (TNO-TPD),

Evaluation of low frequency pulsation
damping devices, 2nd EFRC Symposium,
The Hague, Netherlands, 17. 18. May
2001.

[4]

Peters, M.C.A.M., van Bokhorst, E.,

Limpens, C.H.L. (TNO-TPD), The
impact of flow pulsations on vortex flow
meters, Flomeko 98, Sweden, June 1998.

[5]

Eijk, A., Egas, G., Smeulers, J.P.M. (TNOTPD), Cost-effective and detailed
modelling of compressor manifold
vibrations, ASME Pressure Vessels and
Piping Conference, Montreal, Canada, 21.
26. July 1996, PVP Vol. 328, pp. 415
425.

[6]

Peters, M.C.A.M., van Bokhorst, E. (TNOTPD), Flow-induced pulsations in pipe

systems with closed branches, impact of
flow direction., Flow Induced Vibrations
Symposium, Lucern, Switzerland, 1. June
2000.

[7]

van Bokhorst, E. (TNO-TPD),

Prvention des problmes de pulsations
et de vibrations, Achema 2000, Frankfurt,
Germany, 25. May 2000.

[8]

Eijk, A. (TNO-TPD), Improvements and

Extensions to API 674 Related to Pulsation
and Mechanical Response Studies,
VDMA Pumpentagung, Karlsruhe,
Germany, October 2000.

Further improvements

To improve the PULSIM package, continuous

research and development is necessary. If possible
the R&D is carried out together with our clients. At
this moment, the following developments are
ongoing within the flow and structural dynamics
department of TNO.
- The interaction of compressor valve dynamics
and pipe system. An improved valve model in
PULSIM can lead to:
better
compressor
performance
predictions, both in power consumption
and capacity,
better prediction high frequency pressure
fluctuations caused by valve instabilities
a more detailed pulsation analysis which
can help the manufacturers in the design of
pulsation dampers and the selection of
compressor valves.
- The development of accurate source models for
centrifugal
compressors
and
pumps.
Increasingly, pressure pulsations are observed
in process installations with centrifugal
compressors and pumps caused by an
interaction with the dynamics of the pipe
system. A dynamic source model can lead to
accurate predictions of the stability of these
systems.
- Source models of flow-induced pulsations.
These can be caused by unsteady flow
separation for example at T-joints, bends and
valves. In contrast to external pulsation sources
such as reciprocating compressors the internal
sources are influenced by the flow itself. The
geometry and the gas velocity according to a
specific Strouhal number determine the
pulsation frequency.
- New models for pulsation damping. Apart from
the standard orifice plates, used for the
reduction of low frequency pulsations, new
types of dampers are now available, such as
perforated plates. Experimental validation has
been carried out to evaluate the performance of
various types of damperplates for a wide
frequency range [3].