Wet Gas Compression

Wet Gas Compression
Impeller Rig
Siren Carstens Amundsen
Master of Science in Energy and Environment

Submission date: July 2009
Supervisor: Lars Erik Bakken, EPT
Norwegian University of Science and Technology

Department of Energy and Process Engineering
Problem Description
The overall scope of the thesis is to document compressor performance and stability for single-
and multiphase compression.
Assignment given: 09. February 2009

Supervisor: Lars Erik Bakken, EPT
Preface
This report contains the work performed on my master thesis at the Norwegian
University of Science and Technology, Department of Energy and Process Engineering.
The work was conducted during the spring of 2009.
I would like to thank Professor Lars E. Bakken, PhD yvind Hundseid and PhD graduate
student Trond G. Grner for valuable guidance throughout my work.
Trondheim, July 6th 2009
_________________________________________________________
Siren Carstens Amundsen
i
ii
Abstract
Wet gas compression technology is of great value to the oil and gas industry for boosting
of unprocessed well stream and to reduce investment costs related to equipment and
personnel. The growing interest in wet gas compression leads to a general request for
accurate performance calculation procedures and proper measurement techniques for
multiphase flow metering in compressors.
An impeller rig for examination of single-phase and multiphase performance and

aerodynamic stability is under construction at the test facility at NTNU. The construction
of the compressor rig is behind time due to late deliveries of the compressor components
and instrumentation. The performance calculations are therefore based upon one
compressor test conducted with dry gas at part-load.
The thermodynamic equation of state for ambient air is verified to be consistent with the
ideal gas law in the compressor pressure and temperature range. The calculated
polytropic performance is calculated with ideal gas assumptions and compared to values
estimated by PRO/II. By analyzing the results the sensitivity of the calculation
procedures is identified and the suitability for the ideal polytropic performance
calculations is validated for the actual compressor test and operating range.
A sensitivity analysis is conducted in order to determine the effect of measurement

uncertainties on performance calculations. Due to the low pressures involved for the
compressor test, the performance calculation procedures are highly sensitive to
uncertainties in the pressure measurements. Uncertainties in the temperature
measurements will only slightly influence the polytropic head, but have great influence
on the polytropic efficiency.
The efficiency and operating range of a compressor are constrained by aerodynamic

instabilities. This thesis describes the different flow phenomena associated with
compressor instability and presents recommendations for suitable instrumentation and
measuring techniques. Various visualization techniques are in addition evaluated to
determine the suitability for multiphase compressors.
Dynamic pressure transducers installed in the inlet and discharge piping are
recommended for detection of pressure pulsation throughout the compressor system.
Unsteady internal pressure measurements can be obtained from circumferentially
distributed pressure transducers at various locations within the compressor components.
Vibration probes installed at each end of the rotor are recommended for the vibration
measurements. By analyzing the frequency spectrum for the pressure fluctuation and
radial vibrations one can identify the type of instability phenomenon that occur. Laser
measurement techniques are recommended for the flow visualization in order to obtain
information on the main features of the multiphase flow field.
iii
iv
Abstract in Norwegian
Sammendrag
Vtgass kompresjon er en relativt ny teknologi som olje og gassprodusentene nsker
utvikle for kunne ke trykksttten til eksisterende felt. Trykksttte til produktive
brnner gir direkte fortjeneste ved akselerert produksjon og ved redusering av utstyr og
bemanning. En stadig kende interesse for vtgass teknologi har resultert i et behov for
nyaktige beregninger av ytelse og egnede mleteknikker for flerfase strmning i
kompressorer.
Eksperimentell forsksrigg for underskelse av ytelse og aerodynamisk stabilitet for en-

og flerfase kompresjon er under oppfring ved NTNU. Grunnet forsinkelser i leveranser
og oppfringen av testriggen er det kun gjennomfrt testing for trr gass.
Ytelsesberegningene er dermed basert p mlinger for kun en test for trr gass med
dellast.
Tilstandsligningen for atmosfrisk luft er i overensstemmelse med den ideelle gass lov i
operasjonsomrdet til kompressoren. Kompressorytelsen er dermed beregnet med ideell
polytropisk analyse og sammenlignet med verdier estimert ved simuleringer i PRO/II.
Ved analyse av resultatene er sensitiviteten til ytelsesberegningene dokumentert. Bruk av
ideell polytropisk analyse er validert for kompressor operasjonsrdet.
Sensitivitet med hensyn til mlenyaktighet er inkludert i analysene. Grunnet lave trykk
for kompressortesten er ytelsesberegningene sterkt sensitive for usikkerhet i
trykkmlingene. Unyaktighet i temperaturmlingene vil i liten grad pvirke den
polytropiske lftehyden, men vil ha sterk innvirkning p den polytropiske
virkningsgraden.
Kompressor ytelse og dets operasjonsomrde er begrenset av aerodynamisk ustabilitet i

maskinen. De forskjellige strmningsfenomenene assosiert med aerodynamisk ustabilitet
blir beskrevet i denne oppgaven med pflgende anbefalinger av egnede mlemetoder og
instrumentering. Ulike visualiseringsteknikker er i tilegg evaluert for bruk ved vtgass
kompresjon.
I tilegg til stasjonre mlinger i innlps- og utlpsrr, kan mlinger av trykk

fluktuasjoner implementeres for detektering av aerodynamisk ustabilitet. Dynamiske
trykk transmittere med hy respons anbefales for slike mlinger. Vibrasjonsflere p hver
side av rotoren anbefales for mlinger av radiell vibrasjon. Ved analyse av frekvens
spekteret for trykk- og vibrasjonsfluktuasjoner kan aerodynamisk ustabilitet
dokumenteres. Lasermlinger er anbefalt for visualisering av strmningen.
v
vi
Contents
Preface.................................................................................................................................. i
Abstract .............................................................................................................................. iii
Abstract in Norwegian ........................................................................................................ v
List of Figures .................................................................................................................... ix
List of Tables ..................................................................................................................... xi
Nomenclature................................................................................................................... xiii
1. Introduction..................................................................................................................... 1
1.1 Background ............................................................................................................... 1
1.2 Scope of Thesis ......................................................................................................... 2
1.3 Limitations and Challenges....................................................................................... 2
1.4 Report Structure ........................................................................................................ 2
2. Performance Analysis ..................................................................................................... 5
2.1 Polytropic Analysis................................................................................................... 5
2.2 Schultz Polytropic Analysis...................................................................................... 6
2.3 Wet Gas Performance Analysis ................................................................................ 7
2.4 Equation of State....................................................................................................... 8
2.5 Conclusion Chapter 2.............................................................................................. 10
3. Test Preparation ............................................................................................................ 11
3.1 Test Standards......................................................................................................... 11
3.2 Test Parameters....................................................................................................... 11
3.3 Test Stability ........................................................................................................... 13
3.4 Test Points............................................................................................................... 14
3.5 Test Uncertainty...................................................................................................... 15
4. Compressor Test Facility .............................................................................................. 19
4.1 Impeller Rig ............................................................................................................ 19
4.2 Piping Configuration............................................................................................... 21
4.3 Apparatus and Instrumentation ............................................................................... 22
4.4 Data Acquisition System......................................................................................... 26
5. Compressor Test ........................................................................................................... 27
5.1 Test Matrix.............................................................................................................. 27
5.2 Performance Procedure........................................................................................... 28
5.3 Generation of Performance Curves from Recorded Data Points ............................ 34
5.4 Affinity Laws .......................................................................................................... 37
5.5 Test Stability ........................................................................................................... 39
5.6 Static Measurement Uncertainty............................................................................. 39
5.7 Wet Gas Predictions................................................................................................ 40
6. Measurement Sensitivity............................................................................................... 45
6.1 Case 1...................................................................................................................... 45
6.2 Case 2...................................................................................................................... 49
6.3 Wet Gas Considerations.......................................................................................... 50
vii
7. Aerodynamic Instability................................................................................................ 53
7.1 Performance Characteristics ................................................................................... 53
7.2 Instability Mechanisms ........................................................................................... 54
7.3 Wet Gas Impact on Compressor Stability............................................................... 57
7.4 Test Procedure ........................................................................................................ 58
7.5 Instrumentation ....................................................................................................... 59
7.6 Pressure Characteristic Analysis............................................................................. 60
7.7 Frequency Spectrum Analysis ................................................................................ 61
8. Multiphase Flow Measurements and Visualization Techniques .................................. 65
8.1 Direct Visualization ................................................................................................ 65
8.2 Laser Measurement Techniques ............................................................................. 67
8.3 Pressure Sensitive Paint Measurement Technique ................................................. 70
8.4 CFD......................................................................................................................... 71
9. Conclusion .................................................................................................................... 73
10. Recommendations for Further Work .......................................................................... 75
Bibliography ..................................................................................................................... 77
Appendix A Performance Analysis ............................................................................... 81
A.1 Polytropic Analysis ................................................................................................ 81
A.2 Schultz Polytropic Analysis ................................................................................... 82
A.3 Polytropic Exponent............................................................................................... 84
Appendix B Verification of Discharge Coefficient and Expansibility Factor............... 85
Appendix C PRO/II Setup ............................................................................................. 89
Appendix D Schultz Generalized Compressibility Charts............................................. 91
Appendix E Valid Test Points........................................................................................ 93
Appendix F Computation of Non-Dimensional Head and Flow Coefficients............... 95
Appendix G Measurement Sensitivity Analysis ............................................................ 97
G.1 Measurement sensitivity for case 1........................................................................ 97
G.2 Measurement sensitivity for case 2...................................................................... 101
viii
List of Figures
Figure 3.1: Illustration of measurement errors [11].......................................................... 16
Figure 4.1: Test compressor.............................................................................................. 19

Figure 4.2: Test impeller................................................................................................... 20
Figure 4.3: NTNU test rig................................................................................................. 20
Figure 4.4: Inlet piping and instrumentation..................................................................... 21
Figure 4.5: Discharge piping and instrumentation............................................................ 22
Figure 4.6: APCE-2000 pressure transmitter.................................................................... 23
Figure 4.7: Pt 100 sensor .................................................................................................. 23
Figure 4.8: Pressure and temperature sensors at inlet....................................................... 24
Figure 4.9: Pressure and temperature sensors at outlet..................................................... 24
Figure 4.10: HBM T-12 digital torque transducer ............................................................ 25
Figure 4.11: BSD-MODULFLEX coupling system ......................................................... 25
Figure 5.1: Polytropic head curve..................................................................................... 35

Figure 5.2: Pressure ratio with varying volume flow ....................................................... 36
Figure 5.3: Polytropic head versus flow coefficient performance curve .......................... 37
Figure 5.4: Polytropic head curves given by the affinity laws ......................................... 38
Figure 5.5: Pressure rise with different GVF [15] ............................................................ 40
Figure 5.6: Efficiency with different GVF [15]................................................................ 40
Figure 5.7: Speed of sound with different GVF [7].......................................................... 42
Figure 6.1: Deviation in polytropic head with variation in inlet pressure ........................ 46
Figure 6.2: Deviation in polytropic head with variation in outlet pressure ...................... 46
Figure 6.3: Deviation in polytropic head with variation in inlet temperature .................. 48
Figure 6.4: Deviation in polytropic head with variation in outlet temperature ................ 48
Figure 7.1: Typical centrifugal compressor characteristic................................................ 53

Figure 7.2: Compressor characteristic with surge cycle ................................................... 55
Figure 7.3: Rotating stall in a centrifugal impeller [17] ................................................... 56
Figure 7.4: Flow through a vaneless diffuser [17]............................................................ 57
Figure 7.5: Measurement locations for internal instrumentation [24] .............................. 59
Figure 7.6: Circumferential measuring positions in the diffuser [24] .............................. 60
Figure 7.7: Stall and surge pressure characteristics at compressor inlet/discharge .......... 60
Figure 7.8: Pressure fluctuations with time [25]............................................................... 61
Figure 8.1: Locations for transparent windows ................................................................ 66

Figure 8.2: Test section of a diffuser with an optical window [28].................................. 67
Figure 8.3: PIV measurement system [29] ....................................................................... 69
Figure 8.4: PSP measurement system [29] ....................................................................... 70
Figure 8.5: The schematic PSP measurement................................................................... 71
ix
x
List of Tables
Table 3.1: Stability of compressor test [9]........................................................................ 13

Table 3.2: Permissible fluctuations of test readings [11].................................................. 14
Table 4.1: Impeller test rig instrumentation list............................................................. 22
Table 5.1: Initial test matrix at part load........................................................................... 27

Table 5.2: Measured and calculated test parameters at inlet with Z=1............................. 29
Table 5.3: Measured and calculated test parameters at discharge with Z=1 .................... 29
Table 5.4: Comparison of the polytropic exponent .......................................................... 30
Table 5.5: Comparison of the polytropic head.................................................................. 30
Table 5.6: Deviation between the polytropic exponents for various pressure ratios ........ 31
Table 5.7: Variation in isentropic exponent with varying pressure ratio.......................... 32
Table 5.8: Variation in polytropic efficiency with varying polytropic exponent ............. 32
Table 5.9: Sensitivity of polytropic efficiency with varying polytropic exponent ........... 32
Table 5.10: Variation in polytropic head with different equation of state........................ 33
Table 5.11: Operating data for compressor test ................................................................ 34
Table 5.12: Measured test parameters and calculated test results .................................... 35
Table 5.13: Predicted performance given by the affinity laws ......................................... 38
Table 5.14: Test stability................................................................................................... 39
Table 5.15: Mach numbers at inlet and outlet................................................................... 39
Table 6.1: Reference test parameters Case 1 ................................................................. 45

Table 6.2: Test point constants ......................................................................................... 45
Table 6.3: Sensitivity of polytropic head with varying pressure Case 1 ....................... 47
Table 6.4: Sensitivity of polytropic head with varying temperature Case 1.................. 49
Table 6.5: Reference test parameters Case 2 ................................................................. 49
Table 6.6: Sensitivity of polytropic head with varying pressure Case 2 ....................... 49
Table 6.7: Sensitivity of polytropic head with varying temperature Case 2.................. 50
Table B.1: Discharge coefficients for orifices .................................................................. 86

Table B.2: Expansibility factors for orifices..................................................................... 87
Table F.1: Calculation of performance parameters at test condition ................................ 95
Table G.1: Measurement sensitivity with varying inlet pressure Case 1....................... 97
Table G.2: Measurement sensitivity with varying outlet pressure Case 1..................... 98
Table G.3: Measurement sensitivity with varying inlet temperature Case 1................. 99
Table G.4: Measurement sensitivity with varying outlet temperature Case 1............. 100
Table G.5: Measurement sensitivity with varying inlet pressure Case 2..................... 101
Table G.6: Measurement sensitivity with varying outlet pressure Case 2................... 102
Table G.7: Measurement sensitivity with varying inlet temperature Case 2............... 103
Table G.8: Measurement sensitivity with varying outlet temperature Case 2............. 104
xi
xii
Nomenclature
Symbols
A Area [m2]
a Speed of sound [m/s]
C Discharge coefficient [-]
D Diameter [m]
f Schultz correction factor
H Head [m]
h Enthalpy [J/kg]
I Intensity [W/sr]
i Incidence angle []
Ma Mach number [-]
Mw Molecular weight [kg/kmol]
m Mass flow rate [kg/s]
N Rotational speed [rpm]
n Polytropic exponent [-]
P Power [W]
p Pressure [Pa]
Q Volume flow [m3/s]
R Gas constant [J/kg K]
Re Reynolds number [-]
s Entropy [J/kg K]
T Temperature [K]
U Tangential velocity [m/s]
V Relative velocity [m/s]
v Specific volume [m3/kg]
X Schultz compressibility function
Y Schultz compressibility function
Y Specific work [J/kg]
Z Compressibility factor [-]
Gas volume fraction [-]

Beta ratio [-]
Expansibility factor [-]
Isentropic exponent [-]
Efficiency [-]
Density [kg/m3]
Dynamic viscosity [cP]
Torque [N*m]
Flow coefficient [-]
Head coefficient [-]
Angular velocity [1/s]
xiii
Subscript
1 Inlet, suction state

2 Outlet, discharge state
c Critical
g Gas
l Liquid
m Mechanical
max Maximum
min Minimum
p Polytropic
ref Reference
S Schultz
s Isentropic
v Volume corrected
Abbreviations Description
ASME American Society of Mechanical Engineers

CCD Charge-Coupled Device
CFD Computational Fluid Dynamics
EOS Equation of State
FFT Fast Fourier Transformation
GVF Gas Volume Fraction
ISO International Organization for Standardization
L2F Laser two-focus velocimetry
LDV Laser Doppler Velocimetry
NTNU Norwegian University of Science and Technology
PIV Particle Image Velocimetry
PSP Pressure Sensitive Paint
PTC Performance Test Code
SI International System of Units
SM Surge Margin
Equations of State
BWR Benedict-Webb-Rubin
BWRS Benedict-Webb-Rubin-Starling
GERG University of Bochum (ISO Committee)
LKP Lee-Kesler-Plocker
PR Peng-Robinson
RK Redlich-Kwong
SRK Soave-Redlich-Kwong
Test rig: the total test assembly including all instrumentation.
xiv
Chapter 1
1. Introduction
Chapter 1 explains the background and motivation for the work presented. In addition,
the scope of the thesis, limitations and challenges related are presented.
1.1 Background
An experimental impeller rig for examination of single-phase and multiphase

performance and aerodynamic stability is under construction at the test facility at NTNU.
Accurate measuring results depend on the selection and implementation of proper
instrumentation as well as the applicability of the selected performance procedure for the
actual compressor test. The instrumentation and performance procedures must satisfy
established standards for compressor performance testing in addition to various
requirements for detection of aerodynamic instability.
Various applicable performance test codes are available for dry gas compressor testing.
ASME PTC 10 [1] provides specific guidelines for accurate measuring procedures in
addition to correct installation and location of various measuring devices.
Multiphase flows in compressors have complicated characteristics including interfacial

interactions and relative movement between phases. Determination of actual fluid and
thermodynamic properties is a challenge in multiphase compressors due to this phase
exchange. The instrumentation and measurement techniques employed for single-phase
compression may have insufficient accuracy due to the liquid introduced in wet gas
compression.
Measurement and visualization techniques with careful control of the flow distribution
are desired in wet gas compression to examine the multiphase flow effects under various
conditions in the compressor. Most of the compressor testing described in literature is
performed utilizing single-phase fluid. Techniques commonly used for visualization of
single-phase flow may be difficult to implement in multiphase compressors due to the
inherent non-homogeneity of the flow field.
1
1.2 Scope of Thesis
The overall scope of the thesis is to document compressor performance and stability for
single- and multiphase compression. An extensive literature study forms the foundation
for planning and conducting single- and multiphase compressor tests. Part of this
literature study was presented in the previous work by the author. [2]
The effect of measurement uncertainty on performance calculations is examined to

determine the measurement sensitivity.
Various measurement and visualization techniques are investigated to determine the

suitability for multiphase compressor testing. By achieving increased understanding of
the multiphase effects existing in wet gas compression and the phenomena associated
with aerodynamic instability, one can determine an appropriate instrumentation setup and
suitable measuring techniques.
1.3 Limitations and Challenges
The construction of the compressor rig was behind time due to late deliveries of the
compressor components and instrumentation. The performance calculations are therefore
based upon one compressor test conducted with dry gas at part-load. The results can
nevertheless be utilized to validate the compressor behavior at the test operating
condition.
Instrumentation for detection of compressor instabilities and the injection module for wet
gas testing are not yet implemented. Documentation of compressor performance and
stability for multiphase compression are therefore discarded from the work presented.
Recommended methods, procedures and instrumentation for wet gas is instead presented
for future wet gas testing.
1.4 Report Structure
Chapter 2 describes the relevant performance procedures needed to document and

validate the performance calculations for the compressor rig.
Chapter 3 presents the theoretical foundation for the test preparation based on various
applicable standards and available literature.
Chapter 4 describes the NTNU test facility and documents the experimental
instrumentation set up.
Chapter 5 describes the compressor testing and presents the results concluded from the
performance analysis.
2
Chapter 6 shows the effect of measurement uncertainties on the performance calculations
for the compressor test.
Chapter 7 gives the theoretical foundation for detection of aerodynamic instabilities in

centrifugal compressors. Recommendations for suitable instrumentation are in addition
presented.
Chapter 8 presents various measurement and visualization techniques employed in single-

phase and multiphase systems and determines the suitability for multiphase compression.
Chapter 9 concludes the results that can be drawn based on the work presented.
Chapter 10 contains recommendations for further work.
3
4
Chapter 2
2. Performance Analysis
An accurate method for performance calculations is important to ensure a correct

evaluation of the centrifugal compressor. Performance calculations in ASME PTC 10 [1]
are based on dry gas compression. In default of correct standards and methods for wet gas
compression the test results may deviate considerably when liquid is present in the gas.
The specific work of a compressor is dependent on the compression process and the
compressed gas. For comparison, the ideal isentropic compression process can be
utilized. By introducing the isentropic condition (pv=constant), the specific isentropic
work for a compressor is shown as in equation (2.1).
1

p2
Ys = Z1 RT1 1 (2.1)
1 p1

The term head is generally employed when referring to the specific work done by a
compressor. The compressor actual head, equation (2.2), describe the total change in
enthalpy for the compression process. The relationship between pressure, temperature
and enthalpy are determined by utilizing an appropriate equation of state. The actual head
remains constant independently of the given compression process.
H = h2 h1 = h( p2 , T2 ) h( p1 , T1 ) (2.2)
2.1 Polytropic Analysis
The thermodynamic evaluation of centrifugal compressors is generally based on the

polytropic procedure. Identical compressors operating at different suction pressures will
have variation in isentropic efficiencies due to the deviation in the isobars [(dh/ds)p=T].
This thermodynamic characteristic is taken into account when assuming a polytropic
process. The deduction of the polytropic head is based on the assumption of a constant
polytropic exponent along the compression path. The compressor polytropic head is
given in equation (2.3).
5
n 1

n p2 n
H p = Z1 RT1 1 (2.3)
n 1
1p

The polytropic efficiency is defined as the relationship between polytropic and actual
head as shown in equation (2.4).
Hp
p = (2.4)
H
Equation (2.5) shows the relationship between the polytropic head, efficiency and
compressor power requirement. The mechanical efficiency, m, is usually estimated to
about 97-98.5 %. [3]
1Q1 H p
P= (2.5)
m p
The definitions and equations utilized in the polytropic calculation procedure are given in
Appendix A.1.
2.2 Schultz Polytropic Analysis
At high pressures and temperatures, the ideal gas behaviour is not valid due to changes in
fluid properties. Both ASME PTC 10 [1] and ISO 5389 [4] have implemented the John
M. Schultz polytropic procedure [5] for thermodynamic performance evaluation of a
compressor. Real gas behaviour is taken into account when utilizing the Schultz
procedure.
The procedure assumes a polytropic compression path based on averaged gas properties
of inlet and outlet conditions. Schultz introduced a polytropic volume exponent, nv, to
account for changes in fluid properties. The polytropic volume exponent is defined as a
constant in solving the polytropic head equation due to assumed negligible variation.
The Schultz polytropic head can then be calculated from equation (2.6), where the
correction factor, f, is introduced to account for the slight variation in nv.
6
nv 1

nv p2 nv
H p,S = f * Z1 RT1 1 (2.6)
n
v 1 p
1

Definitions and equations for the Schultz calculation procedure are given in Appendix
A.2.
ASME PTC 10 refers to the Schultz procedure in cases where the outlet conditions are
unknown. An arithmetic mean value between inlet and outlet conditions is utilized for
estimating the compressibility, compressibility function, and the specific heat. This is not
in accordance with the Schultz procedure, where an imaginary midpoint is utilized for
evaluating the exponents and compressibility functions.
2.3 Wet Gas Performance Analysis
The presence of liquid increases the flow complexity in the compressor. Fluid properties
may vary through the compression process due to energy transfer between the phases.
The dry gas performance analysis becomes insufficient when analysing wet gas
compressor performance. [6] A detailed analysis of the fluid properties along the
compression path is essential to assure correct calculations. The phase exchange during
wet gas compression is not accounted for when utilizing averaged gas properties as in the
Schultz procedure.
Direct Integration Analysis
Direct integration of the polytropic process comprehends iteration from suction to

discharge conditions as shown in equation (2.7). The assumed polytropic efficiency is
kept constant during the iteration process.

hp = hs p = constant (2.7)
i =1
The direct integration procedure is independent of the type of fluid being compressed and
involves using real gas properties. [7] Phase changes along the compression path are
included in the procedure and permit a detailed prediction of the actual volumetric flow
through the compressor. The procedure is suitable for wet gas compression and should be
applied in wet gas performance analysis where phase changes along the compression path
are present. [6] The accuracy of the direct integration performance analysis is dependent
7
on the validity of the relevant equation of state and the determination of the fluid
composition.
2.4 Equation of State
The compressor aerodynamic performance is defined by the enthalpy difference over the
compressor. Enthalpies cannot be measured directly and must therefore be determined
from an appropriate equation of state. An equation of state, EOS, is a thermodynamic
equation describing the mathematical relationship between two or more state functions,
such as pressure, temperature and volume.
It is generally not possible to determine the most accurate EOS to predict gas properties.
Neither ASME PTC 10 nor ISO 5389 gives recommendations regarding the preferred
equation of state for compression processes. The calculated performance may vary
depending on the implemented EOS. Frequently used equations of state are Redlich-
Kwong (RK), Soave-Redlich-Kwong (SRK), Peng-Robinson (PR), Lee-Kesler-Plocker
(LKP), Benedict-Webb-Rubin (BWR) and Benedict-Webb-Rubin-Starling (BWRS).
Equation (2.8) gives the general EOS for a gas, where the compressibility factor, Z, is
included to account for compressibility effects.
pv = ZRT (2.8)
To utilize the general EOS, the compressibility factor must be determined for the
specified gas and actual compressor operating range. Equation (2.9) shows a virial EOS
that can be derived from the principles of statistical mechanics to relate the p-v-T
behavior of a gas to the forces between molecules. The virial coefficients B, C, D, etc.
can be determined analytically or from empirical data. The virial equation of state has a
strong theoretical foundation and is free of arbitrary assumptions.
B (T ) C (T ) D (T )
Z = 1+ + 2
+ + ..... (2.9)
v v v3
At states of a gas where the pressure is small relative to the critical pressure, the second,
third and higher terms of equation (2.9) will diminish. The compressibility factor will
thus approach unity at fixed temperature, giving the ideal gas equation of state. To verify
that a gas can be modeled as an ideal gas, the states of interest must be investigated to
determine how well Z=1 is satisfied.
8
At high pressures and temperatures, the ideal gas behavior is not valid. Changes in fluid
properties must be accounted for by implementing an appropriate EOS in the
performance calculations. The modified virial equation of state, the BWRS, is
recommended by Twu et al. [8] for gases in every temperature and pressure range.
An appropriate equation of state is essential when analyzing wet gas compression due to
phase exchanges when liquid is introduced. Phase changes will affect the actual
volumetric flow through the compressor and hence the compressor performance
validation. The commonly used equations of states are not suitable for prediction of such
phase behavior. [8] Hundseid et al [6] demonstrated the suitability for the GERG-2004
equation of state for wet gas applications. The GERG EOS gives accurate density values
in both vapor and liquid phases and should therefore be implemented for the wet gas
analysis.
9
2.5 Conclusion Chapter 2
The ideal polytropic performance analysis is suitable for compression processes where
the compression fluid behaves approximately like an ideal gas. To verify that a gas can be
modeled as an ideal gas, the states of interest must be investigated to determine how well
Z=1 is satisfied.
At high pressures and temperatures, the ideal gas behavior is not valid. Changes in fluid
properties must be accounted for by implementing the Schultz polytropic analysis for
performance calculations.
The direct integration procedure is suitable for wet gas performance analysis where phase
changes along the compression path are present. Phase transitions in wet gas compression
can be assumed negligible at states where the pressures and temperatures are low and the
fluid inlet condition is stable.
The accuracy of the various performance calculations is dependent on the validity of the
implemented equation of state. The ideal EOS provides an acceptable approximation at
states of a gas where the pressure is small relative to the critical pressure, but will be
highly inaccurate at states with high pressures and temperatures. The BWRS equation of
state is recommended by Twu et al. [8] for gases in every temperature and pressure range.
In wet gas compression the GERG EOS gives accurate density values in both vapor and
liquid phases and should be implemented in the performance analysis where phase
changes are present. [6]
10
Chapter 3
3. Test Preparation
3.1 Test Standards
Different applicable performance test codes are available as a guide when testing a
centrifugal compressor. The American Society of Mechanical Engineers (ASME) and the
International Organization for Standardization (ISO) have issued specifications covering
calculation methods, instrumentation, site preparation and the reporting of test results.
Testing of the impeller rig at the test facility at NTNU is conducted in accordance with
ASME PTC 10, Performance Test Code on Compressors and Exhausters. ASME PTC
10 provides specific guidelines for correct installation and location of different measuring
devices and includes requirements and recommendations for performance calculation
procedures. The compressor testing is conducted with the specified gas at or very near the
specified operating conditions, and is classified as Type 1 test according to ASME PTC
10.
ASME PTC 10 is based on single-phase compression. Additional measurement

uncertainties must be accounted for when utilizing this code under wet gas conditions to
achieve valid and accurate test results.
3.2 Test Parameters
The compressor measurement requirements are:
- Inlet pressure
- Inlet temperature
- Discharge pressure
- Discharge temperature
- Compressor flow
- Speed, torque and power
11
Temperature and Pressure Measurements
To evaluate thermodynamic performance, total temperature and pressure must be

determined at the inlet and discharge of the compressor. [1] Total pressure is the sum of
static and velocity pressure as shown in equation (3.1).
p = pstatic + 0.5V 2 (3.1)
The second term in equation (3.1) will diminish at low flow rates and may be neglected
in performance calculations.
Total temperature is the sum of static temperature and velocity temperature. The actual
temperature measured by a sensing element is normally a value between static and total
temperature depending on the ability of the sensor to recover the converted kinetic energy
of the gas stream.
Total and static values for pressure and temperature are assumed to be equal if inlet and
outlet compressor Mach numbers are less than 0.1. [9] For the actual compressor test,
static measurements are utilized due to a low volumetric flow rate and hence low flow
velocity involved.
Flow Measurement
Properly sized orifice meters are suitable for testing centrifugal compressors over a
normal operating range from surge to choke. The required beta ratio of the device
depends on the maximum flow rate to be measured and the range of the differential
pressure transducer available.
The mass flow rate is related to the differential pressure measured over the orifice and
can be determined by equation (3.2), where D is the diameter of the orifice. The equation
assumes a steady flow with fully developed turbulent velocity profile through the orifice.
The volumetric flow rate (Q=m/) can then be determined, where is the fluid density at
the temperature and pressure measured at the orifice.
C
m= D 2 2+ p 1 (3.2)
1 4 4
The expansibility factor and the discharge coefficient C are empirically determined
correction factors. The discharge coefficient relates the actual flow rate to the theoretical
flow rate through the flow-measuring device. The expansibility factor takes into account
12
the compressibility of the fluid being monitored. Both correction factors can be
determined from tables in different applicable standards or from empirical equations.
The discharge coefficient and the expansibility factor utilized for the compressor test are
verified for the actual impeller rig and expected test conditions, as shown in Appendix B.
Torque, Speed and Power
Torque, speed and power are the defining mechanical variables associated with the
functional performance of rotating machinery. The power output of a compressor (P=)
can be determined directly from torque measurements and/or thermodynamically from
the polytropic analysis shown in chapter 2.1, equation (2.5).
3.3 Test Stability
Compressor performance tests should be performed during steady state conditions. [4]
The compressor and the test equipment must be operated for sufficient time at the
specified operating conditions to demonstrate acceptable mechanical operation and stable
values of all measurements to be taken during the compressor testing.
Steady state is achieved if the compressor measurements listed in Table 3.1 apply during
a 10-minute interval. [9]
Table 3.1: Stability of compressor test [9]

Maximum Allowable Variation
Test Reading
During 10-min Interval
Inlet Pressure 1% of Average Value

Outlet Pressure 1% of Average Value
Inlet Temperature 1C
Outlet Temperature 1C
Compressor Speed 10 rpm
Compressor Flow 1% of Average Value
Temperature measurements are especially sensitive to operational fluctuations during the

compressor test. In order to reach thermal equilibrium and measure accurate
temperatures, the compressor system needs sufficient time after any change in the
operating conditions.
13
3.4 Test Points
A minimum of three readings is taken during each test interval to set up a test point. The
readings are summed, and divided by the total number of readings to obtain an average.
This average is utilized as the test point data.
ASME PTC 10 allows some fluctuation in a test point. The fluctuation is defined as the
percent difference between the minimum and maximum test reading divided by the
average of all readings and can be calculated from equation (3.3), where n is the total
number of readings.
100 ( highest reading - lowest reading )

fluctuation (%) = (3.3)
1 n th
i reading
n i =1
The permissible fluctuations of the measured parameters during a test interval are listed
in Table 3.2.
Table 3.2: Permissible fluctuations of test readings [11]

Measurement Symbol Units Fluctuation
Inlet Pressure p1 [Pa] 2%
Inlet Temperature T1 [K] 0.5 %
Discharge Pressure p2 [Pa] 2%
Discharge Temperature T2 [K] 0.5 %
Molecular Weight Mw [kg/kmol] 0.25 %
Speed N [rpm] 0.5 %
Torque [N*m] 1%
For the measurement of inlet and discharge pressure and temperature, an average of the
four independent sensors is calculated. If one recorded observation is inconsistent due to
measurement error, its value should be discarded and the value determined from the
average of the other three. [1]
14
3.5 Test Uncertainty
Test uncertainty is defined as an estimate of the limit of error of a test result. [10] Test
uncertainty must be calculated to determine the accuracy of the compressor test and to
validate the quality of the compressor test results.
Test uncertainty does not refer to the accuracy of a single instrument, but evaluates the
complete range of possible test results given a singular test condition. Data point recorded
during the test should be evaluated individually and redundant calculations should be
performed to check test measurements. A sufficient number of observations will reduce
the random component of uncertainty to an acceptable level.
Classification
Uncertainty sources and errors may be classified by the presumed effect on the
measurement or test results. ASME PTC 19.1 [11] utilizes this classification.
Systematic error refers to an error source whose effect is constant or systematic during
the test. Systematic uncertainty describes the expected limits to a systematic error. The
error due to a random fluctuation of the measured quantity is referred to as random error.
Random uncertainty refers to the expected limit of the scatter of test data. By recording
more measurements of the test quantity this error can be reduced.
The time interval and duration of the test must be clearly specified for classifying an error
as either random or systematic. The total error in a measurement is the combination of
systematic and random errors. The difference between the true and the measured value is
the total error, as shown in Figure 3.1.
Systematic and random uncertainties are both defined at the standard-deviation level as
standard uncertainties. An uncertainty level of 95% is commonly used in ASME PTC
19.1.
15
Figure 3.1: Illustration of measurement errors [11]
Measurement uncertainty
Deviations in the ideal or recommended test conditions and procedures will increase the
individual measurement uncertainties. This can result in a higher total uncertainty for the
centrifugal compressor. The uncertainty in the pressure or temperature measurements
may result from disturbances in the flow field upstream or downstream the compressor.
Installation requirement in ASME PTC 10 ensures that these uncertainties are minimized
or negligible by taking sufficient observation to ensure that the average measurements
will be accurate.
If piping vibration or flow-induced pulsations are high at the location of the static
pressure measurement, the measurement of pressure will show a significantly higher
random uncertainty. Wall conduction heat transfer to and from temperature sensor
contributes to the temperature measurements error.
Calibration
Proper calibration of the instruments can reduce the measurement uncertainty. All
pressure measurements are referred to devices that serve as primary standards of pressure
measurements. A big contribution to the uncertainty of pressure measurements is the
errors accumulated in the propagation of the measurement uncertainty along a calibration
chain and not from the primary standard. [12]
16
Data acquisition
Error in the data acquisition system contributes to the overall uncertainty. Calibration of
the overall system can minimize this effect.
Data reduction uncertainty
Care must be taken when averaging data. Averaging on raw test data may cause
unacceptable deviations and should be avoided. The computational uncertainty is
commonly assumed negligible due to advanced computer systems.
17
18
Chapter 4
4. Compressor Test Facility
4.1 Impeller Rig
The values given for the impeller rig are the values at design point. Some discrepancies
between design point and actual operating point are expected.
The test compressor is composed of an impeller proceeded by a vaneless radial diffuser

and a volute as shown in Figure 4.1.
Figure 4.1: Test compressor
The shrouded test impeller shown in Figure 4.2 contains 18 blades with a back sweep of
50 at the exit. Splitter vanes can be implemented by removing sections of the existing
impeller blades. At design point flow rate of 1.0 m3/s, a work input coefficient of 0.69 is
assumed so that the impeller would have enough pressure-rise to produce a stage pressure
ratio of 1.39 and polytropic efficiency of 80%. The compressor design speed is 10 000
rpm. The frequency of the impeller rotation is 166.78 Hz, with a blade passing frequency
of 3002.04 Hz, half that for potential splitter vanes.
19
Figure 4.2: Test impeller
Atmospheric air is drawn into the compressor through an orifice plate located in the
suction pipe. When testing the compressor for wet gas conditions liquid can be sprayed
into the gas flow from an injection-module placed downstream of the orifice to assure
correct flow measurements.
A high-speed motor drives the compressor with a frequency converter for speed control.
The desired pressure ratio is reached by means of a butterfly valve. The throttling valve is
mounted to a divergent pipe segment installed at the discharge pipe due to dissimilar
diameters at pipe and valve. Adjusting the exit area of the valve with a rotary actuator
controls the compressor mass flow. After passing through the throttle valve, the air is
exhausted to the atmosphere. The test rig is shown in Figure 4.3.
Figure 4.3: NTNU test rig
20
4.2 Piping Configuration
Figure 4.4 shows the pipe length recommendation upstream the compressor according to
ASME PTC 10 [1] and ISO 5167 [13].
The orifice with beta value 0.64 requires a straight pipe length of 5.0m (20D) upstream
and 0.875m (3.5D) downstream the device to assure a correct flow reading. The
requirement of straight pipe length given by the vendor of the orifice is in accordance
with the ISO 5167 recommendations for a single bend pipe. If an orifice with lower beta
value is chosen, shorter pipe lengths are required owing to the lower pressure drop over
the device.
The compressor has an axial inlet that may produce a vortex at the pressure station under
various conditions. The static pressure stations at the inlet are installed at a minimum of
1.0m (4D) upstream the compressor flange to avoid errors in the measurement of inlet
pressure.
Figure 4.4: Inlet piping and instrumentation
The static pressure stations at the discharge are installed 0.6618m (6D) from the
compressor exit due to the un-symmetrical flow produced by the compressor volute.
Figure 4.5 shows the ASME PTC 10 pipe length recommendation downstream the
compressor.
21
Figure 4.5: Discharge piping and instrumentation
4.3 Apparatus and Instrumentation
Table 4.1 shows an overview of the instrumentation utilized for the impeller test rig.
Table 4.1: Impeller test rig instrumentation list

Element Type Number
Pressure Pressure Transmitter, APCE-2000 8
Temperature Pt 100 (IEC 751) 8
Flow Orifice: Plate-Standard, Corner (Beta value 0,6401) 1
Orifice: Plate-Standard, Corner (Beta value 0,4018) 1
HBM T-12 Digital torque transducer, with BSD-

Torque Transducer MODULFLEX Coupling system 1
Pressure Instrumentation
Four pressure measurements are taken at the inlet and discharge piping of the
compressor. Figure 4.6 shows the pressure transmitter utilized for the measurements. The
pressure transmitters work by converting changes in the resistance of a piezoresistance
silicon sensor, which are proportional to the pressure difference being measured, into a
standard current signal. The uncertainty of the pressure transmitter is 0.1%.
22
Figure 4.6: APCE-2000 pressure transmitter
Temperature Instrumentation
Four temperature measurements are taken at the inlet and discharge piping of the
compressor by resistance temperature devices. Figure 4.7 shows the Pt 100 element
utilized for the temperature measurements. The sensors measure the changes in resistance
of a platinum wire that is coiled together to form a ceramic core. The changes in
resistance are converted to a standard voltage signal. Pt 100 elements are assumed to be
highly accurate with great temperature sensitivity. The uncertainty of the pressure
transmitter is 0.05%.
Figure 4.7: Pt 100 sensor
23
Figure 4.8 shows the pressure and temperature sensors located at compressor inlet, while
Figure 4.9 shows the sensors at outlet.
Figure 4.8: Pressure and temperature sensors at inlet
Figure 4.9: Pressure and temperature sensors at outlet
Flow Instrumentation
An orifice with beta value 0.64 located at the inlet piping measures the flow through the
compressor. Two static pressure taps located on opposite sides of the orifice measure the
differential pressure over the orifice. An orifice with beta value 0.40 for measuring
smaller fluid flows can replace the orifice.
24
Orifice meters with beta ratios less than 0.65 have a flow measurement uncertainty of less
than 1.5%. [9] The orifice accuracy will not be influenced by liquid present due to the
downstream location of the injection module. Hence a multiphase correction of the
device is not required.
Torque, Speed and Power Instrumentation
Torque, rotational speed and power are determined by utilizing a digital measurement
system. The torque transducer is connected to the shaft through a torque-sensing
coupling. The torque transducer provides a digital signal proportional to the measured
torque. Figure 4.10 and Figure 4.11 shows respectively the digital torque transducer and
the coupling system.
Figure 4.10: HBM T-12 digital torque transducer
Figure 4.11: BSD-MODULFLEX coupling system
25
Instrumentation for Instability Measurements
Pressure fluctuations caused by instabilities and the resulting vibration of the rotor can be
measured with variation of flow rate and impeller rotating speed to document and
determine the cause of the instability phenomenon.
Due to late delivery of the compressor components, the sensors for instability
measurements are not yet implemented in the compressor rig. The recommended
measurement procedures, instrumentation and location of sensors are described in chapter
7.4-7.7. Multiphase flow measurement and visualization techniques and instrumentation
are described in chapter 8.
4.4 Data Acquisition System
Automatic parameter registration is performed with Lab View-software on a digital

computer. The computer monitors all measured parameters by transforming the analog
signals from the pressure transmitters and temperature sensors into digital data. The
performance equations can be implemented directly into Lab View.
26
Chapter 5
5. Compressor Test
The construction of the compressor rig was behind time due to late deliveries of the
compressor components and instrumentation. The performance calculations are based
upon one compressor test conducted with dry gas at part-load. The results are utilized to
validate the compressor behavior at the test operating condition.
Instrumentation for detection of compressor instabilities is not yet implemented in the

compressor test rig. Documentation of compressor instability is therefore discarded.
Recommended methods, procedures and instrumentation for wet gas is instead presented
in chapter7.4-7.7. Visualization techniques are presented in chapter 8.
5.1 Test Matrix
The compressor test is conducted at part load, with 5000 rpm. By gradually adjusting the
butterfly valve, the compressor mass flow is controlled. Lab View registers the test
parameters with a logging frequency of approximately three second. Table 5.1 shows the
intended test matrix consisting of ten test points at rotational speed of 5000 rpm.
Table 5.1: Initial test matrix at part load

Speed, N Flow, Q1
[rpm] [m3/s]
1 5000 0,40
2 5000 0,45
3 5000 0,50
4 5000 0,55
5 5000 0,60
6 5000 0,65
7 5000 0,70
8 5000 0,75
9 5000 0,80
10 5000 0,85
27
5.2 Performance Procedure
The compressor rig at the NTNU test facility is tested at atmospheric inlet conditions
with a design stage pressure ratio of 1.39. In cases where the pressure is small relative to
the critical pressure (pc=37,7bar) and/or the temperature is large relative to the critical
temperature (Tc=133K), a near ideal gas behavior may be assumed. [14] At such
conditions the ideal polytropic procedure can be utilized to determine the compressor
performance. Assumptions made in advance for the compression gas and performance
procedures must always be verified for the actual test and operating range.
In compression cases where the Schultz polytropic exponent is approximately equal to

the ideal polytropic exponent and the Schultz correction factor is one, the two polytropic
procedures give equivalent results, as shown in Appendix A.3. The ideal polytropic
procedure is, in such cases, often preferred in reference to the Schultz procedure because
of the simplicity of the ideal polytropic calculations. The polytropic procedure is, for this
particular test, preferred because the polytropic calculation procedures are more easily
implemented in the Lab View software.
PRO/II Simulation
The sensitivity and validity of the performance calculations are evaluated by

implementation of the compression routines in the process simulation system PRO/II.
PRO/II has implemented the Schultz polytropic procedure for performance calculations.
The PRO/II simulation setup is shown in Appendix C. By varying the compressor
operating conditions, the sensitivity of the calculation procedures is identified and the
procedures validated for the actual compressor test and operating range.
PRO/II utilizes the mass flow, outlet temperature and pressure to estimate the specific
outlet volume from an appropriate EOS, and hence to determine the polytropic exponent.
The PRO/II calculations are congruent with the calculation procedures given in the ISO
and ASME standards for cases with known outlet conditions. For cases with unknown
outlet conditions PRO/II utilizes iterative processes for the performance calculations.
Comparison of Ideal Polytropic and Schultz Polytropic Procedures
Schultz compressibility functions, X and Y, approach respectively zero and one for the
gas inlet condition, as shown by the generalized compressibility charts given in Appendix
D. Schultz polytropic exponent will thus approach the ideal polytropic exponent.
To verify the assumption of ideal gas behavior for the compressor test, the
compressibility factor, Z, is set equal to 1 to determine the specific volumes and hence
the ideal polytropic exponents from the equations given in chapter 2.1 and Appendix A.1.
The polytropic exponent is in addition determined from simulations in PRO/II for
28
comparison, since PRO/II gives a polytropic exponent corresponding to the Schultz
polytropic exponent.
The accuracy of the thermodynamic properties calculated by PRO/II depends on the

validity of the selected EOS for the specified gas and thermodynamic condition. The
BWRS equation is utilized for the performance simulations. The BWRS equation of state
is found to give approximately equal values to the ideal equation of state by simulations
in PRO/II for the compressor operating range.
Table 5.2 and Table 5.3 show representative measured parameters for the compressor test
and calculated performance data at compressor inlet and discharge with the assumption of
ideal gas behavior.
Table 5.2: Measured and calculated test parameters at inlet with Z=1
p1 T1 Q1 1 v1 m
3 3 3
[bar] [K] [m /s] [kg/m ] [m /kg] [kg/s]
0,937 296,20 0,83 1,10 0,91 0,92
0,941 296,53 0,81 1,11 0,90 0,89
0,949 296,64 0,77 1,12 0,90 0,85
0,973 296,65 0,60 1,14 0,88 0,68
0,978 296,72 0,54 1,15 0,87 0,62
0,983 296,72 0,49 1,15 0,87 0,56
0,990 296,71 0,41 1,16 0,86 0,48
Table 5.3: Measured and calculated test parameters at discharge with Z=1
p2 T2 Q2 2 v2 m
[bar] [K] [m3/s] [kg/m3] 3
[m /kg] [kg/s]
0,999 302,88 0,80 1,15 0,87 0,92
1,007 303,43 0,77 1,16 0,86 0,89
1,023 303,82 0,73 1,17 0,85 0,85
1,065 304,20 0,56 1,22 0,82 0,68
1,074 304,46 0,51 1,23 0,81 0,62
1,082 304,50 0,45 1,24 0,81 0,56
1,091 304,63 0,38 1,25 0,80 0,48
29
Table 5.4 shows a direct comparison between the corresponding calculated ideal
polytropic exponent and the estimated Schultz polytropic exponent given by PRO/II. The
corresponding values for the polytropic head are given in Table 5.5.
Table 5.4: Comparison of the polytropic exponent

Relative
n at Z=1 nv - PRO/II Deviation Deviation
[%]
1,5662 1,5674 0,0012 0,0766
1,4712 1,4720 0,0008 0,0543
1,5091 1,5101 0,0010 0,0662
1,3989 1,3995 0,0006 0,0429
1,4191 1,4198 0,0007 0,0493
1,3644 1,3648 0,0004 0,0293
1,3757 1,3762 0,0005 0,0363
Table 5.5: Comparison of the polytropic head

Relative
Hp [J/kg] Hp,S [J/kg] Deviation
Deviation
[J/kg]
n at Z=1 nv - PRO/II [%]
5507,35 5507,44 0,0865 0,0016
5859,15 5859,22 0,0740 0,0013
6434,15 6434,26 0,1058 0,0016
7774,37 7774,48 0,1079 0,0014
8051,74 8051,87 0,1310 0,0016
8254,13 8254,21 0,0852 0,0010
8415,02 8415,13 0,1088 0,0013
The Schultz polytropic exponent given by PRO/II has a maximum deviation of 0.0766%
from the calculated ideal polytropic exponent. The corresponding deviation for the
Schultz polytropic head is 0.0016%.
The values estimated by PRO/II agree closely with the calculated values determined by
the ideal gas equations and thus verify the use of ideal polytropic calculation procedures
for the compressor test.
The validation of the ideal polytropic procedure will only be accurate for the specified
gas and test condition. For higher pressure and temperatures, the Schultz polytropic
procedure must be implemented and verified for the real gas behavior.
30
Sensitivity of Polytropic Head
Deviation between ideal and Schultz polytropic exponents will for low pressure ratios
only slightly influence the calculation of polytropic head, and can in such cases be
neglected. For higher pressure ratios, similar deviation may strongly influence the
polytropic head and must be accounted for. Table 5.6 shows how deviation between the
polytropic exponents will influence the deviation in polytropic head for different pressure
ratios.
Table 5.6: Deviation between the polytropic exponents for various pressure ratios
p2/p1 = 1,09 p2/p1 = 1,39 (Design) p2/p1 = 10
n Hp n Hp n Hp
Relative deviation [%] Relative deviation [%] Relative deviation [%]
3,24 0,10 0,84 0,10 0,11 0,10
16,21 0,50 4,19 0,50 0,55 0,50
32,47 1,00 8,40 1,00 1,10 1,00
65,16 2,00 16,86 2,00 2,21 2,00
If the polytropic exponents deviate with 3.24%, the polytropic heads will deviate with
0.1% for a pressure ratio of 1.09. To be within the same deviation in head, the polytropic
exponents can only deviate with 0.84% and 0.11% for pressure ratios of respectively 1.39
and 10. This verifies that the sensitivity of the ideal polytropic head calculation is notably
low in the compressor operating range, but will be substantially higher for greater
pressure ratios. The Schultz polytropic analysis must then be implemented to achieve
accurate performance results.
Isentropic Exponent
The Schultz polytropic analysis utilizes the isentropic reference process in predicting the
correction factor, f, and hence the polytropic head. Variation in isentropic exponent may
therefore strongly influence the calculated polytropic head. PRO/II simulations show that
the Schultz correction factor, f, is constant and equal to one in the compressor operating
range, and is not influenced by variation in isentropic exponent.
Table 5.7 shows that variation in the isentropic exponent estimated by PRO/II are small
with varying pressure ratio and can be neglected for performance calculations in the
compressor operating range.
31
Table 5.7: Variation in isentropic exponent with varying pressure ratio
p2/p1
[-] [-]
1,002 1,4027
1,010 1,4031
1,050 1,4031
1,090 1,4031
1,500 1,4030
2,000 1,4028
Sensitivity of Polytropic Efficiency
Variations in polytropic exponent, with approximately constant isentropic exponent, will

highly influence the calculated polytropic efficiency, as seen in Table 5.8
Table 5.8: Variation in polytropic efficiency with varying polytropic exponent

n p
[-] [-] [-]
1,9187 1,4031 0,60
1,7921 1,4031 0,65
1,6961 1,4031 0,70
1,6209 1,4031 0,75
1,5603 1,4031 0,80
1,5106 1,4031 0,85
1,4689 1,4031 0,90
1,4335 1,4031 0,95
1,4031 1,4031 1,00
Table 5.9 shows the sensitivity of the polytropic efficiency with deviation in polytropic
exponents from the first three test points in Table 5.4.
Table 5.9: Sensitivity of polytropic efficiency with varying polytropic exponent

Relative
p [-] p,S [-] Deviation Deviation
n at Z=1 nv - PRO/II [-] [%]

0,7947 0,7936 0,0011 0,1352
0,8970 0,8960 0,0010 0,1153
0,8516 0,8505 0,0011 0,1301
32
The larger difference between the calculated ideal polytropic exponent for the
compressor test and the Schultz exponent estimated from PRO/II is 0.0012, a relative
deviation of 0.0766%. As shown in Table 5.5, the polytropic head then holds a relative
deviation of 0.0016%. In comparison, the calculated polytropic efficiency will have a
relative deviation of 0.1352% at constant isentropic exponent (=1.4031). Hence the
polytropic efficiency is more sensitive for deviations in polytropic exponent than the
polytropic head.
The polytropic efficiency calculated for the compressor test turned out to be extremely
high, indicating that one or more parameters are measured inaccurately or that the
presumed constant isentropic exponent is incorrect. Chapter 6 shows the effect of
measurement uncertainties on the performance calculations for a case representative to
the compressor test. Due to the low pressures involved for the actual compressor test, the
polytropic head and efficiency are highly sensitive to uncertainties in the pressure
measurements. Uncertainties in temperature measurements will only slightly influence
the polytropic head, but will have strong influence on the polytropic efficiency.
EOS Sensitivity
The selected EOS for the performance simulations in PRO/II is the BWRS equation.
Since the calculated performance may vary depending on the implemented EOS, a
sensitivity analysis is conducted.
The polytropic head deviations for the various equation of state compared to the
calculated ideal equation are given in Table 5.10. The operating condition and calculated
parameters for ideal gas behavior are shown in Table 5.11.
Table 5.10: Variation in polytropic head with different equation of state

Rel.
n Hp
PRO/II Deviation, Hp
[-] [J/kg] [%]
Ideal 1,5662 5507,35 0,0000
BWRS 1,5674 5507,44 0,0016
SRK 1,5679 5507,47 0,0022
PR 1,5675 5507,44 0,0017
33
Table 5.11: Operating data for compressor test
p1 [bar] 0,94
T1 [K] 296
p2 [bar] 1,00
T2 [K] 303
[-] 1,4031
n [-] 1,5662
p [-] 0,79
Hp [J/kg] 5507
Table 5.10 demonstrates that deviations in the polytropic head with different EOS are
negligible for the actual compressor performance evaluation. The BWRS equation of
state gives a polytropic head with a relative deviation of 0.0016% to the ideal equation.
The different EOS yield similar results for the polytropic head within an accuracy range
of 0.0017%-0.0022%.
The negligible values for relative deviation with different EOS verifies that at states
where the pressure is small relative to the critical pressure, the ideal EOS will provide a
suitable approximation. For higher pressures and real gases, the ideal EOS is not valid.
The selected EOS must always be verified for the actual compressor operating range and
fluid composition.
5.3 Generation of Performance Curves from Recorded Data Points
The compressor behavior referred to constant operating conditions and constant

compressor geometry is presented in the form of performance curves.
Dimensional analysis techniques are often conducted in order to eliminate the excessive
number of experiments required to obtain a complete presentation of the variables over
the expected operating range. Since the NTNU impeller rig is operated at approximately
constant atmospheric inlet conditions with stable air properties, a dimensional analysis is
not necessary. Curves of delivery pressure and polytropic head are plotted against
volumetric flow for various fixed values of speed are utilized to generate the performance
curves.
34
Test Point
The compressor test points are each determined from three test readings registered by
Lab View for the calculation of average test values. The fluctuations for each test
parameter in the test points are calculated in accordance with ASME standard procedures.
Test points with fluctuations outside the limits given by ASME PTC 10 are discarded.
The valid test point is shown in Appendix E.
Table 5.12 displays the measured parameters together with the calculated pressure ratio
and polytropic head for six representative test points. The corresponding polytropic head
curve is shown in Figure 5.1. The pressure ratio with varying volume flow is shown in
Figure 5.2.
Table 5.12: Measured test parameters and calculated test results

Measured Calculated
Test
p1 T1 Q1 p2 T2 N p2/p1 Hp
Point
[bar] [K] [m3/s] [bar] [C] [rpm] [-] [J/kg]
11 0,94 296,2 0,83 1,00 302,9 4983 1,07 5505
25 0,94 296,5 0,81 1,01 303,4 4951 1,07 5863
35 0,95 296,6 0,77 1,02 303,8 4937 1,08 6430
44 0,97 296,6 0,60 1,06 304,2 4948 1,09 7772
55 0,98 296,7 0,54 1,07 304,5 4951 1,10 8045
60 0,99 296,7 0,41 1,09 304,6 4966 1,10 8447
5000 rpm
9000
8000
7000
Polytropic head [J/kg]
6000
5000
4000
3000
2000
1000
0
0,35 0,45 0,55 0,65 0,75 0,85 0,95
Volum e flow [m 3/s]
Figure 5.1: Polytropic head curve
35
5000 rpm
1,12
1,11
1,10
Pressure ratio [-]
1,09
1,08
1,07
1,06
1,05
0,35 0,45 0,55 0,65 0,75 0,85 0,95
Volum e flow [m 3/s]
Figure 5.2: Pressure ratio with varying volume flow
Non-dimensional head and flow coefficients
The average of the data point at a particular operating condition is used to calculate the
average head and flow coefficients. The non-dimensional head and flow coefficients
describe the aerodynamic and mechanical performance of a centrifugal compressor and
are used to adjust for differences in test conditions in order to match the flow
characteristics when the compressor is tested under varying conditions.
The polytropic head coefficient is given in equation (5.1). The tangential velocity
(U=DN/60) is calculated at the impeller exit diameter (0.385m). The flow coefficient is
given in equation (5.2).
Hp
p = (5.1)
U2
Q1
= (5.2)
N 3
2 D
60
Figure 5.3 shows the head versus flow coefficient curve for the compressor test at
approximately 5000 rpm. The calculated values for the polytropic head coefficient and
flow coefficient are given in Appendix F.
36
Compressor test
0,90
0,85
Polytropic head coefficient [-]
0,80
0,75
0,70
0,65
0,60
0,55
0,50
0,45
0,40
0,0100 0,0130 0,0160 0,0190 0,0220 0,0250 0,0280
Flow coefficient [-]
Figure 5.3: Polytropic head versus flow coefficient performance curve
5.4 Affinity Laws
To predict the compressor performance at different rotational speeds, the affinity laws,
equation (5.3) and equation (5.4), are utilized.
2
N
H = H ref (5.3)
N ref
N
Q = Qref (5.4)
N ref

The affinity laws allow prediction of the head discharge characteristic of a compressor
from a known characteristic measured at a different speed or impeller diameter. The
affinity laws are applicable only if the compressor flow is kinematic and dynamic similar.
Kinematic similarity yields when the velocity diagrams are similar for the different flow
conditions. Dynamic similarity is achieved when the corresponding reduced velocities are
of the same size. The efficiency is held constant in the deduction of the laws.
37
Table 5.13 shows the calculated volume flow and polytropic head for rotational speed
7000 rpm and 10 000 rpm. A comparison of the polytropic head curves is shown in
Figure 5.4.
Table 5.13: Predicted performance given by the affinity laws

5000 rpm 7000 rpm 10 000 rpm
Q1 Hp Q1 Hp Q1 Hp
Test
Point [m3/s] [J/kg] [m3/s] [J/kg] [m3/s] [J/kg]
11 0,83 5505 1,16 7707 1,66 11010
25 0,81 5863 1,13 8208 1,62 11725
35 0,77 6430 1,07 9002 1,53 12860
44 0,60 7772 0,84 10881 1,19 15544
55 0,54 8045 0,76 11262 1,09 16089
60 0,41 8447 0,58 11826 0,83 16894
18000
16000
14000
Polytropic Head [J/kg]
12000
10000
5000 rpm
8000 7000 rpm
10 000 rpm
6000
4000
2000
0
0,3 0,5 0,7 0,9 1,1 1,3 1,5 1,7 1,9
Volume Flow [m3/s]
Figure 5.4: Polytropic head curves given by the affinity laws
38
5.5 Test Stability
Table 5.14 shows the deviations in test parameters compared to the average values of the
test points. The measured parameters are outside the limits given by ASME PTC 10,
Table 3.1.
Due to the low pressures involved for the compressor test, the measurements of pressure
are highly sensitive to fluctuations. Fluctuations in test condition, with regard to
temperature, are most likely caused by the temperature rise in the test facility. As
described in chapter 3.3, the compressor needs sufficient time after any change in
operating condition to reach thermal equilibrium and hence measure accurate
temperatures. The fluctuating speed may be the result of an unstable compressor shaft or
uncertainties in the frequency converter.
Table 5.14: Test stability

Test Average
SI Unit Highest Lowest Deviation Allowable
Parameter Value
Deviation
p1 [bar] 0,96 0,99 0,94 5,87 % 1%
p2 [bar] 1,03 1,09 1,00 9,34 % 1%
T1 [K] 296,5 296,8 295,7 1,13 K 1K
T2 [K] 303,6 304,6 302,2 2,38 K 1K
N [rpm] 4963 5063 4877 186 rpm 10 rpm
5.6 Static Measurement Uncertainty
At Mach numbers < 0.1 the total and static values for pressure and temperature are
assumed to be the same. [9] Table 5.15 gives the inlet and outlet Mach numbers for the
compressor test points.
Table 5.15: Mach numbers at inlet and outlet

Inlet Outlet
Test Q1 V1 a1 Ma1 Q2 V2 a2 Ma2
3 3
Point [m /s] [m/s] [m/s] [-] [m /s] [m/s] [m/s] [-]
11 0,83 17,0 345,0 0,05 0,80 85,5 348,9 0,24
25 0,81 16,5 345,2 0,05 0,77 80,9 349,2 0,23
35 0,77 15,6 345,2 0,05 0,73 76,1 349,4 0,22
44 0,60 12,2 345,2 0,04 0,56 58,6 349,6 0,17
55 0,54 11,1 345,3 0,03 0,51 53,2 349,8 0,15
60 0,41 8,4 345,3 0,02 0,38 40,3 349,8 0,12
39
The Mach number requirement is satisfied for the inlet measurements. The higher Mach
numbers at outlet are due to a smaller outlet diameter that results in a higher velocity at
the measuring station.
The velocity component of the measurements cannot be assumed negligible at the outlet
measuring station. An additional uncertainty must be accounted for due to a greater
difference between total and static pressure and temperature.
5.7 Wet Gas Predictions
The presence of liquid in a compression process will decrease the discharge temperature
and increase the fluid discharge density due to evaporate cooling of the gas phase. [7] A
change in relative volume flow through the compressor due to increased discharge
density results in a change in compressor performance and efficiency as shown in
respectively Figure 5.5 and Figure 5.6.
Figure 5.5: Pressure rise with different GVF [15]
Figure 5.6: Efficiency with different GVF [15]
40
The gas liquid content is given by the gas volume fraction in equation (5.5).
Qg
GVF = = (5.5)
Qg + Ql
Figure 5.6 indicate a more narrow operating range for the wet gas compressor, as the
curves shift to left with increasing GVF. A narrow operating range will result in a
reduced surge margin for the wet gas compressor. Figure 5.7 show that the efficiency for
wet gas compression is expected to decrease with increasing liquid content.
The direct integration procedure, as presented in chapter 2.3, is suitable for wet gas
compression and must be applied in performance analysis where phase changes along the
compression path are present. Phase transitions are dependent on the fluid inlet
composition and compressor operating condition.
The probability for phase changes to occur is limited for the actual compressor test due to
the low compressor pressure ratio and the atmospheric inlet condition, and can thus be
neglected in the performance evaluation. Performance procedures based on the polytropic
assumptions will then be accurate for the actual test. Wet gas testing must be conducted
to ensure the validity of this assumption for the compressor operating range.
Since phase transitions are assumed negligible, the BWRS will be an applicable EOS for
the planned wet gas compressor test. At higher pressure ratios this assumption is not
valid, and an EOS that takes into account any phase transitions must be utilized. The
GERG EOS gives accurate density values in both vapor and liquid phases and is
recommended for the wet gas analysis.
When utilizing static measurements for the wet gas performance calculations, the
uncertainties will be substantially higher. The speed of sound (a=(RT)1/2) will be
severely degraded when liquid is introduced due to greater molecular weight. The Mach
number (Ma=V/a) is inversely proportional to the speed of sound and will increase as the
speed of sound decreases. Total measurements must be implemented for the performance
evaluation, even for low volumetric rates.
Figure 5.7 shows how the gas volume fraction, GVF, will influence the speed of sound in
a typical natural gas composition. [7]
41
Figure 5.7: Speed of sound with different GVF [7]
42
The values estimated for polytropic exponent and polytropic head by PRO/II agree
closely with the calculated values determined by the ideal gas equations and thus verify
the use of ideal polytropic calculation procedures for the compressor test. The sensitivity
of the ideal polytropic head is notably low in the compressor operating range, but will be
substantially higher for greater pressure ratios. The polytropic efficiency is shown to be
more sensitive for deviations in polytropic exponent than the polytropic head for the
compressor test.
The thermodynamic equation of state for ambient air is verified to be consistent with the
ideal gas law in the compressor pressure and temperature range. Deviations in the
performance procedures with different equation of state are negligible for the actual
performance evaluation. The ideal EOS provides an acceptable approximation at many
states, but will be highly inaccurate at states with high pressures and temperatures.
Variations in isentropic exponent estimated by PRO/II are small with varying pressure
ratio and can be neglected for performance calculations in the compressor operating
range.
The probability for phase changes to occur is limited for the planned wet gas compressor
test due to the low compressor pressure ratio and the atmospheric inlet composition and
can thus be neglected in the performance evaluation. Performance procedures based on
polytropic assumptions will be accurate for the planned wet gas compressor test.
When converting the performance results for the compressor test to real gases at high
pressure and temperature conditions, the ideal assumptions will not be applicable. The
Schultz polytropic analysis must be utilized to achieve accurate results. An applicable
EOS for the gas condition and operating range must be implemented in the procedures.
43
44
Chapter 6
6. Measurement Sensitivity
A sensitivity analysis is conducted in order to determine the effect of measurement

uncertainties on the performance calculations. By changing the test parameters one at a
time deviations in polytropic head is analyzed.
The sensitivity analysis is performed for to different cases. The analysis for case 1 is
conducted with outlet conditions representative for the actual compressor test. Case 2 is
conducted to determine the measurement sensitivity for the performance calculations at
higher outlet pressure and temperature.
6.1 Case 1
The test parameters utilized in the sensitivity analysis for case 1 are given in Table 6.1.
Table 6.2 displays the constant input values for the performance calculations. The
calculated values resulting from the sensitivity analysis are shown in Appendix G.1.
Table 6.1: Reference test parameters Case 1

Reference
Test Parameters Symbol Units
Test Point
Inlet Pressure p1 0,95 [bar]

Outlet Pressure p2 1,01 [bar]
Inlet Temperature T1 295 [K]
Outlet Temperature T2 302 [K]
Table 6.2: Test point constants

Reference
Test Point
Compressibility Constant Z 1 [-]
Inlet Temperature R 287 [J/kg*K]
Adiabatic Exponent 1,4 [-]
45
Sensitivity of Pressure Measurements
The sensitivity of polytropic head with varying inlet and outlet pressure is shown in
respectively Figure 6.1 and Figure 6.2. The polytropic head is equally influenced, in
order of relative deviation magnitude, by inlet and outlet pressure measurements.
Deviation in polytropic head with varying inlet pressure Reference inlet pressure
20
Deviation in polytropic head [%]
15
10
0
0,940 0,945 0,950 0,955 0,960
-5
-10
-15
-20
Inlet pressure [bar]
Figure 6.1: Deviation in polytropic head with variation in inlet pressure
Deviation in polytropic head with varying outlet pressure Reference outlet pressure
20
Deviation in polytropic head [%]
15
10
0
1,000 1,005 1,010 1,015 1,020
-5
-10
-15
-20
Outlet pressure [bar]
Figure 6.2: Deviation in polytropic head with variation in outlet pressure
46
A measurement uncertainty of 0.001 bar in inlet pressure gives a relative deviation of
1.7% for the calculated head. An equal uncertainty in outlet pressure results in a relative
deviation of 1.6%, as shown in Appendix G.1.
The sensitivity analysis shows that the performance calculations for the actual
compressor test and operating range are highly influenced by inaccurate measurement of
inlet and outlet pressure.
A summary of the sensitivity analysis for polytropic head with variation in pressure for
case 1 is shown in Table 6.3. The polytropic head deviates 1% when the inlet or outlet
pressure deviate 0.061%, which corresponds to a discrepancy of 0.0006 bar. To achieve a
96% precision, the inlet or outlet pressure can deviate up to 0.245%.
Table 6.3: Sensitivity of polytropic head with varying pressure Case 1

Relative deviation
Hp [%] p1 [%] p1 [bar] p2 [%] p2 [bar]
0,1 0,006 0,0001 0,006 0,0001
0,5 0,031 0,0003 0,031 0,0003
1 0,061 0,0006 0,061 0,0006
2 0,123 0,0012 0,123 0,0012
4 0,245 0,0023 0,245 0,0025
The deviation curves for polytropic efficiency are equal to the curves for polytropic head,
as shown in Table G.1 and Table G.2 in Appendix G.1.
Sensitivity of Temperature Measurements
Figure 6.3 and Figure 6.4 shows the sensitivity of polytropic head with varying inlet and
outlet temperature.
An uncertainty of 1.0 K in inlet or outlet temperature results in a deviation of 0.17% for

the calculated polytropic head, as given in Table G.3 and Table G.4 in Appendix G.1.
The sensitivity analysis shows that the polytropic head is only slightly influenced by
relative deviations of inlet and outlet temperature for case 1.
The polytropic efficiency is, in contrast, highly sensitive to uncertainties in temperature

measurements. Table G.3 and Table G.4, in Appendix G.1, show that a strong influence
on the calculated efficiency when varying inlet and outlet temperature. A 1.18% increase
in inlet temperature results in a 100% increase in polytropic efficiency.
47
Deviation in polytropic head with varying inlet temperature Reference inlet temperature
Deviation in polytropic head [%] 2,0
1,5
1,0
0,5
0,0
290 291 292 293 294 295 296 297 298 299 300
-0,5
-1,0
-1,5
-2,0
Inlet temperature [K]
Figure 6.3: Deviation in polytropic head with variation in inlet temperature
Deviation in polytropic head w ith varying outlet temperature Reference outlet temperature
2,0
Deviation polytropic head [%]
1,5
1,0
0,5
0,0
297 298 299 300 301 302 303 304 305 306 307
-0,5
-1,0
-1,5
-2,0
Outlet tem perature [K]
Figure 6.4: Deviation in polytropic head with variation in outlet temperature
A summary of the sensitivity analysis for polytropic head with variation in temperature
for case 1 is shown in Table 6.4. The polytropic head deviates 1% when the inlet
temperature deviate 2.015%, which corresponds to a discrepancy of 5.94 K. The outlet
temperature can deviate up to 1.999%, 6.04 K, to achieve the same accuracy for the
polytropic head. To achieve a 96% precision, the inlet temperature can deviate up to
8.138%, while the outlet temperature can deviate up to 8.073%.
48
Table 6.4: Sensitivity of polytropic head with varying temperature Case 1
Deviation
Hp [%] T1 [%] T1 [K] T2 [%] T2 [K]
0,1 0,201 0,59 0,199 0,59
0,5 1,006 2,97 0,998 3,01
1 2,015 5,94 1,999 6,04
2 4,043 11,93 4,011 12,11
4 8,138 24,01 8,073 24,38
6.2 Case 2
Sensitivity at Higher Pressure Ratio
The test parameters used in the sensitivity analysis for case 2 are listed in Table 6.5. The
constant input values are the same as for case 1. The calculated values resulting from the
sensitivity analysis are shown in Appendix G.2.
Table 6.5: Reference test parameters Case 2

Reference
Test Point
Inlet Pressure p1 0,95 [bar]

Outlet Pressure p2 2,00 [bar]
Inlet Temperature T1 295 [K]
Outlet Temperature T2 387 [K]
A summary of the sensitivity analysis for polytropic head with variation in pressure for
case 2 is shown in Table 6.6. The polytropic head deviates 1% when the inlet or outlet
pressure deviate 0.747%, which corresponds to a discrepancy of 0.0071 bar for the inlet
pressure and 0.0149 bar for the outlet pressure. To achieve a 96% precision, the inlet or
outlet pressure can deviate up to 3.023%.
Table 6.6: Sensitivity of polytropic head with varying pressure Case 2

Relative deviation
Hp [%] p1 [%] p1 [bar] p2 [%] p2 [bar]
0,1 0,074 0,0007 0,075 0,0015
0,5 0,373 0,0035 0,373 0,0075
1 0,747 0,0071 0,747 0,0149
2 1,500 0,0143 1,500 0,0300
4 3,023 0,0287 3,023 0,0605
49
A summary of the sensitivity analysis for polytropic head with variation in temperature
for case 2 is shown in Table 6.7. The polytropic head deviates 1% when the inlet
temperature deviate 2.102%, which corresponds to a discrepancy of 6.20 K. The outlet
temperature can deviate up to 1.919%, 7.43 K, to achieve the same accuracy for the
polytropic head. To achieve a 96% precision, the inlet temperature can deviate up to
8.499%, while the outlet temperature can deviate up to 7.746%.
Table 6.7: Sensitivity of polytropic head with varying temperature Case 2

Relative deviation
Hp [%] T1 [%] T1 [K] T2 [%] T2 [K]
0,1 0,209 0,62 0,191 0,74
0,5 1,049 3,10 0,958 3,71
1 2,102 6,20 1,919 7,43
2 4,220 12,45 3,850 14,90
4 8,499 25,07 7,746 29,98
The deviations for polytropic efficiency are highly different from case 1, as shown by the
tables in Appendix G.2. For case 1 a 1.18% increase in inlet temperature results in a
100% increase in polytropic efficiency. For case 2 the same increase in inlet temperature
only gives an 11.5% increase for the efficiency.
6.3 Wet Gas Considerations
The presence of liquid in a compression process may decrease the discharge temperature
and specific gas volume, as described in chapter 5.7. Accurate pressure and temperature
measurements are thus essential for wet gas compression analysis. The inlet fluid
condition will determine the degree of evaporate cooling or condensation through the
compression process. [7]
50
Due to the low pressures involved for the actual compressor test, the polytropic head and
efficiency are highly sensitive to uncertainties in pressure measurements. The polytropic
head deviates 1% when the inlet or outlet pressure deviate 0.061%. The sensitivity for
polytropic efficiency is identical to the sensitivity of polytropic head for pressure
measurement uncertainties.
Uncertainties in temperature measurements will only slightly influence the polytropic

head for the actual test. The polytropic head deviates 1% when the inlet temperature
deviates 2.015%. The outlet temperature can deviate up to 1.999% for the same deviation
in head. The polytropic efficiency for case 1 is highly sensitive to uncertainties in
temperature measurements. A 1.18% increase in inlet temperature results in a 100%
increase in polytropic efficiency.
For a compression process with higher outlet pressure and temperature, the polytropic
procedures are less sensitive. The sensitivity is particularly reduced for the polytropic
efficiency. A 1.18% increase in inlet temperature results in an 11.5% increase in
polytropic efficiency.
high, indicating that one or more parameters were measured inaccurately. The sensitivity
analysis presented shows that inaccurate temperature measurements, especially at the
compressor inlet, will highly influence the efficiency calculation for compressors at low
pressure and temperature operating condition.
51
52
Chapter 7
7. Aerodynamic Instability
7.1 Performance Characteristics
The slope of the compressor characteristic curve is strongly influenced by the gas
velocity relative to the impeller blades. For a typical backswept impeller the head
increases with decreasing flow due to a reduction in relative velocity. The head increase,
with decreasing flow, is what causes the basic slope to the centrifugal compressor
performance curve as shown in Figure 7.1.
Figure 7.1: Typical centrifugal compressor characteristic
Most compressors reveal specific changes in behavior and flow pattern when operated at
flow rates sufficiently below the design value. Generally a point is reached at which the
pressure rise is at a maximum and further reduction in mass flow leads to a sudden
change in the compressor flow pattern. The head curve will then be positively sloped and
have a theoretically unstable characteristic. [16]
The centrifugal compressor characteristics are determined by the geometry of the

compressor components and the compressor operating condition. Each of the compressor
components will have an optimum operating condition where its losses are at a minimum.
These losses will increase as the compressor operates further from the optimum operation
point and hence impose definite limits on the operating range of the compressor.
The common terminology for the onset of instability is the surge point, regardless of the
type of instability occurring in the compressor. A line showing the instability location for
different rotational speeds is known as the surge line. The surge line is referred to as the
compressor stability limit as shown in Figure 7.1.
53
A safety margin is set between the surge line and the closest allowable operating point to
avoid the compressor from being directed into the surge region. The surge margin,
denoted SM in Figure 7.1, is usually set to 10% from the surge line. [17] The surge
margin ensures stable operation of the compressor but may prohibit operation in regions
of highest efficiency or pressure rise.
7.2 Instability Mechanisms
The efficiency and performance of a compressor are constrained by aerodynamic

instabilities. Flow instabilities limit the stable operating range and may prohibit operation
in regions with highest efficiency. In addition to the adverse performance effects, flow
instabilities may cause unacceptable levels of subsynchronous rotor vibration. [18]
Flow instabilities are perturbations of the steady, axis-symmetric flow during normal
operating conditions. The instability mechanisms are commonly divided into three
categories; stall, rotating stall and surge. Stall and rotating stall are unsteady propagation
of flow fluctuations limited to one or few components of a compressor, while surge
accompanies the propagation and resonance of fluctuations throughout a compressor
system.
Surge
Surge is a system phenomenon that may occur in compressors due to changes in

operating conditions. Surging is associated with full flow reversal and fluctuations in
inlet and discharge pressure and temperature. This essentially one-dimensional instability
can results in a limit cycle oscillation in the compressor characteristic depending on the
amplitude of flow and pressure fluctuations as shown in Figure 7.2.
Surge may be classified according to the amplitude of the flow fluctuation. Deep surge
usually refers to the condition when full flow reversal occurs. The aerodynamic pulsation
caused by surging is transmitted throughout the compression system.
Operation during surge may result in extensive loss of performance and efficiency. Flow
reversal may in addition result in reverse bending on compressor components and high
radial and axial vibrations.
54
Figure 7.2: Compressor characteristic with surge cycle
Stall
Stall is a local instability phenomenon in which a region of stagnant flow occurs within
one or few components of the compressor. Stationary stall cells form in vaned diffusers,
return channels, guide vanes etc. The phenomenon may occur at any point on the
compressor characteristic curve but is more common at very high or very low flow rates.
[19]
Rotating Stall
Rotating stall refers to the class of stall cells in which one or more regions of stagnant
flow rotate around the circumference of the compressor. The most common forms of
rotating stall occur in impellers or diffusers at very low flow rates. [19]
Both impeller and diffuser rotating stall can have significant effects on mechanical and
aerodynamic performance. The flow separation and the high losses associated with stall
cells may result in degraded performance. [20] Rotating stall cells may in addition subject
the rotor to unbalanced pressure forces and hence cause unacceptable levels of
subsynchronous vibration of the rotor. [18]
Impeller Rotating Stall
There exist numerous theories in the literature to the nature and origin of impeller
rotating stall. Probable causes of the phenomenon are flow separation near the impeller
exit, high incidence angles at the impeller leading edge or pressure disturbances caused
by the impeller blade geometry.
Off-design operation changes the incident angle to the impeller blades and flow
separation may occur. Breakdown of the flow in one impeller channel will cause a
deflection to the incoming flow. The deflection will cause the incidence to increase to
one side and reduce it on the other. The blade with increased incidence will tend to stall
55
at the suction side of the blade leading edge. [20] The flow separation will continuously
shift around the impeller blades in the direction in which incidence are increased as
shown in Figure 7.3.
Figure 7.3: Rotating stall in a centrifugal impeller [17]
Diffuser Rotating Stall
Compressor performance during rotating stall is often dominated by the flow field in the
diffuser, which may be highly non-uniform [21] There is always a tendency for the flow
to break away from the boundary in a diffusing process and hence the radial diffuser has
an important role in establishing the overall efficiency and pressure-rise of a centrifugal
compressor. The diffuser may be the component limiting the operating range, depending
on the diffuser design and its matching to the impeller. A distorted flow field from the
impeller will affect the diffuser inlet condition and hence have a strong effect on
compressor performance in the unstable operating region. [21]
The pressure field in the diffuser inlet is highly sensitive to changes in flow rate. [16] The
onset of diffuser rotating stall is strongly influenced by the diffuser flow angle, which is a
function of the diffuser geometry, gas Reynolds Number and compressor operating
conditions. [21] The diffuser flow path is shown in Figure 7.4.
56
Figure 7.4: Flow through a vaneless diffuser [17]
When the flow angle exceeds some critical angle, the diffuser walls will dissipate the
flow momentum of the gas by friction to the point where frictional forces are increasing
faster than the compressor head. The growth in boundary layer thickness results in
increased blockage in the diffuser channels. The diffuser pressure recovery and stability
decreases as inlet blockage increases. [22] Off-design operation continuously changes the
diffuser flow angle and results in periodic flow unsteadiness in the diffuser inlet.
Matching between Components
Instability phenomena are often more complicated in radial compressors due to matching
between compressor elements. The more disturbed the impeller exit flow profile, the
greater the likelihood that the downstream diffuser will stall. Stall cells formed in the
diffuser may excite the formation of rotating stall in the impeller. [21]
Rotating stall usually occurs prior to surge, but may also exist in the nominally stable
operating range. [18] Surge is usually the flow instability that is of most concern in
centrifugal compressors since the centrifugal compressor may operate fairly satisfactory
with rotating stall present. Analyzing the instability effect of each compressor component
is important to determine the total system instability.
7.3 Wet Gas Impact on Compressor Stability
The presence of liquid will alter the flow regime within the compressor. Multiphase flow
within wet gas compressors is expected to be annular due to the high gas volume fraction
and the high flow velocity involved. [23] The annular flow pattern will consists of a thin
liquid film and dispersed droplets in the core gas flow.
57
Identification of the liquid distribution is highly important in understanding the different
mechanisms leading to compressor instability. Multiphase flows in compressors have
complicated characteristics including interfacial interactions and relative movement
between phases. [23]
Several parameters are considered to influence the onset and severity of compressor
instability when exposed to wet gas. A reduced stall margin is expected in wet gas
compression due to premature boundary layer separation. The interface between the gas
and liquid film is characterised by a pattern of waves introducing an additional roughness
to the flow. [23] Increased blockage induced by the roughness result in a reduced flow
area and premature boundary layer separation. [23] Premature separation will impose
considerable limitation on the stability margin and hence on the operational characteristic
of the compressor. Recovery of potential energy in the diffuser will in addition be highly
disturbed by the liquid particles and premature boundary layer separation. [23]
Changes in compressor velocity distribution and flow angles contribute to reducing the
stability margin by inducing high Mach numbers. High Mach numbers are associated
with the possibility of aerodynamic shock waves followed by increased boundary layer
separation.
7.4 Test Procedure
Pressure fluctuations caused by instabilities and the resulting vibration of the rotor should
be measured with variation of flow rate and impeller rotating speed to document the
different instability phenomena.
The surge line should be approached slowly, by throttling suction or discharge flow while
maintaining compressor speed. The compressor flow rate should be varied from near
surge to choke conditions to evaluate the entire operating range of the compressor.
Determination of head, flow and efficiency close to the surge line may be highly
inaccurate due to fluctuating measurement readings.
58
7.5 Instrumentation
Dynamic Pressure Measurements
Dynamic pressure transducers should be installed in the inlet and outlet piping for
detection of pressure pulsations throughout the compressor system.
In order to investigate the unstable characteristics within different compressor

components, internal instrumentation should be implemented. Circumferentially
distributed pressure transducers at the locations shown in Figure 7.5 are recommended.
Figure 7.5: Measurement locations for internal instrumentation [24]
The recommended positions for the pressure transducers at the entrance to the vaneless
diffuser, position 3, are shown in Figure 7.6. [24] Due to narrow flow channels in the
compressor the sensors must have as small dimensions as possible and be flush mounted
in the shroud wall to minimize disturbances accumulated from the installed sensors.
59
Figure 7.6: Circumferential measuring positions in the diffuser [24]
Measurement of Radial Vibration
Vibration probes installed at each end of the rotor are recommended to obtain
measurements of the rotor vibration. The Kitsler sensors are suitable for such
measurements.
7.6 Pressure Characteristic Analysis
The pressure characteristics can be utilized to document the compressor behavior and to
indicate whether surge or rotating stall occur. Compressor surge appear as large high to
low pressure fluctuations in the pressure characteristic at compressor inlet/discharge,
whereas rotating stall displays fluctuations more frequently, as shown in Figure 7.7.
Figure 7.7: Stall and surge pressure characteristics at compressor inlet/discharge
60
Figure 7.8 shows a typical pressure characteristic. An indication of rotating stall is seen
as steadily increasing pressure amplitude just prior to the first surge cycle.
Figure 7.8: Pressure fluctuations with time [25]
Since rotating stall is a local pressure disturbance, in contrast to surge which cause
fluctuations throughout the compressor system, pressure characteristic deduced from
internal measurements within a diffuser can determine the type of instability that occur.
Surge differs from rotating stall by simultaneous occurrence of high to low pressure
fluctuations on all the pressure transducers located circumferentially in the diffuser.
7.7 Frequency Spectrum Analysis
Pressure Fluctuation Frequency
The frequency content of a pressure reading can be analyzed by utilizing a fast Fourier
transformation (FFT). The relative magnitude of the dominant instability frequency is
compared to the magnitude related to the blade passing frequency to determine type of
instability that is present. [26] Rotating stall appears as a broad and varied peak at more
than one frequency in the FFT spectrum, whereas surge is seen as a distinct frequency.
[26]
Radial Vibration Frequency
M. Kita et al. [18] found a correlation between the pressure fluctuation and the rotor
vibration during rotating stall. The rotor vibration amplitude is proportional to the
pressure fluctuation amplitude and may therefore be an indication of compressor
instability. The increased vibration is in most cases due to the onset of rotating stall rather
than full surge.
Impeller and diffuser rotating stall generally manifest themselves in different frequency
ranges. By analyzing the frequency spectrum of the radial vibrations one may identify the
61
type of rotating stall that occurs. The subsynchronous radial vibration frequency for
impeller rotating stall lies in the range of 50-80% of the compressor running speed, while
the range for diffuser rotating stall is 6-33%. [19]
A determination of whether the radial vibration is flow sensitive can be an indicator when
identifying the flow instabilities. There will be a distinct flow rate at which the impeller
rotating stall will come and go. Unlike when impeller stall is present, diffuser rotating
stall will appear at some distinct flow rate but will not disappear by some reduction in
flow rate. A significant increase in flow rate is necessary to get rid of the stall cells.
Since various rotordynamic elements may yield vibration signals that approximate stall
characteristics, it is important to rule out such rotor effects before attributing
subsynchronous rotor vibration to rotating stall. [19] Vibration due to rotordynamic
elements can be observed with relatively constant response frequency despite changes in
discharge pressure, in contrast to instability caused vibration.
62
The efficiency and performance of a compressor are constrained by aerodynamic

instabilities. In addition to the adverse performance effects, instabilities may cause
unacceptable levels of subsynchronous rotor vibration. A reduced stall margin is expected
in wet gas compression due to premature boundary layer separation.
Pressure fluctuations caused by instabilities and the resulting vibration of the rotor should
be measured with variation of flow rate and impeller rotating speed to document the
different instability phenomena over the entire compressor operating range.
Dynamic pressure transducers in the inlet and discharge piping are recommended for
detection of pressure pulsation throughout the compressor system. Circumferentially
distributed pressure transducers at different locations within the compressor components
are recommended to document the internal flow behavior. Vibration probes installed at
each end of the rotor are recommended to obtain vibration measurements.
By analyzing the frequency spectrum for the pressure fluctuations and radial vibrations
one can identify the cause and type of instability phenomenon that occurs. The frequency
content can be analyzed by utilizing a fast Fourier transformation (FFT). Rotating stall
appears as a broad and varied peak at more than one frequency in the FFT spectrum,
whereas surge is seen as a distinct frequency. [26] The subsynchronous radial vibration
frequency for impeller rotating stall lies in the range of 50-80% of the compressor
running speed, while the range for diffuser rotating stall is 6-33%. [19]
63
64
Chapter 8
8. Multiphase Flow Measurements and Visualization

Techniques
Complex flow structure with local randomness is characteristic in multiphase flow

systems. The understanding of such complex flow patterns requires reliable data over the
entire flow field, which in turn depends on the implementation of proper measuring and
visualization techniques.
Visualization techniques with careful control of the flow distribution are desired in wet
gas compression to document aerodynamic instability under various compressor
operating conditions and to identify the dynamic changes occurring in the flow field prior
to rotating stall and surge.
Techniques commonly used for visualization of single-phase flows may be difficult to

implement in multiphase compressors due to the inherent non-homogeneity of the
multiphase flow field. A thorough evaluation of the visualization techniques is essential
to outline their range of applicability, along with advantages and limitations inherent to
each technique.
8.1 Direct Visualization
Direct visualization of the flow field may provide useful information about the flow
behavior within the compressor components. Identification of the mechanism leading to
degraded performance and operational limitations may be obtained by rendering certain
flow characteristics directly accessible to visual perception. The prevalent flow regime,
the liquid distribution, and the interaction between the phases can then be documented
and further analyzed.
The compressor flow field can be visualized directly by replacing sections of the
compressor casing with transparent material, Plexiglas, as shown in Figure 8.1.
65
Figure 8.1: Locations for transparent windows
Since the compression fluid is transparent and its flow pattern may not be seen directly,
an injection of particles or dye which scatters light should be inserted to better visualize
the fluid motion.
A high-speed camera that captures the scattered light from the particles can be used to
reconstruct a stereoscopic view of the flow. Quantitative data from the flow field can be
derived by analyzing the obtained picture. Computers that allow calculation in reasonable
time, and with the high number of operations required by image processing, must be
utilized.
High seed particle concentration is recommended for compressor flow field studies since
high spatial resolution measurements are required. [25] Determination of optimal particle
size and density is a major challenge when utilizing tracer particles for flow visualization
in compressors. The particle size should be small enough to accurately follow the flow
through the compressor channels. The difference between the fluid and the particle
motion can be minimized by giving the particles a density almost coinciding to that of the
fluid. The method gives accurate results in stationary flow, but may fail in unsteady and
transient flows due to the finite size of the particles.
A disadvantage with particle tracer methods is that the injection process and the injected
material may influence the flow field. Flush-mounted injection probe in a close distance
to the measuring point will minimize the probe disturbances. Window deposition of
particles is an obstacle to tracer-based measurement techniques since the seeding
particles have a tendency to adhere to the transparent windows.
66
Multiphase Application
The observer may not be able to fully visualize the different wet gas flow phenomena due
to liquid film formation on the Plexiglas windows. The liquid deposition on the
transparent windows depends on the GVF and the relative concentration of dispersed
droplets in the gas core flow.
Particle tracer methods in wet gas compression have a challenge owing to the varying
density of the phases involved and to the turbulent and unsteady state of the fluids. This
difficulty is further discussed in chapter 8.2 when laser measurement techniques are
addressed for multiphase applications. An idea of the origin and development of the flow
regime and phase distribution can nevertheless be obtained.
8.2 Laser Measurement Techniques
Laser measurement techniques are suitable for flow investigation in the narrow flow
channels of a centrifugal compressor. [27] Velocimetric measurements within a
centrifugal impeller, diffuser and volute can provide information of the main features of
the fluid flow. The volume in which the measurements are made is only limited by
optical means so that measurements can be made even in boundary layers and at places
that cannot be reached by probes.
Laser measurement techniques are limited to transparent measurement locations. Optical

access to the internal flow of a centrifugal compressor can be provided as described for
direct visualization. Figure 8.2 shows an optical window utilized for PIV measurements
in a vaned diffuser. The technical limitations caused by the surrounding of the
compressor may be an obstacle when implementing the optical apparatus.
Figure 8.2: Test section of a diffuser with an optical window [28]
67
Laser velocimeters enable non-intrusive measurements of the flow field. Unlike the probe
measurement methods, the laser velocimetric measurements are independent of the
properties of the fluid and environmental conditions such as temperature, pressure and
density.
Laser Doppler and Laser two-focus Velocimetry
Laser Doppler Velocimetry (LDV) and Laser two-focus velocimetry (L2F) have been
widely used for flow measurements in compressors. LDV is a technique for measuring
the direction and speed of fluids. By analyzing the Doppler-equivalent frequency of the
laser light scattered by the seeded particles within the flow, the velocity field of the fluid
can be determined.
Laser-two-focus measurements (L2F) collects light scattered from particles convected by

the flow. The time-of-flight of the particles between two focused laser spots is measured.
In contrast to the LDV, the L2F velocimetry is a statistical based measurement system;
hence the temporal fluctuations of the flow cannot be measured.
Particle Image Velocimetry
Particle Image Velocimetry, PIV technique, has major features of other velocimetric
methods. The technique can present extensive velocity fields in a very short time and can
provide simultaneous and multipoint measurements. Hayami et al. [27] performed
detailed flow measurements within a centrifugal impeller using PIV. Ibaraki et al. [28]
applied PIV for vaned diffuser measurements.
PIV is an imaging technique used to measure instantaneous velocities and related

properties in fluid flows. The PIV method measures the velocity components across a
planar 2D matrix of points within a flow field in order to produce a 2D velocity vector
field.
The fluid must be seeded with tracer particles assumed to accurately follow the flow. By
illuminating the particles with a laser light sheet, an instantaneous image is produced.
The positions of the moving particles are recorded with a CCD camera.
To obtain velocity information the laser must be pulsed at least two times at a known
time separation, creating an image of the laser light sheet in which the particles appear as
multiples spots. The separation of the recorded spots is a measure of the distance traveled
by the particles between the laser light pulses.
Typical PIV apparatus consists of a digital camera, a high power laser, for example a
double-pulsed Nd:YAG laser and an optical arrangement to convert the laser output light
to a light sheet. A PIV measurement system located at the compressor inlet in shown in
Figure 8.3.
68
Figure 8.3: PIV measurement system [29]
Laser measurements within wet gas compressors are more challenging than for single-
phase compressors. Liquid film formation and possible window deposition of particles
during the compressor testing can prevent a clear optical path to the measurement zone.
No particle seeding is required if the dispersed phase itself is the object of interest. [30]
Optical properties such as light refraction changes in sections within the flow where there
are big local differences in flow density. Fluid density is a function of the refractive index
of the flowing medium, which is a measure of how much the speed of light is reduced
inside the medium.
If the phase velocities are to be independently measured, to determine local slip velocities
and superficial velocities, the gas flow must be seeded and some discrimination between
the seed particles and the dispersed phase is necessary. [30]
A velocity map of the flow field can be obtained from PIV measurements, but it may be
difficult to identify which velocity vectors correspond to which phase of the flow. Towers
et al. [31] examined the use of fluorescent particles for two-phase flow measurements.
The technique operates by identifying fluorescent tracer particles inserted in one phase of
the flow in order to separate the particle images obtained from both phases. Fluorescent
tracer particles are appropriate for high speed unsteady gas flows and hence appropriate
for studying non-stationary flow field during compressor surge. [30]
PIV measurements at the compressor inlet can give an indication of the phase behavior,
liquid distribution and possible development of the flow field. PIV apparatus are more
easily implemented at the compressor inlet, in contrast to compressor components
downstream, due to insufficiently supply of seeding particles downstream. The
measurements may in addition be more accurate at the inlet due to the highly disturbed
and transient flow field in the downstream components.
69
8.3 Pressure Sensitive Paint Measurement Technique
Pressure Sensitive Paint (PSP) techniques can be adopted for detailed measurements in
the compressor components where optical access is limited. Visualization of the unsteady
pressure field in the diffuser at stall and surge conditions will provide useful information
when documenting compressor aerodynamic instabilities. Figure 8.4 shows a typical PSP
measurement system.
Figure 8.4: PSP measurement system [29]
PSP consists of luminescent molecules dispersed in an oxygen permeable binder. The

mixture is applied to the compressor surface under investigation. Ultraviolet light is
beamed onto the surface, raising the molecules to a higher energy level. The binder emits
light at a visible and longer wavelength to return to the original energy state. A CCD
camera, capturing the intensity of emission from the luminescent molecules, as shown in
Figure 8.5, obtains images of the surface pressure distribution. The small size of the PSP
molecules allows for detailed pressure data to be obtained inside the diffuser.
PSP measurements exploit the oxygen sensitivity of the luminescent molecules dispersed
within the binder. [32] The excited molecules are deactivated with oxygen, making the
emission intensity proportional to the amount of oxygen in the nearby environment.
Higher pressures fields mean that more oxygen molecules are being squeezed into the
flow which means lower intensity of light being emitted.
The Stern-Volmer equation, given in equation (8.1), can be used for intensity-based
measurements. The calibration coefficient A and B are temperature-dependent and can be
obtained experimentally for a given PSP formulation. [32]
p
I ref = A(T ) + B(T ) (8.1)
pref
70
Figure 8.5: The schematic PSP measurement
There are some probable difficulties when utilizing the PSP technique for visualization in
a wet gas compressor. Liquid film formation on the optical window and on diffuser
surface will influence the intensity-based measurements. Dispersed droplets in the
gaseous phase are likely to disturb the excitation and emission of light and adversely
affect the achievable measurement accuracy.
An incompressible liquid film on the binder surface can hinder the oxygen sensitive
molecules to react to different pressure levels. The intensity of the emitted luminescence
captured by the CCD camera will then be inaccurate due to the reduced amount of
oxygen available to deactivate the exited molecules.
8.4 CFD
Computational fluid dynamics (CFD) uses numerical methods and algorithms to simulate
and analyze behavior of fluids under a wide range of conditions. The objective is to
analyze the simulated vector field to identify features such as turbulence, vortices, and
other forms of structure in the compressor flow field. Computers are utilized to perform
the calculations required to simulate the interaction of fluids and gases in a complex flow
field.
Although computational fluid dynamics can obtain predictive solutions for the complex
flow pattern in compressors, experimental measurement should always be performed to
validate the computed values.
71
Visualization of the flow field can provide useful information about the flow behavior
within the compressor components. Pressure and velocity fluctuations caused by
aerodynamic instabilities can be measured and visualized to document instability
phenomena.
The visualization methods presented require optical access to the internal flow. This can
be accomplished by replacing sections of the compressor casing with Plexiglas. Liquid
film formation on the Plexiglas windows may prevent this optical access and cause
inaccurate measurements when utilizing the techniques in wet gas compressors.
Laser measurement techniques are recommended for flow investigation in centrifugal

compressors. PIV techniques can present extensive velocity fields in a very short time
and can provide simultaneous and multipoint measurements, which is of importance
when documenting aerodynamic instabilities.
No particle seeding is required if the dispersed phase itself is the object of interest. If the
phase velocities are to be independently measured, the gas flow must be seeded and some
discrimination between the seed particles and the dispersed phase is necessary.
Fluorescent tracer particles are recommended for investigating the non-stationary flow
field during both single- and multiphase compressor instability.
PSP measurements are suitable for obtaining information of the pressure distribution
during single-phase compression, but will be highly inaccurate for wet gas applications.
The intensity-based measurements will be influenced by liquid film formation on the
binder surface and disturbed by the dispersed droplets in the gaseous phase.
72
Chapter 9
9. Conclusion
The ideal polytropic performance analysis is suitable for the actual compressor test. The
values estimated for the polytropic exponent and polytropic head by PRO/II agree closely
with the calculated values determined from the ideal gas equations and thus verify the use
of ideal polytropic calculation procedures for the compressor test.
When converting the performance results for the compressor test to real gases at high
pressure and temperature conditions, the ideal assumptions will not be applicable. The
sensitivity of the ideal polytropic head with deviations in polytropic exponent is notably
low for the actual compressor test, but will be substantially higher for greater pressure
ratios. The Schultz polytropic analysis must then be utilized to achieve accurate results.
The probability for phase changes to occur is limited for the planned wet gas compressor
test due to the low compressor pressure ratio and the atmospheric inlet condition, and can
thus be neglected in the performance evaluation. Performance procedures based on
polytropic assumptions will then be accurate. For higher pressure ratios and with real
gases the direct integration procedure must be implemented to account for any phase
transitions along the compression path.
high, indicating that one or more parameters were measured inaccurately. The sensitivity
analysis presented shows that inaccurate temperature measurements, especially at the
compressor inlet, will highly influence the efficiency calculation for compressors at low
pressure and temperature operating condition.
The polytropic head and efficiency are highly sensitive to uncertainties in the pressure
measurements. The polytropic head deviates 1% when the inlet or outlet pressure deviate
0.061%. Uncertainties in temperature measurements will only slightly influence the
polytropic head, but have great influence on the polytropic efficiency. The polytropic
head deviates 1% when the inlet temperature deviate 2.015%. The outlet temperature can
deviate up to 1.999% for the same deviation in head. A 1.179% increase in inlet
temperature results in a 100% increase in polytropic efficiency.
For a compression process with higher outlet pressure and temperature, the polytropic
procedures are less sensitive. The sensitivity is particularly reduced for the polytropic
efficiency. A 1.18% increase in inlet temperature results in an 11.5% increase in
polytropic efficiency.
73
A reduced stall margin is expected in wet gas compression due to premature boundary
layer separation. In addition to the adverse performance effects, instabilities may cause
unacceptable levels of subsynchronous rotor vibration.
The pressure fluctuations caused by instabilities and the resulting vibration of the rotor
should be measured with variation of flow rate and impeller rotating speed to document
the different instability phenomena over the entire compressor operating range. The
frequency content can be analyzed by utilizing a fast Fourier transformation (FFT).
Dynamic pressure transducers in the inlet and discharge piping are recommended for
detection of pressure pulsation throughout the compressor system. Circumferentially
distributed pressure transducers at different locations within the compressor components
are recommended to document the internal flow behavior. Vibration probes installed at
each end of the rotor are recommended to obtain vibration measurements.
Pressure and velocity fluctuations caused by aerodynamic instabilities can be visualized

to document instability phenomena. Laser measurement techniques are recommended for
flow investigation in centrifugal compressors. PIV techniques can present extensive
velocity fields in a very short time and can provide simultaneous and multipoint
measurements, which is of importance when documenting aerodynamic instabilities.
Optical access can be accomplished by replacing sections of the compressor casing with
Plexiglas. Fluorescent tracer particles are recommended for investigating the non-
stationary flow field during both single- and multiphase compressor instability. No
particle seeding is required if the dispersed phase itself is the object of interest. If the
phase velocities are to be independently measured, the gas flow must be seeded and some
discrimination between the seed particles and the dispersed phase is necessary.
74
Chapter 10
10. Recommendations for Further Work
Since only one test at part-load has been conducted, several tests at part- and full-load
must be carried out to document and validate the performance procedure over the entire
operating range.
The effects and causes of measurement uncertainties should be further investigated in

order to increase the accuracy of the performance calculations.
At higher pressures and temperatures the ideal polytropic performance analysis is not
suitable for the evaluation of the compressor. The Schultz calculation procedures should
be implemented in the performance analysis to render possible a direct comparison of the
compressor performance at different operating conditions.
Testing at wet gas conditions must be conducted to determine the accuracy and validity
for the performance procedures and measurements techniques. In order to determine the
suitability for the performance procedures, the GERG equation of state should be
implemented and validated for wet gas applications.
By implementation of dynamic pressure transducers and PIV apparatus, additional insight

into the origin and cause of flow instabilities can be obtained.
75
76
Bibliography
[1] American Society of Mechanical Engineers, Performance Test Code on

Compressors and Exhausters, ASME PTC 10, 1997
[2] S.C. Amundsen, Wet Gas Compression, NTNU, Project Thesis, 2008
[3] J.M. verli, Strmningsmaskiner, Termiske maskiner, Bind 3, 1992
[4] International Organization for Standardization, Turbocompressors

Performance test code, ISO 5389:1992, 1992
[5] J.M. Schultz, The Polytropic Analysis of Centrifugal Compressors, Journal of

Engineering for Power, pp. 69-82, 1962
[6] . Hundseid, L.E. Bakken, T. Helde, A Revised Compressor Polytropic

Performance Analysis, ASME GT2006-91033, 2006
[7] . Hundseid, L.E. Bakken, Wet Gas Performance Analysis, ASME GT2006-
91035, 2006
[8] C. Twu, H. Kusch, Selection of Equations of State Models for Process

Simulator, Simsci Inc., 1994
[9] R. Kurz, K. Brun, D.D Legrand, Field Performance Testing of Gas Turbine
Driven Compressor Sets, Proceedings 28th Texas A&M Turbomachinery
Symposium, 1999
[10] American Society of Mechanical Engineers, General Instructions, ASME PTC

1, 1999
[11] American Society of Mechanical Engineers, Test Uncertainty, ASME PTC

19.1, 1998
[12] American Society of Mechanical Engineers, Pressure Measurement, ASME

PTC 19.2, 1998
[13] International Organization for Standardization, Measurement of Fluid Flow by

Means of Pressure Differential Devices Inserted in Circular Cross-Section
Conduits Running Full, ISO 5167-1:2003, 2003
[14] M.J. Moran, H.N. Shapiro, Fundamentals of Engineering Thermodynamics, 5th

Edition, 2004
77
[15] L.E. Bakken, Status og utfordringer innen flerfase maskiner, Norske
Sivilingenirers Forening, 1993
[16] R. Hunziker, G. Gyarmathy, The Operational Stability of a Centrifugal

Compressor and Its Dependence on the Characteristics of the Subcomponents,
ASME Journal of Turbomachinery, Vol. 116, 250-257, 1994
[17] M.T. Gresh, Compressor Performance: Aerodynamic for the User, Newnes,
2001
[18] M. Kita, S. Iwamoto, D. Kiuchi, R. Kawashita, Prediction of Subsynchronous

Rotor Vibration Amplitude Caused by Rotating Stall, Proceedings of the Thirty-
seventh Turbomachinery Symposium, 97-102, 2008
[19] J.M. Sorokes, Rotating Stall An Overview of Dresser-Rand Experience,

Dresser-Rand
[20] H.I.H. Saravanamutto, H. Cohen, G.F.C Rogers, Gas Turbine Theory, 5th
edition, 2001
[21] K.U. Ziegler, H.E. Gallus, R. Niehuis, A Study on Impeller-Diffuser Interaction

Part I: Influence on the Performance, ASME Journal of Turbomachinery, Vol.
125, 173-182, 2003
[22] V.G. Filipenco, S. Deniz, J.M. Johnston, E.M. Greitzer, N.A. Cumpsty, Effects
of Inlet Flow Field Conditions on the Performance of Centrifugal Compressor
Diffusers: Part 1 - Discrete-Passage Diffuser, ASME Journal of
Turbomachinery, Vol. 122, 1-10, 2000
[23] T.G. Grner, L.E. Bakken, L. Brenne, T. Bjrge, An Experimental Investigation

of Airfoil Performance in Wet Gas Flow, ASME GT2008-50483, 2008
[24] S. Mizuki, Y. Oosawa, Unsteady Flow within Centrifugal Compressor Channels

Under Rotating Stall and Surge, ASME Journal of Turbomachinery, Vol. 114,
312-320, 1992
[25] M.P. Wernet, M.M. Bright, G.J. Skoch, An Investigation of Surge in a High-
Speed Centrifugal Compressor Using Digital PIV, ASME Journal of
Turbomachinery, Vol. 123, 418-428, 2001
[26] M. Schleer, S.J. Song, R.S. Abhari, Clearance Effects on the Onset of Instability
in a Centrifugal Compressor, ASME Journal of Turbomachinery, Vol. 130,
2008
78
[27] H. Hayami, M. Hojo, S. Aramaki, Flow Measurements in a Transonic
Centrifugal Impeller Using a PIV, Journal of Visualization, Vol. 5, Number 3,
255-261, 2002
[28] S. Ibaraki, T. Matsuo, T. Yokoyama, Investigation of Unsteady Flow Field in a

Vaned Diffuser of a Transonic Centrifugal Compressor, ASME Journal of
Turbomachinery, Vol. 129, Issue 4, 686-693, 2007
[29] H. Hayami, M. Hojo, S. Aramaki, Flow Measurement in a Transonic

Centrifugal Impeller Using a PIV, The Visualization Society of Japan and
Ohmsha, LTD, Journal of Visualization, Vol. 5, No. 3, 255- 261, 2002
[30] H.E. Albrecht, M. Borys, N. Damaschke, C. Tropea, Laser Doppler and Phase
Measurement Techniques, ISBN 3540678387, 9783540678380, Springer 2003
[31] D.P. Towers, C.E. Towers, C.H. Buckberry, M. Reeves, A Colour PIV System
Employing Fluorescent Particles for Two-Phase Flow Measurements, Meas.
Sci. Technol. 10, 824-830, 1999
[32] K.R. Navarra, D.C. Rabe, S.D. Fonov, L.P. Goss, C. Hah, The Application of
Pressure- and Temperature-Sensitive Paints to an Advanced Compressor,
ASME Journal of Turbomachinery, Vol. 123, 823-829, 2001
79
80
Appendix A Performance Analysis
A.1 Polytropic Analysis
The polytropic exponent is given by equation (A.1).
p2
ln
p1
n= (A.1)
v
ln 1
v2
The polytropic exponent can be determined when inlet and discharge pressure and
temperature is known by the thermodynamic relationship given in equation (A.2).
n 1

T2 p2 n
= (A.2)
T1 p1
The efficiency can be calculated when suction and discharge pressure and temperature is
known, as shown in equation (A.3).
p
ln 2
1 p1
p = (A.3)
T
ln 2
T1
In cases of known polytropic efficiency, equation (A.4) can be utilized to determine the
polytropic exponent.
1 n
p = (A.4)
n 1
81
A.2 Schultz Polytropic Analysis
Compressibility functions X and Y are included in the procedure to supplement the

compressibility factor Z for changes in fluid properties. When utilizing the
compressibility functions in the performance analysis, the polytropic head and efficiency
are determined explicitly from the performance equations. Values for the compressibility
factors can be obtained from Schultzs compressibility charts. The definitions of X and Y
is shown in equation (A.5) and equation (A.6).
T dv
X = 1 (A.5)
v dT P
p dv
Y = (A.6)
v dp T
Schultz defined a polytropic relationship as shown in equation (A.7), where the Schultz
polytropic volume exponent is given in equation (A.8).
pv nv = constant (A.7)
(1 + X )
nv = (A.8)
1 1 1
Y + X 1
p
p

A correction factor, as defined in equation (A.9), is introduced to account for the slight
variation in nv. The volume corrected isentropic exponent is given in equation (A.10).
( v 1)( h2 s h1 )
f = (A.9)
v ( p2 v2 s p1v1 )

v = (A.10)
Y
82
Schultz polytropic head becomes as shown in equation (A.11).
n
H p , S = f v [ p2 v2 p1v1 ] (A.11)
nv 1
By introducing equation (A.7) and pv=RT in combination with equation (A.11), Schultz
polytropic head becomes as shown in equation (A.12).
nv 1

nv p2 nv
H p,S = f * Z1 RT1 1 (A.12)
nv 1 p1

The polytropic head determined from equation (A.11) and equation (A.12) is only equal
if the polytropic volume exponent is defined as in equation (A.1).
83
A.3 Polytropic Exponent
When Schultz compressibility functions X and Y approach respectively zero and one,
Schultz polytropic exponent approach the ideal polytropic exponent.
Ideal Polytropic exponent, deduced from equation (A.4):
p
n=
1 + p
Schultz polytropic exponent:
(1 + X )
nv =
1 1 1
Y + X 1

p p
X = 0
Y = 1
(1 + 0 ) 1 p p
nv = = = =
1 1 1 1 1 p 1 (1 p ) 1 + p
1* + 0 1

p p p p
84
Appendix B Verification of Discharge Coefficient and
Expansibility Factor
Discharge Coefficient
The Reader-Harris/Gallagher equation for calculating the discharge coefficient is given in

equation (B.1). [53]
0,7
106
C = 0,5961 + 0, 0261 0, 216 + 0, 000521
2 8

Re D
0,3
106 4
+ ( 0, 0188 + 0, 0063 A ) 3,5
+ ( 0, 043 + 0, 080e 10 L1
0,123e 7 L1
) (1 0,11A) 1 4
Re D
0, 031( M 2 0,8M 21,1 ) 1,3
(B.1)
L1 (l1/d) and L2 (l2/d) are the quotients of the distance from the upstream and downstream
tapping to the upstream and downstream face of the plate and pipe diameter. Due to the
orifice corner tapping utilized in the compressor test these quotients become zero. M2
((2L2)/(1-)) will in addition be zero as a consequence of the corner tapping. The
simplified Reader-Harris/Gallagher equation is given in equation (B.2).
0,7
106
C = 0,5961 + 0, 0261 0, 216 + 0, 000521
2 8

Re D
0,3
106 4
+ ( 0, 0188 + 0, 0063 A ) 3,5
+ ( 0, 043 + 0, 080 *1 0,123*1)(1 0,11A )
Re D 1 4
(B.2)
The discharge coefficient is calculated for an orifice with beta value 0.6401 and for an
orifice with beta value 0.4018. The discharge coefficients are estimated at maximal and
minimal mass flow rate to show the effect of varying Reynolds number. The calculated
values are given in Table B.1.
85
Table B.1: Discharge coefficients for orifices
0,6401 [-] 0,4018 [-]
1,19 [kg/m3] 1,19 [kg/m3]

Dpipe 0,25 [m] Dpipe 0,25 [m]
2
Apipe 0,05 [m ] Apipe 0,05 [m2]
0,02 [cp] 0,02 [cp]
mmax 3,00 [kg/s] mmax 1,10 [kg/s]

3
Qmax 2,52 [m /s] Qmax 0,92 [m3/s]
Remax 763943,73 [-] Remax 280112,70 [-]
C 0,606 [-] C 0,602 [-]
mmin 1,20 [kg/s] mmin 0,50 [kg/s]

3
Qmin 1,01 [m /s] Qmin 0,42 [m3/s]
Remin 305577,49 [-] Remin 127323,95 [-]
C 0,607 [-] C 0,603 [-]
86
Expansibility Factor
Equation (B.3) shows the empirical formula for computing the expansibility factor, where
(p2/p1) denotes the pressure drop over the orifice. The calculated values are given in
Table B.2.The expansibility factor is calculated for an orifice with beta value 0.6401 and
for an orifice with beta value 0.4018.
p 1/
= 1 ( 0,351 + 0, 256 + 0,93 ) 1 2
4 8
(B.3)
p1
Table B.2: Expansibility factors for orifices
0,6401 [-] 0,4018 [-]

1,19 [kg/m3] 1,19 [kg/m3]
1,40 [-] 1,40 [-]
mmax 3,00 [kg/s] mmax 1,10 [kg/s]

3
Qmax 2,52 [m /s] Qmax 0,92 [m3/s]
Diff pressure 250,00 [millibar] Diff pressure 250,00 [millibar]
P2/p1 (orifice) 0,75 [-] p2/p1 (orifice) 0,75 [-]
0,922 [-] 0,934 [-]
mmin 1,20 [kg/s] mmin 0,50 [kg/s]

3
Qmin 1,01 [m /s] Qmin 0,42 [m3/s]
Diff pressure 34,52 [millibar] Diff pressure 45,98 [millibar]
p2/p1 (orifice) 0,97 [-] p2/p1 (orifice) 0,95 [-]
0,990 [-] 0,988 [-]
87
88
Appendix C PRO/II Setup
89
90
Appendix D Schultz Generalized Compressibility
Charts
91
92
Appendix E Valid Test Points
Test Point
Temperature Discharge, T2 29,73
Test Interval Temperature Inlet, T1 23,05
11 Pressure Discharge, p2 1,00
Pressure Inlet, p1 0,94
Speed 4983,21
Test Point
Speed 4941,49
Test Point
Speed 4989,98
Test Point
Speed 4950,66
Test Point
Speed 4937,28
Test Point
Speed 4947,21
Test Point
Speed 4948,18
93
Test Point
Speed 4957,66
Test Point
Speed 4939,39
Test Point
Speed 4951,19
Test Point
Speed 4953,22
Test Point
Speed 4965,77
Test Point
Speed 4938,86
94
Appendix F Computation of Non-Dimensional Head
and Flow Coefficients
Table F.1: Calculation of performance parameters at test condition
Measured Calculated
Test Q1 N Hp p
3
Point [m /s] [rpm] [J/kg] [-] [-]
11 0,83 4983 5505 0,55 0,0279
25 0,81 4951 5863 0,59 0,0274
35 0,77 4937 6430 0,65 0,0261
44 0,6 4948 7772 0,78 0,0203
55 0,54 4951 8045 0,81 0,0183
60 0,41 4966 8447 0,84 0,0138
95
96
Appendix G Measurement Sensitivity Analysis
G.1 Measurement sensitivity for case 1
Inlet Pressure
Table G.1: Measurement sensitivity with varying inlet pressure Case 1
Inlet Pressure, p1
Deviation p1 Deviation p1 Deviation Hp Deviation p

p1 [bar] n [-] Hp [J/kg] p [-]
[bar] [%] [%] [%]
0,940 -0,010 -1,05 1,48 6156 17,3 0,88 17,3
0,941 -0,009 -0,95 1,50 6062 15,5 0,86 15,5
0,942 -0,008 -0,84 1,51 5971 13,8 0,85 13,8
0,943 -0,007 -0,74 1,52 5880 12,1 0,84 12,1
0,944 -0,006 -0,63 1,53 5789 10,3 0,82 10,4
0,945 -0,005 -0,53 1,54 5699 8,6 0,81 8,6
0,946 -0,004 -0,42 1,56 5608 6,9 0,80 6,9
0,947 -0,003 -0,32 1,57 5517 5,2 0,78 5,2
0,948 -0,002 -0,21 1,59 5427 3,4 0,77 3,4
0,949 -0,001 -0,11 1,60 5337 1,7 0,76 1,7
0,950 0,000 0,00 1,62 5246 0,0 0,75 0,0
0,951 0,001 0,11 1,64 5156 -1,7 0,73 -1,7
0,952 0,002 0,21 1,66 5066 -3,4 0,72 -3,4
0,953 0,003 0,32 1,68 4976 -5,1 0,71 -5,1
0,954 0,004 0,42 1,70 4887 -6,9 0,69 -6,9
0,955 0,005 0,53 1,72 4797 -8,6 0,68 -8,6
0,956 0,006 0,63 1,74 4707 -10,3 0,67 -10,3
0,957 0,007 0,74 1,77 4618 -12,0 0,66 -12,0
0,958 0,008 0,84 1,80 4528 -13,7 0,64 -13,7
0,959 0,009 0,95 1,83 4439 -15,4 0,63 -15,4
0,960 0,010 1,05 1,86 4349 -17,1 0,62 -17,1
97
Outlet Pressure
Table G.2: Measurement sensitivity with varying outlet pressure Case 1
Outlet Pressure, p2

p2 [bar] n [-] Hp [J/kg] p [-]
[bar] [%] [%] [%]
1,000 -0,010 -0,990 1,84 4396 -16,2 0,62 -16,2
1,001 -0,009 -0,891 1,81 4480 -14,6 0,64 -14,6
1,002 -0,008 -0,792 1,79 4565 -13,0 0,65 -13,0
1,003 -0,007 -0,693 1,76 4651 -11,4 0,66 -11,4
1,004 -0,006 -0,594 1,74 4736 -9,7 0,67 -9,7
1,005 -0,005 -0,495 1,71 4821 -8,1 0,69 -8,1
1,006 -0,004 -0,396 1,69 4907 -6,5 0,70 -6,5
1,007 -0,003 -0,297 1,67 4992 -4,9 0,71 -4,9
1,008 -0,002 -0,198 1,65 5077 -3,2 0,72 -3,2
1,009 -0,001 -0,099 1,64 5162 -1,6 0,73 -1,6
1,010 0,000 0,000 1,62 5246 0,0 0,75 0,0
1,011 0,001 0,099 1,60 5331 1,6 0,76 1,6
1,012 0,002 0,198 1,59 5416 3,2 0,77 3,2
1,013 0,003 0,297 1,58 5501 4,8 0,78 4,8
1,014 0,004 0,396 1,56 5585 6,5 0,79 6,5
1,015 0,005 0,495 1,55 5670 8,1 0,81 8,1
1,016 0,006 0,594 1,54 5754 9,7 0,82 9,7
1,017 0,007 0,693 1,52 5838 11,3 0,83 11,3
1,018 0,008 0,792 1,51 5922 12,9 0,84 12,9
1,019 0,009 0,891 1,50 6006 14,5 0,85 14,5
1,020 0,010 0,990 1,49 6090 16,1 0,87 16,1
98
Inlet Temperature
Table G.3: Measurement sensitivity with varying inlet temperature Case 1
Inlet Temperature, T1
Deviation T1 Deviation Hp Deviation p

T1 [K] Deviation T1 [K] n [-] Hp [J/kg] p [-]
[%] [%] [%]
290,0 -5,0 -1,69 2,96 5202 -0,85 0,43 -42,2
290,5 -4,5 -1,53 2,73 5207 -0,76 0,45 -39,6
291,0 -4,0 -1,36 2,54 5211 -0,68 0,47 -36,8
291,5 -3,5 -1,19 2,37 5215 -0,59 0,49 -33,7
292,0 -3,0 -1,02 2,22 5220 -0,51 0,52 -30,4
292,5 -2,5 -0,85 2,09 5224 -0,42 0,55 -26,6
293,0 -2,0 -0,68 1,98 5229 -0,34 0,58 -22,5
293,5 -1,5 -0,51 1,87 5233 -0,25 0,61 -17,9
294,0 -1,0 -0,34 1,78 5238 -0,17 0,65 -12,6
294,5 -0,5 -0,17 1,70 5242 -0,08 0,70 -6,7
295,0 0,0 0,00 1,62 5246 0,00 0,75 0,0
295,5 0,5 0,17 1,55 5251 0,08 0,80 7,8
296,0 1,0 0,34 1,49 5255 0,17 0,87 16,9
296,5 1,5 0,51 1,43 5260 0,25 0,95 27,6
297,0 2,0 0,68 1,37 5264 0,34 1,05 40,5
297,5 2,5 0,85 1,32 5269 0,42 1,17 56,2
298,0 3,0 1,02 1,28 5273 0,51 1,31 75,9
298,5 3,5 1,19 1,24 5277 0,59 1,50 101,2
299,0 4,0 1,36 1,19 5282 0,67 1,75 134,9
299,5 4,5 1,53 1,16 5286 0,76 2,11 182,1
300,0 5,0 1,69 1,12 5291 0,84 2,63 253,0
99
Outlet Temperature
Table G.4: Measurement sensitivity with varying outlet temperature Case 1
Outlet Temperature, T2
Deviation T2 Deviation, Hp Deviation, p

[%] [%] [%]
297,0 -5,0 -1,66 1,12 5203 -0,83 2,59 247,1
297,5 -4,5 -1,49 1,16 5207 -0,75 2,07 177,8
298,0 -4,0 -1,32 1,20 5212 -0,67 1,73 131,7
298,5 -3,5 -1,16 1,24 5216 -0,58 1,48 98,8
299,0 -3,0 -0,99 1,28 5220 -0,50 1,30 74,1
299,5 -2,5 -0,83 1,33 5225 -0,42 1,16 54,9
300,0 -2,0 -0,66 1,38 5229 -0,33 1,04 39,5
300,5 -1,5 -0,50 1,43 5233 -0,25 0,95 26,9
301,0 -1,0 -0,33 1,49 5238 -0,17 0,87 16,5
301,5 -0,5 -0,17 1,55 5242 -0,08 0,80 7,6
302,0 0,0 0,00 1,62 5246 0,00 0,75 0,0
302,5 0,5 0,17 1,69 5251 0,08 0,70 -6,6
303,0 1,0 0,33 1,78 5255 0,17 0,65 -12,3
303,5 1,5 0,50 1,87 5260 0,25 0,62 -17,4
304,0 2,0 0,66 1,96 5264 0,33 0,58 -21,9
304,5 2,5 0,83 2,07 5268 0,41 0,55 -26,0
305,0 3,0 0,99 2,19 5273 0,50 0,52 -29,6
305,5 3,5 1,16 2,33 5277 0,58 0,50 -32,9
306,0 4,0 1,32 2,49 5281 0,66 0,48 -35,9
306,5 4,5 1,49 2,66 5286 0,75 0,46 -38,7
307,0 5,0 1,66 2,87 5290 0,83 0,44 -41,2
100
G.2 Measurement sensitivity for case 2
Inlet Pressure
Table G.5: Measurement sensitivity with varying inlet pressure Case 2
Inlet Pressure, P1
p1 [bar] n [-] Hp [J/kg] p [-]
[bar] [%] [%] [%]
0,940 -0,010 -1,053 1,56 73479 1,5 0,79 0,9
0,941 -0,009 -0,947 1,56 73338 1,3 0,79 0,8
0,942 -0,008 -0,842 1,56 73235 1,1 0,79 0,7
0,943 -0,007 -0,737 1,57 73131 1,0 0,79 0,5
0,944 -0,006 -0,632 1,57 73028 0,9 0,79 0,4
0,945 -0,005 -0,526 1,57 72925 0,7 0,79 0,2
0,946 -0,004 -0,421 1,57 72822 0,6 0,79 0,1
0,947 -0,003 -0,316 1,57 72720 0,4 0,79 -0,1
0,948 -0,002 -0,211 1,57 72617 0,3 0,79 -0,2
0,949 -0,001 -0,105 1,57 72515 0,1 0,78 -0,3
0,950 0,000 0,000 1,57 72412 0,0 0,78 -0,5
0,951 0,001 0,105 1,58 72310 -0,1 0,78 -0,6
0,952 0,002 0,211 1,58 72207 -0,3 0,78 -0,8
0,953 0,003 0,316 1,58 72105 -0,4 0,78 -0,9
0,954 0,004 0,421 1,58 72003 -0,6 0,78 -1,0
0,955 0,005 0,526 1,58 71901 -0,7 0,78 -1,2
0,956 0,006 0,632 1,58 71800 -0,8 0,78 -1,3
0,957 0,007 0,737 1,58 71698 -1,0 0,78 -1,5
0,958 0,008 0,842 1,58 71596 -1,1 0,77 -1,6
0,959 0,009 0,947 1,59 71495 -1,3 0,77 -1,7
0,960 0,010 1,053 1,59 71394 -1,4 0,77 -1,9
101
Outlet Pressure
Table G.6: Measurement sensitivity with varying outlet pressure Case 2
Outlet Pressure, p2
p2 [bar] n [-] Hp [J/kg] p [-]
[bar] [%] [%] [%]
1,990 -0,010 -0,500 1,58 71961 -0,6 0,78 -0,7
1,991 -0,009 -0,450 1,58 71973 -0,6 0,78 -0,6
1,992 -0,008 -0,400 1,58 72022 -0,5 0,78 -0,5
1,993 -0,007 -0,350 1,58 72071 -0,5 0,78 -0,5
1,994 -0,006 -0,300 1,58 72120 -0,4 0,78 -0,4
1,995 -0,005 -0,250 1,58 72169 -0,3 0,78 -0,3
1,996 -0,004 -0,200 1,58 72217 -0,3 0,78 -0,3
1,997 -0,003 -0,150 1,58 72266 -0,2 0,78 -0,2
1,998 -0,002 -0,100 1,58 72315 -0,1 0,78 -0,1
1,999 -0,001 -0,050 1,57 72363 -0,1 0,78 -0,1
2,000 0,000 0,000 1,57 72412 0,0 0,78 0,0
2,001 0,001 0,050 1,57 72461 0,1 0,78 0,1
2,002 0,002 0,100 1,57 72509 0,1 0,78 0,1
2,003 0,003 0,150 1,57 72558 0,2 0,79 0,2
2,004 0,004 0,200 1,57 72606 0,3 0,79 0,3
2,005 0,005 0,250 1,57 72655 0,3 0,79 0,3
2,006 0,006 0,300 1,57 72703 0,4 0,79 0,4
2,007 0,007 0,350 1,57 72752 0,5 0,79 0,5
2,008 0,008 0,400 1,57 72800 0,5 0,79 0,5
2,009 0,009 0,450 1,57 72849 0,6 0,79 0,6
2,010 0,010 0,500 1,57 72897 0,7 0,79 0,7
102
Inlet Temperature
Table G.7: Measurement sensitivity with varying inlet temperature Case 2
Inlet Temperature, T1
Deviation T1 Deviation Hp Deviation p
[%] [%] [%]
290,0 -5,0 -1,69 1,63 71824 -0,81 0,74 -5,9
290,5 -4,5 -1,53 1,63 71883 -0,73 0,74 -5,4
291,0 -4,0 -1,36 1,62 71942 -0,65 0,75 -4,8
291,5 -3,5 -1,19 1,61 72001 -0,57 0,75 -4,2
292,0 -3,0 -1,02 1,61 72060 -0,49 0,76 -3,6
292,5 -2,5 -0,85 1,60 72119 -0,41 0,76 -3,0
293,0 -2,0 -0,68 1,60 72177 -0,32 0,76 -2,4
293,5 -1,5 -0,51 1,59 72236 -0,24 0,77 -1,8
294,0 -1,0 -0,34 1,59 72295 -0,16 0,77 -1,2
294,5 -0,5 -0,17 1,58 72353 -0,08 0,78 -0,6
295,0 0,0 0,00 1,57 72412 0,00 0,78 0,0
295,5 0,5 0,17 1,57 72471 0,08 0,79 0,6
296,0 1,0 0,34 1,56 72529 0,16 0,79 1,3
296,5 1,5 0,51 1,56 72588 0,24 0,80 1,9
297,0 2,0 0,68 1,55 72646 0,32 0,80 2,5
297,5 2,5 0,85 1,55 72705 0,40 0,81 3,2
298,0 3,0 1,02 1,54 72763 0,48 0,81 3,9
298,5 3,5 1,19 1,54 72821 0,57 0,82 4,5
299,0 4,0 1,36 1,53 72880 0,65 0,82 5,2
299,5 4,5 1,53 1,53 72938 0,73 0,83 5,9
300,0 5,0 1,69 1,52 72996 0,81 0,84 6,6
103
Outlet Temperature
Table G.8: Measurement sensitivity with varying outlet temperature Case 2
Outlet Temperature, T2
Deviation T2 Deviation, Hp Deviation, p
[%] [%] [%]
382,0 -5,0 -1,29 1,53 71922 -0,68 0,82 5,0
382,5 -4,5 -1,16 1,54 71971 -0,61 0,82 4,5
383,0 -4,0 -1,03 1,54 72020 -0,54 0,81 4,0
383,5 -3,5 -0,90 1,54 72069 -0,47 0,81 3,5
384,0 -3,0 -0,78 1,55 72118 -0,41 0,81 2,9
384,5 -2,5 -0,65 1,55 72167 -0,34 0,80 2,4
385,0 -2,0 -0,52 1,56 72216 -0,27 0,80 1,9
385,5 -1,5 -0,39 1,56 72265 -0,20 0,79 1,4
386,0 -1,0 -0,26 1,57 72314 -0,14 0,79 1,0
386,5 -0,5 -0,13 1,57 72363 -0,07 0,79 0,5
387,0 0,0 0,00 1,57 72412 0,00 0,78 0,0
387,5 0,5 0,13 1,58 72461 0,07 0,78 -0,5
388,0 1,0 0,26 1,58 72510 0,13 0,78 -0,9
388,5 1,5 0,39 1,59 72559 0,20 0,77 -1,4
389,0 2,0 0,52 1,59 72607 0,27 0,77 -1,9
389,5 2,5 0,65 1,60 72656 0,34 0,77 -2,3
390,0 3,0 0,78 1,60 72705 0,40 0,76 -2,8
390,5 3,5 0,90 1,60 72754 0,47 0,76 -3,2
391,0 4,0 1,03 1,61 72803 0,54 0,75 -3,7
391,5 4,5 1,16 1,61 72851 0,61 0,75 -4,1
392,0 5,0 1,29 1,62 72900 0,67 0,75 -4,5
104

Wet Gas Compression

Uploaded by

Copyright:

Available Formats

Wet Gas Compression

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wet Gas Compression

Uploaded by

Copyright:

Available Formats

Wet Gas Compression

Siren Carstens Amundsen

Master of Science in Energy and Environment

Norwegian University of Science and Technology

Assignment given: 09. February 2009

Trondheim, July 6th 2009

Siren Carstens Amundsen

An impeller rig for examination of single-phase and multiphase performance and

A sensitivity analysis is conducted in order to determine the effect of measurement

The efficiency and operating range of a compressor are constrained by aerodynamic

Eksperimentell forsksrigg for underskelse av ytelse og aerodynamisk stabilitet for en-

Kompressor ytelse og dets operasjonsomrde er begrenset av aerodynamisk ustabilitet i

I tilegg til stasjonre mlinger i innlps- og utlpsrr, kan mlinger av trykk

Figure 3.1: Illustration of measurement errors [11].......................................................... 16

Figure 4.1: Test compressor.............................................................................................. 19

Figure 5.1: Polytropic head curve..................................................................................... 35

Figure 7.1: Typical centrifugal compressor characteristic................................................ 53

Figure 8.1: Locations for transparent windows ................................................................ 66

Table 3.1: Stability of compressor test [9]........................................................................ 13

Table 4.1: Impeller test rig instrumentation list............................................................. 22

Table 5.1: Initial test matrix at part load........................................................................... 27

Table 6.1: Reference test parameters Case 1 ................................................................. 45

Table B.1: Discharge coefficients for orifices .................................................................. 86

Table F.1: Calculation of performance parameters at test condition ................................ 95

Gas volume fraction [-]

1 Inlet, suction state

ASME American Society of Mechanical Engineers

Test rig: the total test assembly including all instrumentation.

An experimental impeller rig for examination of single-phase and multiphase

Multiphase flows in compressors have complicated characteristics including interfacial

The effect of measurement uncertainty on performance calculations is examined to

Various measurement and visualization techniques are investigated to determine the

1.3 Limitations and Challenges

1.4 Report Structure

Chapter 2 describes the relevant performance procedures needed to document and

Chapter 7 gives the theoretical foundation for detection of aerodynamic instabilities in

Chapter 8 presents various measurement and visualization techniques employed in single-

Chapter 10 contains recommendations for further work.

An accurate method for performance calculations is important to ensure a correct

2.1 Polytropic Analysis

The thermodynamic evaluation of centrifugal compressors is generally based on the

2.2 Schultz Polytropic Analysis

2.3 Wet Gas Performance Analysis

Direct Integration Analysis

Direct integration of the polytropic process comprehends iteration from suction to

2.4 Equation of State

3.1 Test Standards

ASME PTC 10 is based on single-phase compression. Additional measurement

3.2 Test Parameters

The compressor measurement requirements are:

To evaluate thermodynamic performance, total temperature and pressure must be

p = pstatic + 0.5V 2 (3.1)

Torque, Speed and Power

3.3 Test Stability

Table 3.1: Stability of compressor test [9]

Inlet Pressure 1% of Average Value

Temperature measurements are especially sensitive to operational fluctuations during the

100 ( highest reading - lowest reading )

Table 3.2: Permissible fluctuations of test readings [11]