A Moore-Greitzer axial compressor model with spool dynamics.

Department of Engineering Cybernetics Norwegian University of Science and Technology O.S. Bragstads plass 8, N-7034 Trondheim, Norway Email: fTommy.Gravdahl,Olav.Egelandg@itk.ntnu.no
Abstract
The compressor stall/surge model of Moore and Greitzer 11] is extended to include the spool dynamics. This results in a model with four states, where the new state is the B-parameter. The e ect of a time varying B-parameter on the post stall transients of an axial compressor is demonstrated through simulations of a compression system with speed control. Author(s) states Greitzer 6] Hansen et.al 8] several (see 1]), incl. Gravdahl/Egeland 3] Moore/Greitzer 11] J several (see 1]), incl. J Gravdahl/Egeland 4] Fink et.al 2] B Gravdahl/Egeland 5] B Gravdahl/Egeland JB (this paper) A/C A C A/C A A C C A M/C/S M M C M C M MCS MS

Jan Tommy Gravdahl and Olav Egeland

1 Introduction
The low order model of Moore and Greitzer 11] for the post stall transients of axial compression systems has been used extensively in stall/surge analysis and control. In the original work of Moore and Greitzer, the compressor speed is assumed constant. If the equilibrium of the compression systems is located to the left of the surge line, which passes through the local maxima of the compressor characteristic, the ow becomes unstable. Dependent on certain system parameters, and as will be demonstrated, compressor speed, the instability can take the form of either rotating stall, surge or both. Surge is an axisymmetric oscillation of the ow and rotating stall is characterized by regions of reduced ow that rotates around the annulus of the compressor. These oscillations severely reduces the compressor e ciency and can possibly damage the compressor. A number of approaches to control of surge and rotating stall have been proposed. A review of the di erent approaches can be found in 1]. In 7] it was concluded that low values of Greitzer's Bparameter B leads to rotating stall, while high values of B leads to surge. As B is proportional to the angular speed of the machine, it is of major concern to stall/surge controller design to include the spool dynamics in a stall/surge-model. A model for centrifugal compressors with nonconstant speed was presented in 2]. In 5] a similar model was derived, and surge and speed control was investigated. However, as the models of 2] and 5] are developed for centrifugal machines, they do not include the equations for rotating stall. In table 1, the development in stall/surge modeling and control is outlined. In 8] it was demonstrated that the model of 6] also applies to centrifugal compressors. It seems that the modeling and control of an axial compression system including both rotating stall and spool speed is an open problem. The problem is also listed among topics for further research in 7]. In this paper we propose an extension to the Moore/Greitzer model that includes the B-parameter as a state.

Table 1: Development in compressor stall/surgeLI pT C

01

control. A=Axial, C=Centrifugal, M=Modeling, C=stall/surge control, S=Speed-control

E Lc

LE

Station numbers

HH H
x

ps Vp

A A pT

Inlet duct

Exit duct IGV Compressor Plenum Throttle

Figure 1: Compression system 2 Preliminaries

The compression system consists of an inlet duct, inlet guide vanes (IGV), variable speed axial compressor, exit duct, plenum volume and a throttle. The throttle can be regarded as a simpli ed model of a turbine. The system is shown in gure 1. Our aim is to develop a model for this system in the form _ = f (z ) z (1) T where z = ( J B ) and is the circumferentially averaged ow coe cient, is the total-to-static pressure rise coe cient, J is the squared amplitude of angular variation (rotating stall) and B is Greitzer's B-parameter, which is proportional to the speed of the compressor. The modeling of the compression system relies heavily on the modeling in 11], however the assumption of constant angular speed !, and thus constant B , is relaxed and a momentum balance for the

spool is included. In 11] nondimensional time was dened as Ut (2) MG = R where U is the spool speed at mean radius, R is the mean compressor radius and t is time. As we in this paper consider time varying U , this normalization will not be used. Instead we propose to use dt =U (3) R where Ud is the desired constant velocity of the wheel. Note that if Ud = U = const, we have that = MG . All distances are nondimensionalized wrt R, that is the nondimensional duct lengths, are de ned as LE : I lI = L and l (4) E= R R The axial coordinate is denoted and the circumferential coordinate is the wheel angle . Greitzer's Bparameter is de ned as r Vp 4 U (5) B= 2as Ac Lc where as is the speed of sound, Vp is the plenum volume, Ac is the compressor duct ow area and Lc is the total length of compressor and ducts. Bs and U is related as 4 2a Ac Lc : U = bB where b = (6)
s

equation (12) can be written

c

tip = mc R U: (14) For simplicity, the inlet tangential velocity of the gas has been ignored in c . The compressor mass ow is given by mc = Ac U : (15) Combining (12) and (15) gives tip 2 = Ac R U (16) and by (9) the nondimensional compressor torque is c

R2

R2

;c =

Rtip R

(17)

3.2 Compressor
where
2

The pressure rise over a single blade row is 10] pE ; p1 = F ( ) ; d 1 2 dt

(18)

Vp

The momentum balance of the spool can be written (7) I d! dt = t ; c where ! is the angular speed, I is the spool moment of inertia, and t and c are the drive (turbine) torque and compressor torque respectively. Using ! = 2U=R, (3) and (5), the spool sdynamics (7) can be written 4as IUd Ac Lc dB = ; : (8) As in 2], torques are nondimensionalized according to t; c (9) ; = ;t ; ;c = A c RU 2 where is the constant inlet density. Now, equation (8) can be written dB = B 2 (; ; ; ) (10) 1 t c d where the constant 1 is de ned as 4 R3Ac (11) 1 = 2IU b: d As the compressor torque equals the change of angular momentum of the uid 9], the compressor torque can be written (12) c = mc Rtip Ctip where Rtip is the radius of the rotor and Ctip is the tangential velocity of the uid as it leaves the rotor. Using the slip factor 4 Ctip (13) = U
tip

3 Modeling 3.1 Spool dynamics

x =C (19) U is the local axial ow coe cient, F ( ) is the pressure rise coe cient in the blade passage and is a coe cient of pressure rise lag. According to 11], d dt can be calculated as d = @ + @ : (20) dt @t @t

Using (3) it is seen that

rotor

stator

@ @t

rotor

@ @ @ = @ @ @t + @ @t = Ud @ + @ U (t) R@ U = Rd @ @ @ R

(21) (22)

R2

Vp d

where the unsteadiness of the ow through the stator passage re ects the accelerations associated with transients e ects. For the rotor there is also unsteadiness due to the rotor blades moving with velocity U (t) through a circumferential nonuniform ow. Considering a compressor of N stages, we get pE ; p1 = NF ( ) ; 1 2 @ + U @ (23) 1 2 2a @ U @ where
2

@ @t

stator

4 R a= (24) N Ud is a constant. Note that if U ( ) = const and Ud U

such that = MG , equation (23) is reduced to equation (5) in 11]. It is noted that the ow coe cient can depend on both and , even though the atmospheric stagnation pressure pT is constant. The average of around the wheel is de ned as Z 1 2 ( )d = 4 ( ): (25) 2 0 Further = ( ) + g( ) and h = h( ) (26) where h is a circumferential coe cient.

The fact that the rotational speed of the wheel now is assumed time varying does not change the conditions upstream of the compressor. Therefore the equations stated in 11] are still valid, and will be presented here. The pressure di erence over the IGVs, where the ow is axial, can be written p1 ; p0 = 1 K h2 (27) U2 2 G where 0 < KG 1 is the entrance recovery coe cient. If the IGVs are lossless KG = 1. Upstream of the IGV irrotational ow is assumed so that a (unsteady) velocity potential ~ exists. The gradient of ~ gives axial and circumferential velocity coe cients everywhere in the entrance duct. At the IGV entrance point (denoted by subscript '0') we have ( ~ )0 = ( ) + g( ) and ( ~ )0 = h( ) (28) where partial derivation wrt and is denoted by subscripts. Applying Bernoulli's equation we get pT ; p0 = 1 ( 2 + h2 ) + ( ~ ) (29) 0 U2 2 where the term ( ~ )0 is due to unsteadiness in and g. A straight inlet duct of nondimensional length lI is considered, and the velocity potential can be written ~ = ( + lI ) ( ) + ~0 ( ) (30) 0 ~ where is a disturbance velocity potential such that ~0 = 0 ( ~0 )0 = g( ) and ( ~0 )0 = h( ): =;lI (31) Equation (29) can now be written pT ; p0 = 1 ( 2 + h2 ) + l d + ( ~0 ) : (32) I 0 U2 2 d Downstream of the compressor the ow is complicated and rotational. As in 11], the pressure p in the exit duct is assumed to di er only slightly from the static plenum pressure ps ( ), such that the pressure coe cient P satis es Laplace's equation. 4 ps ( ) ; p r2 P = 0: P= (33) U2 The axial Euler equation 12] is used to nd the pressure drop across the exit duct. Generally we have in the xcoordinate dp = dCx (34) ; dx dt where Cx is the velocity component along the x-axis, the axial ow velocity. Employing the chosen nondimensionalization of time and distance, d = Ud d and d = 1 d (35) dt R d dx R d we get dpE = ; U 2 1 (P ) (36) dx R E (37) Inserting (36) and (37) in (34) we get the following expression for axial Euler equation, evaluated at E , where time varying U has been taken into account Ud d n( ~ ) U o (P )E = U 0 2d

3.3 Entrance duct and guide vanes

Ud ( ~ ) U + ( ~ ) dU : (38) = U 0 0d 2 From (28) we have ( ~ )0 = ( ) + ( ~0 )0 : (39) Inserting (28) and (39) into (38) we get dU h ( ) + ( ~0 ) i : d d ~0 )0 + Ud (P )E = U + ( 0 2 U d U d (40) We want that P = 0 for at the duct exit. Thus, d ; ~0 d P = UU ( ; l E) () d
Finally we get ps ; pE = (P ) = Ud ;l d ; (m ; 1)( ~0 ) E 0 U2 U( ) E d h i Ud dU ;l ( ) ; (m ; 1)( ~0 ) (42) +U E 0 2 d where, as in 10] and 11], the compressor duct ow parameter m has been included. It is noted that if U ( ) =const and Ud U such that = MG , equation (42) is reduced to equation (20) in 11]. Using the preceding calculations, we now want to calculate the net pressure rise from the upstream reservoir total pressure pT to the plenum static pressure ps at the discharge of the exit duct. This is done by combining equations (23), (27), (32) and (42) according to ps ; pT = (NF ( ) ; 1 2 ) ; (l + l Ud + 1 ) d I E U2 2 U a d 1 d 2 ~0 + (1 ; m) U U ; 1 ( )0 ; 2 (1 ; KG)h Ud dU ;l ( ) ; (m ; 1)( ~0 ) +U E 0 2 d U ( ~0 ) ; 21a 2( ~0 )0 + U (43) 0 d where @ @ @g @g 2@ @ + @ = 2 @ + 2@ + @ ~0 ~0 = 2@ @ + 2( )0 + ( )0 (44) has been used. By de ning 1 pT 1 2 ( ) = ps ; c ( ) = NF ( ) ; U2 2 (45) l (U ) = l + l Ud + 1 m (U ) = (1 ; m) Ud ; 1 and assuming KG 1, (43) can be written ~0 ( ) = c ( ) ; lc (U ) d d + mU (U )( )0
c I E

Ud dU ( ; l ) ( ) ; ~0 : +U E 2 d

(41)

3.5 Overall pressure balance

3.4 Exit ducts and guide vanes

dCx = Ud d n( ~ ) U o : 0 dt Rd

1 Notice that l = L R. As in 2] L used in the de nition of c c c B is a constant. See equation (5).

6

Ud dU ;l ( ) ; (m ; 1)( ~0 ) +U E 0 2 d U ( ~0 ) ; 21a 2( ~0 ) + U (46) d 0

which, with the usual assumptions, reduces to equation (26) in 11]. Equation (46) requires knowledge of the disturbance velocity potential ~0 and its derivatives. As ~0 satis es Laplace's equation,r2 ~0 = 0, it has a Fourier series. As an approximation only the rst term of this series is used, which means ( ~0 )0 = ;( ~0 )0 : (47) Switching notation to 4 ( ~0 ) ) ( ~0 ) = ;Y Y( )= (48) 0 0 equation (46) is written () =
c(

3.7 Characteristics

Ud dU (;l ( ) ; (m ; 1)Y ( )) +U E 2 d UY : ; 21a 2Y + U (49) d We now want to express equation (49) in terms of varying B instead of varying U . Using (5) it is seen that
( ) = c ( ; Y ) ; lc (B ) d d + mB (B )Y 1 + Ud ; b (;lE ( ) ; (m ; 1)Y ( )) ; 21a 2Y + bB (50) Ud Y :

; Y ) ; lc (U ) d d + mU (U )Y

The compression systems characteristics are taken from 11]. The usual 3rd order polynomial steady state compressor characteristic is employed: 1 3 3 (57) c ( ) = c0 + H 1+ ( 2 W ; 1) ; 2 ( W ; 1) where the parameters c0 , H and W are de ned in 11]. The compressor characteristic can also be plotted as a family of curves, one line for each rotational speed, with the surge line passing through the local maxima of the constant speed lines. However, when using the nondimensionalizing employed here , these lines collapse into a single curve c ( ) 2], and the surge line collapses into the local maximum c ( ). When the compressor is in rotating stall the characteristic is given by 3 5 3 s ( ) = c0 + H 1 ; ( 2 W ; 1) + 2 ( W ; 1) : (58) The throttle characteristics is ptaken to be : (59) T =

3.8 Galerkin procedure

Integrating (50) over one cycle wrt we get Ud; 1 ( ) = 1 Z 2 ( ;Y )d : ( )+lc(B ) d + l d E b 2 0 c (51) The mass balance in the plenum can be written d ( V )=m ;m (52) c t dt p p where p is the plenum density, mc is the mass ow entering the plenum from the compressor, and mt is the mass ow leaving through the throttle. Assuming the pressure variations in the plenum isentropic, we have By nondimensionalizing pressure with U 2 , mass ow with UAc , transforming to nondimensional time , and taking account for (5), (53) can be written d = 2 ( ; ) ; 2 dB (54) where

As previously assumed Y is approximated with the rst term in a Fourier series, that is Y is represented by a single harmonic function Y of unknown time varying amplitude A( ): Y = WA( ) sin( ; r( )) (60) where r( ) is an unknown phase angle. From (50), we get Ud ; 1 ( ) Y = m 1(B ) ( ) + lc(B ) d + l E d b B U d; 1 ; ( ;Y )+ (m ; 1)Y ( ) (61) ; 21a 2Y + bB Ud Y De ning = ; r( ), we get from (60) dr : Y = W dA sin ; WA ( ) cos (62) d d The model developed thus far includes partial derivatives of Y . A Galerkin approximation is to be used to produce a set of ordinary di erential equations. A residue R, is de ned as 4 Y ;Y : R= (63) The Galerkin approximation is calculated using the weight functions h1 = 1 h1 = sin h2 = cos (64) and the inner product Z2 hR hi i = 21 R( )hi ( )d : (65) 0 Calculating hR hi i = 0 for i = 1 2 3 and using (57) give after some lengthy algebra 1 Z 2 ( ) d = ( ) + l ( B ) d + l Ud ; 1 ( ) c E 2 0 c d b Z 2 Ud; 1 (m ; 1) A( ) + dA m0 1 c ( ) sin d = B
c

3.6 Plenum mass balance

dpp = a2 s dt Vp (mc ; mt ):

(53)

Bd

2=

Using (10) we get

R LcUd b:
T)

(55)

(56) The model of the compression system now consists of the torque balance for the spool (10), the local momentum balance (50), the annulus averaged momentum balance (51) and the mass balance of the plenum (56).

d = 2( ; d B

; 2 1;B :

W 0 b d Z 2 1 dr 0 bB )A( ) (66) c ( ) cos d =;( mB ; W 0 d 2aUd 1 ; mB ) and = + WA( ) sin . Exwhere m0B = ( a
amining the last equation in (66), it is recognized that

1 b dr = 2 Ud (67) d 1 ; mB (B )a B Notice that time varying B implicates that the phase angle r is not a constant, which was the case in 11].

the integral vanishes, and assuming A 6= 0 the phase angle r satis es

theorem. Work on this subject for the model (68)-(71) is underway. The desired speed was set to Ud = 215m/s in both the following simulations. Numerical values for the parameters in the model are given in appendix A.

4.1 Unstable equilibrium, = 0:5

3.9 Final model

By evaluating the integrals in (66), using (10), (54) and rearranging, we get the following model for the compression system d = H ; ; c0 + 1 + 3 ( ; 1)(1 ; J ) d l (B ) H 2 W 2

d = 2 ( ; ) ; 2 ;B (69) T 1 d B dJ = J 1 ; ( ; 1)2 ; J d W 4 2 U ; ( m ; 1) W 3aH 1 ; d 3 (70) bH (1 ; mB a)W dB = ;B 2 = (u ; )B 2 (71) 1 1 c d where J is de ned as the square of the stall amplitude 4 A2 ( ) J( ) = (72) and u = t is the speed control. The model (68) is in the desired form of (1). Notice that the model is highly coupled, so that an acceleration of U should have
impact on all states. It should be noted that the model developed in this paper can be reduced to other well known compression system models. This is summed up in table 2. Changes made Rede ne Gives the to (68)-(71) nondim. model of time acc. to J 0, dB := Ut Greitzer 6] d =0 R U = Ud , ; = 0 dB = 0, U = U := Ut Moore/Greitzer d d R ;=0 11] J 0 := t!H Fink et.al. 2]

3 lE Ud ; ;1 2 ( W ; 1) ; bH

(68)

In gure 2, a simulation of the system is shown. The throttle gain was set at = 0:5 so that the equilibrium is to the left of the local maximum of the characteristic. As can be seen from gure 2, the compressor goes into rotating stall as B (and thus compressor speed U ) is low. Moreover, when the applied torque from the speed controller cause B to increase, the stall amplitude J falls o and the compressor goes into surge. This is what could be expected according to 7]. The surge oscillations have a period of 180, which correspond to a surge frequency of about 10Hz. A desired speed of Ud = 215 corresponds to a desired B-parameter of Bd = Ud=b = 2:23. After 1500 this value is reached. As the compressor torque ;c varies with , see equation (17), we would expect oscillations in speed U as the compressor is in surge. This is con rmed by the lower right plot in gure 2. In the upper plot of gure 3, the trajectory starts on the stable part of the characteristic, then rotating stall occurs and the trajectory approaches the intersection of the throttle and in-stall characteristics. As B increases the resulting surge oscillations are clearly visible.
0.8 0.6 0.4 0.2 0 0.2 1.4 1.2 1 0.8 0.6 0.4 0.2 0 500 1000 1500 2000 0 0 500 1000 1500 2000

3 0 0 500 1000 1500 2000

500

1000

1500

2000

B
1 215 214 213 1500 1600 1700 1800 1900 2000

Table 2: Relation with other models. (!H is the Helmholtz

frequency)

Figure 2: Simulation of the system (68)-(71). Low B leads

to rotating stall, and high B leads to surge.

4.2 Stable equilibrium = 0:65 4 Simulations

Here some simulations of the model will be presented. For speed control, a simple P-type controller of the form ;t = c(Ud ; U ) (73) will be used. The nondimensional drive torque ;t is used as the control, and feedback from compressor speed U is assumed. Initial values were chosen as ( 0 0 J0 B0 ) = (0:55 0:65 0:05 0:1) (74) such that the ( )-trajectory starts on the stable part of the compressor characteristic as can be seen in gure 2. In 5] compressor speed was controlled in a similar manner, and stability was proven using Lyapunovs

In gure 4, the impact of the spool dynamics on a stable equilibrium of the compression system is illustrated. Initial values were unchanged. The throttle gain was set at = 0:65, giving a stable equilibrium, and the speed controller gain was chosen as c = 1. The dotted trajectories show the system response to a speed change from U = 0:05 to U = 215. It can be seen that the acceleration of U a ects the other states of the model. This is due to the couplings with speed U and torque ; in the model. Of special interest is the stall amplitude. The initial value of J (0) = 0:05 grows to nearly fully developed rotating stall as the machine is accelerating, but is quickly damped out as desired speed is reached. The response is also plotted in the lower plot of gure 3. Simulations show that this stalling can be avoided by accelerating the compressor at a lower rate, that is by

1 0.8 0.6 0.4 0.2 0 0.2

c
0.2 0.3 0.4 0.5 0.6 0.7 0.8

0.1

0.1

Compressor speed was controlled with a simple proportional control law. Further work includes 1) Stability analysis and stall/surge control design for variable speed compressors, 2) The use of both speed control and stall/surge control to achieve rapid acceleration without stalling, and 3) Inclusion of higher harmonics in rotating stall, that is include more terms in the Fourier series of ~0 . This will be the subject of future publications.

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.2 0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

A Numerical values used in simulations

s c

Figure 3: Simulations result superimposed on the com-

pression system characteristics. The compressor characteristic, the in-stall characteristic and the throttle characteristic are drawn with solid, dashed and dash-dot lines respectively.

Sym. value R 0:1m lE 8 Vp 1:5m3 H 0:18 I 0:03kgm2

Sym. value kg 1:15 m 3 lI 2 Ac 0:01m2 W 0:25 m 1:75

Sym. value as 340 m s Lc 3m a 0:3 0:3 c0 0.9

Table 3: Compression system parameters used in simulations

1] B. de Jager. Rotating stall and surge control : A survey. In Proceedings of the 35th Conference on Decision and Control, pages 1857{1862, New Orleans, LA, 1995. 2] D.A. Fink, N.A. Cumpsty, and E.M. Greitzer. Surge dynamics in a free-spool centrifugal compressor system. Journal of Turbomachinery, 114:321{332, 1992. 3] J.T. Gravdahl and O. Egeland. Compressor surge control using a close-coupled valve and backstepping. In Proceedings of the 1997 American Control Conference, Albuquerque, NM, June 1997. 4] J.T. Gravdahl and O. Egeland. Control of the three-state Moore-Greitzer compressor model using a close-coupled valve. In Proceedings of the 1997 European Control Conference, July 1997. 5] J.T. Gravdahl and O. Egeland. Speed and surge control for a low order centrifugal compressor model. Accepted as a regular paper to 1997 IEEE Conference on Control Applications, October 1997. 6] E.M. Greitzer. Surge and Rotating stall in axial ow compressors, Part I: Theoretical compression system model. Journal of Engineering for Power, 98:190{ 198, 1976. 7] E.M. Greitzer and F.K. Moore. A theory of poststall transients in a axial compressor systems: Part II{ Application. Journal of Engineering for Gas Turbines and Power, 108:231{239, 1986. 8] K.E. Hansen, P. J rgensen, and P.S. Larsen. Experimental and theoretical study of surge in a small centrifugal compressor. Journal of Fluids Engineering, 103:391{394, 1981. 9] J.D. Mattingly. Elements of gas turbine propulsion. McGraw-Hill series in aeronautical and aerospace engineering. McGraw-Hill, 1996. 10] F.K. Moore. A theory of rotating stall of multistage axial compressors: Parts I,II,III. J. of Engineering for Gas Turbines and Power, 106:313{336, 1984. 11] F.K. Moore and E.M. Greitzer. A theory of poststall transients in a axial compressor systems: Part I{ Development of equations. Journal of Engineering for Gas Turbines and Power, 108:68{76, 1986. 12] F.M. White. Fluid mechanics. McGraw-Hill, New York, 2nd edition, 1986.

References

using a smaller c. In contrast, the solid trajectories in gure 4 show the response without the spool dynamics. Now, the initial value J (0) = 0:05 is damped out very quickly, and and converges to their equilibrium values. The only transient e ects are due to the initial conditions.
0.7 0.65 0.6 0.55 0.5 0.45 0.4 0.35 0 200 400 600 800 1000 0.3 0.2 0.5 0.4 0.7 0.6

200

400

600

800

1000

3 2.5 2

2.5 2 1.5

1.5 1 0.5 0 0 200 400 600 800 1000

B
1 0.5 0 0 200 400

600

800

1000

Figure 4: Stable equilibrium with and without Bdynamics

5 Conclusion and further work

A Moore-Greitzer axial compressor model with spool dynamics has been derived. This results in a model with time varying B -parameter. Through simulations it has been demonstrated that the model is capable of demonstrating both rotating stall and surge, and that the type of instability depends on the compressor speed.