Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

TECH

corner
Influence Of Inlet Parameters On
Centrifugal Compressor Surge
Limit Line > Case study uses numerical analysis to
examine effects of variations
BY MASSIMILIANO DI FEBO
AND PASQUALE PAGANINI

S
urge is an instability phenomenon that consists of
a rapid oscillation of the mass flow, exacerbated by
the compressor and accompanied by characteristic
noise and high vibrations.
During the surge, the flow is suddenly reversed from the
discharge to the suction. This reverse flow through the com-
pressor causes high mechanical stresses on the machine’s
internal components. The surge is, then, an abnormal op-
erating condition that can be destructive and could cause
damage or failure when the machine remains in this condi-
tion without adequate protection.
Actual protection methods have been designed with the
intention of operating the compressor safely far from surge
points. Today the state-of-the-art anti-surge systems provide
protection consisting of opening, partially or totally, a special
control valve (anti-surge valve) located on a line that recy-
cles the gas from the discharge to the compressor suction.
In this way the control system reduces the overall line
resistance and increases the flow, moving the compressor
operating point to the right of the characteristic curve cor- ! Figure 1. This schematic shows an anti-surge system.
responding to the actual operating speed. Figure 1 shows a
typical plant layout showing a recirculation line that includes The anti-surge valve is commanded by a dedicated pro-
a recycle control valve (anti-surge valve). portional integral derivative (PID) controller, which is usually
embedded in the protection system PLC. The surge protec-
Massimiliano Di Febo is operations manager for Industrial tion logic embeds the surge limit line (SLL) and the corre-
Plant Consultants S.r.l. (IPC). He holds a master’s degree lated surge control line (SCL). Figure 2 shows an example
in mechanical engineering. He has more than 10 years of of an SLL and the SCL that usually is calculated with a 10%
professional background, with significant experience in flow safety margin from the SLL.
centrifugal pump testing, GE Nuovo Pignone specifications, Using field readings, the protection logic calculates the
applications, installations, diagnostics and maintenance. actual operative compression ratio, B, and uses the stored
Pasquale Paganini is technical manager for IPC. He holds a SCL data to determine the corresponding actual flow limit
master’s degree in mechanical engineering and is a spe- parameter. This value is then used as set point for the
cialist on thermodynamic and energy systems. He has more anti-surge valve PID controller. The process variable in-
than 10 years experience in the automotive industry (test- stead is the actual flow parameter pp also obtained from
1

ing, instrumentation, manufacturing and quality control) and the field readings.
in IPC applications. continued on page 62

JUNE 2013 60 COMPRESSORtech2


conditions (hence it’s sometimes called a universal surge
line) and being simple, it presents the possibility of easy
implemention on process computers.
This appeared as a great advantage in the early stages
of anti-surge applications (in the 1960s), when engineers
were looking for a very simple surge locus formulation to
be implemented with a very modest calculation capabil-
ity available at that time. For this reason, it has been used
largely in the past and it is still used today on actual surge
protection systems. But it also has some limitations. In fact,
recalling the base hypothesis it is possible to argue that:
1. The Hp – Q flow curve depends upon the gas suction
!"Figure 2. This is an example of an SLL and the SCL that usually condition.
is calculated with a 10% flow safety margin from the SLL. 2. The affinity law is valid for liquid flow, and can be con-
sidered applicable to gas for low compressibility flows
The preceding description illustrates how actual surge i.e., low Mach flows.
protection systems rely completely on the concept of the 3. Also, when applicable, the affinity law is valid in a small
SLL. This is basically the central element of the actual sys- range around a reference point and cannot be used
tem protection action. At this point, it is important to recall to describe the overall range of the compressor surge
that the SLL concept is based on some main hypothesis. point without violating the considered hypothesis ap-
In fact the SLL is a simplified correlation between the com- plicability limits.
pression ratio, B, and the flow parameter pp that identifies the
1
4. Also equation 5 is valid for low, B, even if it introduces
surge points. This simplified correlation is derived applying: conservative errors.
a) the hydraulic affinity laws (fan laws). These considerations show how, finally, the SSL depends
(equation 1) on the inlet suction conditions; even this is much easier to
see only in compressible flow conditions. Point one and the
H p = K1Q 2
affinity laws appear to be the main hypothesis. These can
b) the expression of polytrophic head for perfect gas: be considered applicable for single-stage compressors op-
(equation 2) erating at low Mach numbers.
n
n 1 For multistage compressors and for compressors run-
n
Hp = z1 RT1 1 ning at higher Mach numbers, a surge locus different from
n 1
the one obtained extending the fan law validity on all surge
c) the expression of the volumetric flow rate through the points should be considered. Also, the real surge locus will
suction flow meter: depend on the real behavior of the gas in the actual gas in-
(equation 3) let condition, i.e., inlet pressure, inlet temperature and inlet
gas mix composition.
p
Q = K0
1 Numerical investigation
d) the real gas state equilibrium The purpose of the investigation is to check the effect of
(equation 4) variation of inlet parameters on the surge line and in partic-
p1 ular check if the surge line remains constant or if it changes.
1 =
z1 RT1 Input data for the analysis are the compressor performance
Using the preceding hypothesis and introducing a simpli- map in design conditions and the relevant inlet gas con-
fying approximation (conservative for protection purposes) dition (gas composition, inlet pressure and temperature).
(equation 5) Based on these inputs the calculation method proceeds
n 1 with the following main steps:
n n
1 1 1. Availability of compressor design condition perfor-
n 1 mance map.
p
It’s possible to obtain the SLL final expression as correla- 2. Derivation of the surge line B – p .
1

tion between the flow parameter pp and the compression


1
3. Calculation of compressor off-design performance map.
ratio, B 4. Analytical determination of the surge points for each
(equation 6) operating speed.
5. Derivation of the numerically evaluated surge line.
p
=K ( 1) 6. Check between the surge line initial (point 2) and nu-
p1 merically evaluated (point 5)
This expression of the SSL seems to vary with gas inlet continued on page 64

JUNE 2013 62 COMPRESSORtech2


Step 1, the starting point, is the availability of the com- ferent inlet pressures, temperature and gas composition.
pressor performance map at design condition. From this The following cases have been than studied:
map the surge locus is derived. The equation to calculate C1 — DC (Case 1, design condition)
the surge points is: C1 — ODC (Case 1, off-design condition)
Pd (1) C2 — DC (Case 2, design condition)
PS C2 — ODC (Case 2, off-design condition)
=0
V For both the two cases, this study will show how the limit
Where: Pd= Discharge pressure; Ps= Suction pressure; and, surge line changes when inlet conditions move from design
v̇ = Volume flow values to off-design values.
Step 2, with the off-design inlet conditions, the new com-
pressor performance map is calculated. This step is developed Compressor with low-pressure suction
using the Cmap software. Cmap integrates aeromechanical In this case study, a centrifugal compressor is running
and thermodynamic calculations that are useful to predict the with different inlet conditions. Starting from the DC condi-
compressor performance in off-design conditions. The soft- tion, the surge line has been calculated for changes in mo-
ware algorithm is characterized by the capability to consider lecular weight (ODC 1 and ODC 2), pressure (ODC 3), tem-
the compressed gas mix as real gas. All mixture properties perature (ODC 4) and both properties (ODC 5), as shown
are then derived using the equation of state. In this analysis in the following table.
the Lee-Kesler equation of state has been used. Case 1: Low-Pressure Suction Condition
Among other very important considerations is the cal-
Operative
culation for each performance point of the gas mixture, Conditions p1 [bar a] T1 [°C] M [g/mole] Gas Mix (*)
compressibility and real gas polytrophic exponents. The
availability of these accurate thermodynamic properties as DC 67.5 38.5 16.4 Mix Design
functions of the gas composition and inlet and outlet pres-
sure and temperatures allows the software to run the aero- ODC 1 67.5 38.5 19.24 Mix 1
mechanical routines that provide final results for compres-
sor performances with higher precision. ODC 2 67.5 38.5 23.79 Mix 2
These calculations are not based on affinity law or other
approximations. They are based on the availability of a non- ODC 3 83.0 38.5 16.4 Mix Design
dimensional model of the compressor based on correlations
of work coefficient, flow coefficient parameterized by the ODC 4 67.5 23.0 16 .04 Mix Design
Mach number. Cmap software produced output performance ODC 5 83.0 23.0 23.79 Mix 2
maps are then analyzed for finding the surge points with cri-
teria (1) for several operating speeds. This numerical pro- !" Table 1. (*) Gas mix has been obtained modifying the gas mix
cess enables obtaining the numerically evaluated surge line. design composition. Calculations have been developed using the
The surge line is then calculated from the compressor per- complete gas composition. Gas mixes used are available in the Ap-
formance map with reference to inlet off design conditions. pendix. P1-inlet pressure at centrifugal compressor; T1-inlet tem-
The two obtained surge lines (DC and ODC) are then perature at centrifugal compressor; M-molecular weight; G/mole-
drawn on a plan having as x-axes the ratio between $p grams per mole; DC-design condition; ODC-off-design condition.
across the orifice and the suction pressure (Ps) and as y-
axes the compression ratio (Pd/Ps). The following graph shows the surge limit lines obtained
The error between the design and off design universal with different inlet conditions.
surge lines is then calculated with reference to the com-
pressor inlet volume flow. With equal compression ratio,
$p
calculations show the percentage error between Ps in case
off-design condition with respect design condition.

Case studies
Two real cases will be presented.
Case 1 (C1) — Compressor working with low-pressure
suction condition.
Case 2 (C2) — Compressor working with high-pressure
suction condition.
In each one of these cases, running conditions have
been considered for two different compressors: DC (design
condition), ODC (off-design condition). For the off-design !"Figure 3. The surge limit lines were obtained with different inlet conditions.
condition many sub-cases have been investigated with dif- continued on page 66

JUNE 2013 64 COMPRESSORtech2


The following table compares the design conditions ver-
sus the off-design conditions.

Case 1: Low-Pressure Suction Condition

Comparison $p % $T [°C] DM % Pd error %


Ps

DC - ODC 1 0 0 +20 2.5-3.5 - 1.1

DC - ODC 2 0 0 +48 2.5-3.5 - 2.6

DC - ODC 3 23 0 0 2.5-3.5 +1.7


!" Figure 4. This graph shows the surge limit lines obtained with
DC - ODC 4 0 -15.5 0 2.5-3.5 +0.9 different inlet conditions.

DC - ODC 5 23 -15.5 +48 2.5 +4.0


Case 2: High-Pressure Suction Condition

3.5 +7.0
Comparison $p % $T [°C] DM % Pd Error %
!" Table 2. This shows a design conditions comparison with off- Ps
design conditions.
DC - ODC 1 0 0 +20 1.5 5.5
In the previous table, it can be noted that for the off-de-
sign conditions considered (obtained varying separately 2.7 6.8
the pressure, temperature or the molecular weight), the
maximum error calculated is 2.6%. If the variation on the DC - ODC 2 0 0 +48 1.5 16.2
inlet conditions is applied simultaneous on pressure, tem-
perature and mix, the maximum error calculated is 7% (with 2.7 18.6
compression ratio of 3.5).
DC - ODC 3 +6.5 0 0 1.5 1.5
Compressor with high-pressure suction
Also, in this case study, a centrifugal compressor is run- 2.7 2.1
ning with different inlet conditions. Starting from the DC
condition the surge line has been calculated changing mo- DC - ODC 4 0 -15 0 1.5 3.4
lecular weight (ODC 1 and ODC 2), pressure (ODC 3), tem-
perature (ODC 4) and both properties (ODC 5), as shown 2.7 4.1
in the following table.
DC - ODC 5 +6,5 -15 +48 1.5 27.6
Case 2: High-Pressure Suction Condition 2.7 31.2

Operative p1 [bar a] T1 [°C] M [g/mole] Gas mix (*) !" Table 4. This table compares the design conditions versus the
Conditions off-design conditions.
DC 267.5 60 16.04 Mix Design
In the previous table, note that for the off-design conditions
ODC 1 267.5 60 19.24 Mix 1 considered (obtained varying separately the pressure, tem-
perature or the molecular weight), the maximum error calcu-
ODC 2 267.5 60 23.79 Mix 2
lated is 18.6%. If the variation on the inlet conditions is applied
ODC 3 285.0 60 16.4 Mix Design simultaneously on pressure, temperature and mix, the maxi-
mum error calculated is 31.2% (with compression ratio of 2.7).
ODC 4 267.5 45 16.4 Mix Design
It is interesting to note that the correct surge flow line moves
ODC 5 285.0 45 23.79 Mix 2 to right with respect to the initial (design) surge line. This fact
implies that the control system is underestimating the surge
!"Table 3. Gas mix (*) has been obtained modifying the gas mix de- flow, and then the compressor could be potentially exposed
sign composition. Calculations have been developed using the com- to surge events.
plete gas composition. Gas mixes used are available in the Appendix. continued on page 68

JUNE 2013 66 COMPRESSORtech2


KIENE INDICATOR VALVES
FOR COMPRESSOR ANALYSIS Conclusions
The analysis developed indicates that in a low-pressure
compressor with changing inlet conditions, surge is not a
problem. However, a reduction in operating efficiency occurs.
In a high-pressure compressor, with varying the inlet
conditions, the errors calculated (more 30%) are greater
than the safety margin (usually 10%).
Analysis proposed and described in this article enables
affirming that the actual centrifugal compressors’ protection
systems based on simplified hypothesis such as the ideal
%""&> $" gas law, hydraulic affinity laws are valid only in a certain
#   !## #! !" 
range of applications.
&& ' 86'/-.8!.649-. "''$''
For a high-pressure compressor, if varying the inlet con-
'2'</2/=+7  #
ditions, the errors calculated are greater than the safety
&& 425')8 ''7',+ (''''''
margin. Consequently, the protection becomes inefficient

and potentially can cause damage to the compressor.
&& #+38#'1:+ #'!' '''!' 

&& ' 
'%' '#
The software developed by Industrial Plants Consul-
&&  '','78 45+6'8/43
tants allows the prediction of performances for a cen-
trifugal compressor under varying thermodynamic condi-
#'4)0/3-'3*1+ '''!
tions of the inlet gas. The prediction is accurate even at
!#  !##   
high pressures, where the ideal gas theory commonly
& &  
used introduces considerable errors and implements ad-
$!
vanced protection from surge, overcoming limits of the
KIENE
   
current technology.

!+1+5.43+  

'< 




2'/1/3,40/+3+*/+7+1)42$+(7/8+;;;0/+3+*/+7+1)42
References
Lee, B.I; Kesler, M.G., “A generalized thermodynamic corre-
lation based on three-parameter corresponding states,” AiChE
Journal 1975, 21 (3), 510-527.
Plocker, U.; Knapp, H.; Prausnitz, J, “Calculation of high-
pressure vapor-liquid equilibria from a corresponding-states
correlation with emphasis on asymmetric mixtures,” Ind.
Eng. Chem. Process Des. Dev. 1978, 17 (3), 324-322.
Lee, Kesler, AIChE Journal, Volume 21, No. 3.
Kouremenos & Antenopoulos, “Isentropic Exponents of
Real Gases and Application for the Air at Temperatures
From 150 K to 450 K.”
Maric, Flow Measurement and Instrumentation 16 (2005).
K.H. Lüdtke, Process Centrifugal Compressors.
Cengel, Boles, “Thermodynamics An Engineering Ap-
proach.”
Di Febo, Paganini, Pedone, Esposito, “Prediction of cen-
trifugal compressor performance and application for test,
surge protection and machinery diagnostic” (2012). CT2

Appendix
Gas composition of the gas used:

Symbol % mol
Methane 100
Total 100
M (g/mole) 16.04

! Table 5. Mix design.

JUNE 2013 68 COMPRESSORtech2


Certified Quality
Management System
ISO 9001: 2008

Symbol % mol
Precise Boring …
Nitrogen 1.46
Methane 80.51
Ethane 14.69
Propane 3.19
i-butane 0.07
n-butane 0.08
Total 100
M (g/mole) 19.24
… to exceed your
! Table 6. Mix 1.
expectations!!
Symbol % mol Visit our website at www.HahnMfg.com

Nitrogen 0.08 PRECISION BORING, MILLING, DRILLING, TAPPING


AND GRINDING ON FERROUS AND NON-FERROUS
Carbon dioxide 3.62 MATERIAL UP TO 30,000 LBS.

Methane 73.11
Ethane 10.44
Propane 6.5
i-butane 0.83
n-butane 1.82
i-pentane 0.46
n-pentane 0.5
n-Hexanes 0.74 We Manufacture and
n-Heptane 1.04 Remanufacture the
Hydrogen sulphide 0.57 World’s Largest
n-Octanes 0.25 Crankshafts
n-Nonanes 0.04 Ellwood Crankshaft Group
Irvine, PA, USA 16329
Total 100 Hermitage, PA, USA 16148
1-800-247-1326
M (g/mole) 23.79 or 724-347-0250
ecgsales@elwd.com
www.ellwoodcrankshaftgroup.com
! Table 7. Mix 2.

JUNE 2013 69 COMPRESSORtech2

You might also like