Chapter Axial Compressor Stages Aw axial compressor" js a pressure producing machine. The energy level of air or gas flowing through it is increased by the action of the rotor blades which exert a torque on the fluid. This torque is supplied by an external source—an electric motor or a steam or gas turbine. Besides numerous industrial applications the multistage axial compressor is the principal element of all gas turbine power plants (see Chapter 3 and 5) for land and aeronautical applications, An axial compressor stage was defined in Sec. 1.7 and its merits discussed in Sec. 1.9. In contrast to the axial turbine stages, an axial compressor stage is a relatively low temperature and:yx-ssure machine with no serious material problems. While impulse type of stages are commonly used in turbines, these are rare in compressors except in supersonic stages. The flow in a compressor stage is throughout decelerating as in a diffuser. In this respect the nature of flow in a compressor stage is largely identical to the diffuser flow which has been discussed in Sec. 2.4. As Pointed out earlier, the blade profiles in compressor blade rows are more critical than in turbine stages. This is on account of the flow occurring against the pressure rise and the susceptibility of the flow to separation, The arrangement of blade rows (cascades) in compressors and the nature of flow through them have been discussed at length in Sec, 8.5. Figures 8.26 to 8.36 dipict various geometrical and aerodynamic aspects of compressor blade rows and the flow through them. This chapter deals with the acrothermodynamic analysis of flow through complete compressor stages. Methods of estimating the stage Work and efficiency are given here (see Sec. 2.6). Flow problems through both long (low pressure stages) and short blades (high pressure) have been considered. Various problems of low pressure axial compressor stages are identical {o those of axial-flow fans which have been dealt with in Chapter 14 Slight amount of overlap between these chapters has been considered inevitable and tolerated. However, some aspects have been covered only ‘none chapter. Some striking differences that would be observed between the two chapters are the use of “multi-stages” and the term “pressure [Link]Axial Compressor Stages 457 ratio” in this chapter and “single stage fans” and “pressure rise in millimeters of water gauge” in Chapter 14. > 11.1 Stage Velocity Triangles Pressure and velocity variations through a compressor stage (with inlet guide vanes) are shown in figure 11.1(a). IGVs Rotor Diffuser a | | |_| Pressure Velocity Fig. 14.1 (a) _ Pressure and velocity variation through a compressor stage . Velocity triangles for axial compressor stages have been briefly dis- cussed in Chapters | and 8. The velocity triangles shown in Fig. 11.1(b) are for a general stage which receives air or gas with an absolute velocity ¢, and angle @ (from the axial direction) from the previous stage. In the case of the first stage in a multi-stage machine, the axial direction of the [Link]Compressors and Fans Upstream guide vanes Py Py Fon Entry velocity triangle ", Fig. 11.1 (b) Velocity triangles for a compressor stage approaching flow is changed to the desired direction (a,) by providing a ow of blades upstream guide vanes (UGV). Therefore, the first stage experiences additional losses arising from flow through the UGVs. For a general stage, the entry to the rotor, exit from the rotor and the diffuser blade row (stator) are designated as stations 1, 2 and 3, Tespectively. The air angles in the absolute and the relative systems are denoted by a, o, a and B,, B, respectively. If the flow is repeated in another stage. y= cy and a = a Subscripts x andy denote axial and tangential directions respectively. Thus the absolnte switl or whirl vectors 1 and cyy ate the tangential [Link]Axial Compressor Stages 459 components of absolute velocities cy and c,, respectively. Similarly, w. and wp are the tangential components of the relative velocities w, and ws, respectively. Peripheral velocity at the mean diameter of the rotor at stations 1 and 2 is taken as w= uy =i ‘The following trignometrical relations obtained from velocity triangles (Fig, 11.1b) will be used throughout this chapter. From velocity triangles at the entry: Cy = C1 COS Oh = W, Cos By Gua ey = 6; Sin 0 = Cy tam Of (11.2) wey =, sin By = 6 tan By (11.3) w= oy tM (11.4a) w=; sin oy + w, sin By (11.4b) w= cy (tan a, + tan B,) (11.4c) From velocity triangles at the exit: Cy = C2 COS = W COS By (5) Cy = 2 SiN Oy = Cy tan (11.6) Why = Wp Sin By = Cxp tan By ay uot Wye (11.8a) = cz sin a + w, sin By (11.86) w= Cg (tan op + tan fy) (11.80) For constant axial velocity through the stage: Cx = Gn = Oa = x ans) 3 C08 f, = Ww; c08 By (11.10) Equations (11.40) and (11.8c) give e, = €4 608 % = w; cos By = 1 Sian oy, + tan B, = tan @ + tan B, GLb ce @ This relation can also be presented in another form using Eqs. (11.4a) and (11.8a), Ct Wy = Gat Wye Ga- G1 = > Me (11.12a) oc, (tan of — tan ot) = ¢, (tan By ~ tan B) (11.126) Equations 11.12 (a and b) give the change in the swirl components across the rotor blade row. For steady flow in an axial machine, this is [Link]460 Turbines, Compressors and Fans Proportional to the torque exerted on the fluid by the rotor (Sec. 6.9, Eq. (6.145b)). 14.1.4 Work The specific work in a compressor stage is given by Eq. (6.1476). In the present notation it is w=uea- 6) (11.13a) Using Eq. (11.12b) it can also be expressed as wwe, (tan a — tan a) = we, (tan B,~tan B,) (11 13b) If the alphas and betas are actual air angles, Eq. (11.13b) gives the actual value of the stage work. The difference between the actual, isentropic and Euler’s work has already been explained in Sec. 6.9. For axial-flow compressor (v= u, = u,), the specific work equation identical to Eq. (6.153a) is 1 2 1 Wa 5 (-G)+ 5 wh wa) (11.130) For a desired pressure rise in a compressor, the work input should be minimized to obtain higher efficiencies. In this respect the selection of the optimum blade and flow geometries (Sec. 8.5.7) is important, 11.1.2 Blade Loading and Flow Coefficients The blade loading coefficient for an axial compressor stage is defined as (11.14) This is a dimensionless quantity used for comparing stages of differing sizes and speeds. Head, pressure or loading coefficients have been discused in See. 7.4.1. For fan applications, y is defined as a pressure coefficient in Eq, (14.7) In Sec. 7.4.2 another dimensionless coefficient known as the capacity Coefficient is defined. An expression for the flow coefficient is derived from this Eq. (7.17): = _ (11.13) Equations (11.13a) and (11.13b), when pat in Eq. (11.14), give success: ively = fa _ on ¥ u u Y= 9 (tan oy — tan a) = 9 (tan B, tan B,) (11.16) [Link]__ Axial Compresior Stages 461 The performance of axial compressor stages is presented in terms of — wplots. Figure 7.6 depicts the variation of pressure coefficient (loading coefficient) with the flow coefficient for an axial compressor stage. 11.1.3 Static Pressure Rise ‘The main function of a compressor is to raise the static pressure of air or gas, The static pressure rise in the stage depends on the flow geometry and the speed of the rotor. The total static pressure rise across the stage is the sum of static pressure rises in the rotor and diffuser (stator) blade rows; expressions for these values are derived here assuming reversible adiabatic flow and constant axial velocity through the stage. Further, in view of the small pressure rises over blade rows of axial compressor stages the flow is assumed incompressible, i.e. p ~ constant. ‘The Bernoulli equation across the rotor blade row gives 1 a Pit 5 PM = Pat > PM 1 (pe =P2- P= x POW WA) (u.l7) Using velocity triangles of Fig. 11.1b, whe wh = ch + whi — ea = Mpa = M1 w= w3 = c? (tan? B, ~ tan? Bo) ‘This when put in Eq. (11.17) gives the pressure rise across the rotor as 1 2 (Ap)a = 5 Pe; (tan? By ~ tan? B,) (11.18) Similarly, the Bernoulli equation across the diffuser blade row gives (Ap)p = P3~ P2* 3 p(ch- ec) (11.19) tno botds-29 (Pp) = 3 pc2 (tan?e, — tan’a,) Assuming = a, (Ap)p = ; pe? (tan?or, — tan?oy) (1.20) ‘The stage pressure rise is PsP) = APs = APR + APD [Link]462 Turbines, » Compressors and Fans Substituting from Eqs. (11.18) and (11.20), APy = 4 pez {(tan"B, ~ tan") + (taney - tan®ay)} Using Eq. (11.126) Apa = % pe; (tan By ~ tan B) Le, (an B+ tan By +c, (tan @ + tan @,)} This, on rearrangement and using Eqs. (11.4c) and (11.8), gives Apa, = pe, u (tan By, — tan By) (11.21) Apz = pe, u (tan 0% - tan ay) (41.22) These relations for isentropic flow can also be obtained direct from Eq, (11.13b). For such a flow, changes in pressure and enthalpy and work are related by APs = An, = w (41.23) p > 11.2. Enthalpy-Entropy Diagram Figure 11.2 shows the enthalpy-entropy diagrarp-for a ‘general axial-flow Compressor stage. Static and stagnation values of hresstees and enthalpies at various stations (shown in Fig. 11.1b) have bekn indicated. ” The stagnation state 0, at rotor entry in the absolute system is fixed by the pressure p, and velocity c,. The isentropic flow over the rotor and diffuser blades is represented by 1-2, and 2,~3,,. The stagnation point 0;,, Corresponding to the final state at the end of an isentropic compression is obtained from Mosse = The irreversible adiabatic or actual compression process is represented by curve 1-2-3, Here energy transformation processes (1-2) and (2-3) in the blade rows occur with stagnation pressure loss and increase in entropy. For the rotor blades in the relative system: Stagnation pressure loss = (AP), = Pore ~ Poort The stagnation enthalpy remains constant. Foret = haga | (11.24) Bs 1 1 11.25) tpwieh thud (11.25) [Link]_ Axial Compressor Stages 463 Pos = 03 hop hog j<— ast | $8 ps wa Enthalpy Ee Entropy Fig. 11.2 Enthalpy-entropy diagram for flow through a compres sor stage ‘The pressure (stagnation) and enthalpy loss coefficients are defined by — (Apo)R = Porret = Porret (11.26) I 1 F yew 7PM 20, (h -T = 2 GEFs) 1.27) [Link]464 Turbines, Compressors and ns For the diffuser blades (absolute system): Stagnation pressure loss = (Ap9)p = Po» — Pos The stagnation enthalpy remains constant. hah 1 (11.28) Int Gahyt z gS . The pressure and enthalpy loss coefficients are defined by p= GMa. Pa= Pa cies ze 3 Pee 2 ‘T, — T, f= tote . 2¢pG-Ty) (11.30) Ta a td 3 11.2.1 Efficiencies The efficiency of the compression process can now be defined on the basis of ideal (isentropic) and actual (adiabatic) processes defined in the previous section. The ideal work in the stage is Ws = hoses ~ hin = p (Toszs — Tor) (1131) This is the minimum value of the stage work required to obtain a static Pressure rise of (AP). = Ps — Py However, the actual process, on account of losses and the associated inreversibilities, will require a larger magnitude of work for the same pressure rise, This is given by Wa = hon ~ hor = Figg — hig = ¢, (Tos ~ Tox) (11.32) Thus the total-to-total efficiency of the stage is defined by ideal stage work between total conditions at entry and exit actual stage work roses hoy. Tones ~ Toy 1 igs hy Tos = Toy ‘The magnitude of the stage work obtained from velocity triangles with actual velocities and air angles equals the actual work. Thus Eqs. (11.13) when put into Eq. (11.32) give le Ne (11.33) Ios ~ Tig = 4 (2 ~ ey) (11.34a) Hos ~ hoy = wey (am of —tan of.) = ue, (tan By tan 8) (11.346) 1 1 figs ~ hoy = 3 (3-3) + y Ow) (11.34e) [Link]=—_ Asal Compresor Stages 465 Similarly, the actual value of the loading coefficient in Eq. (11.14) is given by ys tata 2 £p Mos = To) 1.35) w a Equation (11.33) when put into Eq, (11:35) gives - a ty ya emt So Tone To) (11.36) we he Equation (11.23) is not valid for an actual flow because > OP : For isentropic (ideal) and incompressible flow, we have from Eq, (11.31) (AP) Pp With the help of this equation the following relations for the stage efficiency are obtained: hoass ~ hor = &p Tosss ~ Tor) (11.37) (AP) a 11.37 We Big = hod ee (AP) st - APs 11.376) MW Be, Cas ~ Tod) eee = OD AED) pu (Cyr — ey) The stage pressure rise (Ap),; can easily be measured on water or mercury manometers and the actual work input can be measured through torque measuring devices. Further insight into this aspect can be obtained from the energy flow diagram given in Fig. 11.9. If the axial velocity through the stage remains constant, then for a = Gz, C1 = C3 Tigg — Boy = hs — fy (11.38) Tog - To = T3- T These relations are also valid approximately for small values of c, and cy. Similarly, for the isentropic process Frases — Hor = Hass ~ hi (11,39) Tosas~ Tor = Tree - Th [Link]466 Turbines, Compressors and Fans With the help of Eqs. (11.38) and (11.39), the efficiency of the comp- ressor stage can be defined as a static-to-static efficiency. = fos hy _ Bo -T Me = Et (11.40) 11.2.2 Degree of Reaction ‘The degree of reaction prescribes the distribution of the stage pressure rise between the rotor and the diffuser blade rows. This in turn determines the cascade losses in each of these blade rows. AS in axial flow turbines (Sec 9.5.2), the degree of reaction for axial compressors can also be defined in 4 number of ways; it can be expressed either in terms of enthalpies, pressures ot flow geometry. (a) For a reversible stage the degree of reaction is defined as = Sentropic change of enthalpy in the rotor isentropic change of enthalpy in the stage From Fig. 11.2 this gives (41) For relatively small pressure changes flow can be assumed incom- pressible, i.e. p ~ constant. 25 (11.42)Asial Compressor Stages 467 In compressor applications pressures are of greater interest. There- fore, in practice the degree of reaction is more frequently defined in terms of pressures which are generally known. (b) For a real or actual compressor stage the degree of reaction is defined as actual change of enthalpy in the rotor actual change of enthalpy in the stage moh. ho 114 in=h Ts ey For ¢) = 3, hy ~ hy = hos — hor = # (2 Sn) Equation (11.25) gives 1 hy hy = 3 = 3) ‘These relations when used in Eq, (11.43) give (11.442) 242 Wi M2 (11.44b) Re Jule 6) ‘This expression can be further expressed in terms of air angles — ef (tan? By ~tan* B2) 2ue, (tan B, ~ tan B) n-3(2) (tan B, + tan B;) (11.440) But & = gand + (tan By + tan B;) = tan By, Therefore, R= tan By (144d) Equation (11.44c) can be rearranged to give R 4s) {an By + tan cy) ~ (tan cy ~tan B)} From Eq, (11.11) tan By + tan oy = ‘Therefore, a=} 2(S) can tan 8) (11.45) [Link]468 Turbines, Compressors and Fans This is a useful relation in terms of the geometry of flow and can be used to study the effect of air angles and the required cascade geometry (to provide these air angles) on the degree of reaction of an axial compressor stage 11.2.3 Low Reaction Stages A low reaction stage has a lesser pressure rise in its rotor compared to that in the diffuser, i.e. (4p)q < (Ap)p. In such a stage, the quantity (tan o% ~ tan f;) is positive or, in other words, a > B; (Fig. 11.1b and Eq, (11.43) The same effect can be explained in another manner. In Eq, (11.45). 6, tan Oy = 6, =u Wy, é tan By = Wa =u oy Therefore, after substituting these values in Eq. (11.45) neke 3(2-%2) (11.46) i 1 Badd omy «aay This equation relates the degree of reaction to the magnitudes of switl or the whirl components approaching the rotor and the diffuser. Thus a low degree of reaction is obtained when the rotor blade rows remove less swit! compared to the diffuser blade rows, i.e. Wy <0 Figure 11.3 shows the enthalpy-entropy diagram for such a stage. The swirl removing ability of a blade row is reflected in the static pressure rise across it. In a low-degree reaction stage the diffuser blade rows are burdened by @ comparatively larger static pressure rise which is not desirable for obtaining higher efficiencies. 11.2.4 Fifty Per Cent Reaction Stages One of the ways to reduce the burden of a large pressure rise in a blade row is to divide the stage pressure rise equally between the rotor and diffuser. To approach this condition (Fig. 11.4) aay = I~ y= Ey ~ hy) (11.48) This when put in Eq, (11.43) gives R= } (fifty per cent reaction) [Link]__Asial Compressor Stages 469 Enthalpy Entropy Fig. 11.3 _Enthalpy-entropy diagram for flow through a low reaction stage (R< 3) Equation (11.44b) for R= 3 gives 1 wav} 2 Butea ew) Substituting from Eg. (11.13¢) we-wh=ch- 4 (1.49) For R= ; . Eg, (11.45) gives o% = By (11.50a) This when substituted in Eq. (11.11) gives % = By (11.50) [Link]420 Trine, Compressors and Fans a ha yi) 2 Po g gita—m) my Entropy Fig. 11.4 Enthalpy-entropy diagram for a fifty per cent reaction stage Equation (11.50) along with Eq, (11.49) yield — Ww =, (151) These relations show that the velocity triangles at the entry and exit of the rotor of a fifty per cent reaction stage are symmettical; these have been shown in Fig. 1.17. The ‘whirl or swirl components at the entries of the rotor and diffuser blade rows are also same. Sy = Wye Wa = 6, (11.52) 11.2.5 High Reaction Stages Zhe static pressure rise in the rotor of a high reaction stage is larger compared to-that in the diffuser, ic. (Ap)_ > (Ap)p, For such a stage the uantity (tan a - tan B,) in Eq, (10.45) is negative, giving R > i. [Link]‘Asial Compressor Stages 471 Therefore, for such a stage B,> m% My? Go Figure 11.5 shows the velocity triangles for such a stage. It can be observed that the rotor blade row generates a higher static pressure on account of the larger magnitude of the swirl component w,, at its entry. ‘The swirl component (c,9) passed on to the diffuser blade row is relatively smaller, resulting in a lower static pressure rise therein. Fig. 11.5 High reaction stage (R> 4, 0% < Ba) Since the rotor blade rows have relatively higher efficiencies, it is advantageous to have a slightly greater pressure rise in them compared to the diffuser. ‘A further discussion on the various degrees of reaction of the axial fan stages has been given in Chapter 14. [Link]TZ Turbines, Compressors and Fans *> 11.3. Flow Through Blade Rows After studying the geometry and thermodynamics of the flow through a compressor stage, further insight can be obtained by looking at the flow in the individual blade rows. Therefore, the two parts of the hs diagram in Fig. 11.2 (for the stage) are redrawn in Figs. 11.6 and 11.7. The similarity between these diagrams must be noted Some aspects of incompressible and compressible flow through diffusers have already been discussed in Sec. 2.4. The flow over a small pressure rise can be considered incompressible, ic. density can be assumed to remain constant with little sacrifice in accuracy. 11.3.1 Rotor Blade Row The flow process as observed by an observer sitting on the rotor is depicted in Fig. 11.6. The initial and final pressures are p, and p, for both isentropic and adiabatic processes. In the isentropic process the flow will diffuse to a velocity w,, giving the stagnation enthalpy and pressure as hoi, and Poi, Fespectively. fay y= 5 Ox P= 5 (133) Enthalpy Entropy Fig. 11.6 Enthalpy-entropy diagram for flow through rotor blade row [Link]__ _ Axial Compressor Stages 473 ‘The actual process gives the final velocity w, and stagnation pressure Pow. Here the same static pressure rise (Ap), occurs with a greater change in the kinetic energy ; (4 - w3). In the ideal or isentropic process this is 1 1 3 OW -WA< 3 OD) This difference is due to the losses and the increase in entropy. ‘The efficiency of the rotor blade row can now be defined by y= eh = (hy = hy) = = Ins) ey hy In -hy =) aches te Re (aLs4) ‘Assuming perfect gas and substituting from Eq. (11.27) 2 eee m1 FBR & (11.55a) w 2 (11.55b) i The assumption of incompressible flow is not required in Eqs. (11.54) and (1155) hy — fgg = (hy — hy) — Ging ty) For incompressible flow, substituting from Eq. (11.53) 1 2 n= Ing 1 Ot -¥D- + @-Pd 2 p 1 1 Ia- bas = 5 {(n +3 put)-(02 +43) Pode 1 hy Ing = = ~ Pooset) = (11.56) Mas = 5 (Pose ~ Poet) = “pg Substituting this in Eq. (11.54) (Apo)x af (11.57a) 2 P (hy —hy) Substitution from Eq. (11.26) gives (11.576)474 Turbines, Compressors and Fans 7 11.3.2 Stator Blade Row The ideal and actual flow processes occurring in the diffuser blade row are shown in Fig. 11.7. Its efficiency is again defined as for the rotor blade row. iy _ (hy ~ hy) ~ (hts ~ sy) fy fy hy (11.58) Enthalpy Entropy Fig. 11.7 Enthalpy-entropy diagram for flow through diffuser (stator) blade row Substituting in terms of the enthalpy loss coefficient 2 & (11.59) — a 2c, (TF; -T) (11.59b) For incompressible flow, ‘A, hy ~ hy = i @o2 ~ Pos) = Code (11.60) [Link]ee Axial Compressor Stages 475 Therefore, the diffuser efficiency can be expressed in terms of the diffuser stagnation pressure loss coefficient (11.61a) (1L.61b) > 11.4. Stage Losses and Efficiency Losses occurring in axial turbine and compressor cascades have been discussed in Chapter 8. Some methods of estimating these losses for compressor blade rows are given in Sec. 8.5.6. In a complete compressor, stage losses due to bearing and disc friction (shaft losses) also occur. Cascade losses*”' depend on a number of aerodynamic and cascade geometrical parameters. Figure 11.8 shows the variation of exit air angles and losses with Reynolds number. It is observed that beyond the Reynolds number of 2 x 10°, the variations are not significant. 40 0.15 > 2 z $35 0.10 8 5 S S 8 a § % 30 0.05 y 25 —1____]00 0 40 20. #30 40 50 Rex 10> Fig. 11.8 Typical variation of exit air angles and cascade losses with Reynolds number [Link]476 Turbines, Compressors and Fans _ a Figure 11.9 shows the energy flow diagram for an axial compressor stage. Figures in the brackets indicate the order of energy or loss corresponding to 100 units of energy supplied at the shaft. Energy from the prime mover or shaft work (100) Stage work Shaft losses (M93 ~ hos) Q (98) Disc Bearing friction loss loss Rotor aerodynamic losses (9) Secondary Profle Annulus Tip Energy at the loss loss loss leakage stator entry (€9) Isentropic work Stator aerodynamic losses (2) @ Secondary _Profie Annulus loss loss loss. Fig. 11.9 Energy flow diagram for an axial flow compressor stage The stage work (igs ~ fg) is less than the energy supplied to the shaft by the prime mover on account of bearing and disc friction losses. All the stage work does not appear as energy at the stator entry on account of aerodynamic losses in the rotor blade row. After deducting the stator (diffuser) blade row losses from the energy at its entry, the value of the ideal or isentropic work required to obtain the stage pressure rise is obtained. The cascade losses in the rotor and stator would depend on the degree of reaction. The values shown in the the energy flow diagram are only to give an example. The ratio of the isentropic work (82) and the actual stage work (98) gives the stage efficiency, whereas the overall efficiency is directly obtained as 82%. [Link]ee ___ Asta Compressor Stages _ 47 11.4.1 Stage Losses Itis seen from Figs. 11.2 and 11.9 that stage losses are made up of the cascade losses (see Sec. 11.2) occurring in the rotor and diffuser blade rows. The loss coefficients (Y or ) ate proportional to the square of the entry velocities. For incompressible flow over compressor blade rows, the pressure loss coefficient is determined by cascade tests. As stated in Sec. 9.5.1, the pressure loss coefficient (Y) can be taken as equal to the enthalpy loss coefficient (£) for a majority of cases. The losses in the rotor and diffuser blade rows are now determined from Fig. 11.2 using Eqs. (11.26) and (11.27). The slopes of the constant pressure lines p, and py with some approximation are given by * hy Ins 2 62} eee (11.62) Stag tao as T; (11.63) Eliminating As from these equations 1, hy, Is = a tad) Substituting from Eq. (11.27) T, 1 fag haoe = Fe BR 9 wi (11.64) From Eq. (11.30) hy = 63 (11.65) Assuming ©) ~ ns Thos — Porss = 3 ~ has hy — ggg = (Its — his.) + (tgs — ass) Substituting from Eqs. (11.64) and (11.65) 1 LR, t hy hay => 5 At BSR MT (11.662) The stagnation state is represented by 03,, at the end of isentropic comp- ression from pressure p, 10 ps. However, due to stage losses, the actual state is represented by point 0; (Fig. 11.2). Therefore, the total stage losses are 1 q fy boss = bo 2 * SRM (11.666) [Link]478 Turbines, Compressors and Fans 11.4.2 Stage Efficiency The total-to-total stage efficiency (7, = 1,) is given by Eq. (11.33) as For (ins = Fos) = (lgs — hose) os — Io Tos — Por Fro — Haass (11.67) Figs ign Equation (11.66b) when put into Eq. (11.67) yields st = Ns Na ie Sin (11.68a) 3 ~ hoy From velocity triangles (Fig. 11.1b), for constant axial velocity © =, sec a w= 6, see B, These relations along with Eq. (11.34b) give Sa cf sec? B, bp ed sec? aay + oe 4) A | Sx sec? B, L Mw =1- 36 oa (21.686) i ~ tan Similar expressions in terms of pressure loss coefficients can also be derived using Eqs. (11.56) and (11.60), = (Apo), 1 In- hy, eon ee (11.69a) GZ t, fae Ig = O= ta) = 3(2) Yaw} (11.69) Iy— hy, = Sede eS ; Ye (11.70) Therefore, the total stage losses are ee lt roa — hoses = 5 Yp 03 + 3(2) Yew} (171) tos~ Ise = 5 2 (% sec? or, + 2 Yy sec? a] (72) 2 - [Link]Axial Compressor Stages 479 These equations when substituted in Eq. (11.67) yield Yat ead 11.73 ny eae (11.73) t= 1-4 ¢— (1.74) 2 Fora given geometry of the stage and flow conditions, the values of the air angles 0, B, and B, and the loss coefficients Yp and Yq can be determined from cascade tests. Alternatively, these values can also be determined by empirical correlations. > 11.5 Work Done Factor‘??? Owing to secondary flows and the growth of boundary layers on the hub and casing of the compressor annulus, the axial velocity along the blade height is far from uniform. This effect is not so prominent in the first stage ofa multi-stage machine but is quite significant in the subsequent stages. Figure 11.10 depicts the axial velocity distributions in the first and last stages of a multistage axial compressor. The degrees of distortion in the axial velocity distributions will depend on the number of the stage (Ist, 2nd, ..., 14th, ete.). On account of this, the axial velocity in the hub and tip regions is much less than the mean value, whereas in the central region its value is higher than the mean. Annulus \ height First Fig. 11.10 Axial velocity distributions along the blade heights in the first and last blade rows of a multi-stage compressor (typical curves) [Link]480 Turbines, Compressors and Fans The effect of this phenomenon on the work absorbing capacity of the stages can be studied through Eq. (11.13b): w= ue, (tan B; ~ tan B,) w=u {e, (tan o% + tan B,)— c, (tan a + tan B,)} Substituting from Eq. (11.11) w=u {u—c, (tan o + tan B)} (1.75) The air angles , and ay are fixed by the cascade geometry of the rotor blades and the upstream blade row. Therefore, assuming (tan a; + tan B,) and w as constant, Eq. (11.75) relates work to the axial velocity at various sections along the blade height. The velocity triangles of Fig. 11.1b are redrawn in Fig. 11.11 for the design value (mean value shown in Fig. 11.10), and the reduced (c, ~ Ac,) and increased (c, + Ac,) values of the axial velocity. Increased incidence Reduced c, Coniiglmwnnnet ee & os Increased cy, incidence Reduced c,, = Design cy Increased c, Fig. 11.11 Effect of axial velocity on the stage velocity triangles and the work It may be seen from the velocity triangles that the work (Eqs. (L.13a) and (11.75)) decreases with an increase in the axial velocity and vice versa, Therefore, the work capacity of the stage is reduced in the central region of the annulus and increased in the hub and tip regions, However, the expected increase in the work at the hub and tip is not obtained in actual practice on account of higher losses. Therefore, the net result is that the stage work is less than that given by Euler’s equation based on a constant value of the axial velocity along the blade height. This reduction in the work absorbing capaciaty of the stage is taken into account by a “workdone factor” O. This varies from 0.98 to 0.85 depending on the [Link].____ Axial Compressor Stages 481 number of stages. Therefore, the work expressions in Eqs. (11.34a) and (11.34b) are modified to fog — bag = 2 U (Ga - Oy) (11.76a) igs — oy = Qu cy (tam Oy — tan O%) = Que, (tan B, - tan By) (11.76b) Figure 11.12 gives the mean values of the work-done factor‘? to be ‘used for each stage in a given number of stages (shown on the X-axis). 1.0 S095 3 = 0.90 3 = 0.85 0.80 0 4 8 12 16 Number of stages. Fig. 11.12 Variation of work-done factor with number of stages of axial flow compressor (from Howell and Bonham’ oF by conurtesy of the Instn. Mech. Engrs. London) > 11.6 Low Hub-Tip Ratio Stages For higher flow tates, the cross-sectional area of the low pressure stage in axial compressors must be large. This requires relatively larger mean diameters and blade heights leading to hub-tip ratios (d,/d,) much below unity (see Sec. 9.9). The variation of various flow parameters in the radial direction (along the blade height) in such stages are significant and depend on the conditions imposed in their design. Some of these designs including the radial equilibrium (Sec. 9.9.1) and free vortex (See. 9.9.2) are discussed in the following sections. 11.6.1. Radial Equilibrium Flow through turbine blade rows with radial equilibrium has already been discussed in Sec. 9.9.1. Equation (9.109b) is equally applicable to [Link]482 Turbines, Compress. and Fans compressor blade rows. However, for further understanding this equation is derived here in a different manner. From Euler’s momentum equation (6.46) for c, . rea 2 (11.77) For isentropic incompressible flow, Po=P* 5 pe= p+ } p(dreh+dy For radial equilibrium (c, = 0) PoP 5 pleh+ 2) i.78) d Lod bobsled cee) 1 dp) _ 1 dp dey de, 1dpo dp, deg, (pli pir ane Substituting from Eq, (11.77) 2 cy 1 4, Mey Ld (11.79) ee dee re i For some conditions in the flow through compressor blade rows, the Stagnation pressure can be assumed constant along the blade height, ie py = ra This condition when applied in Eq. (11.79a) gives dX de, | cb Mo 4g Me Coe te Gt S =0 (11.798) This can be reduced to the following form: 3 ceeF 4 2 (e,F=0 (11.79¢) 11.6.2 Free Vortex Flow A fice vortex turbine stage was discussed in Sec. 9.9.2. Compressor stages have also been designed for free vortex flow. This condition requires re, = constant, Therefore, Eq (11.79) gives diye yee =0 ¢, = constant along the blade height [Link]=¢, an Cath = Cet = Colm xan = Cran = Cram = Cx Axial Compressor Stages 483 (11.80) (11.806) ‘The variations of the tangential velocity components cg, and cg, are governed by the following relations: Ta Copa = "Loy = MnO = C1 Ty Cogn = FE gt = TibOam = C2 (11.81a) (11.816) Constants C, and C, are known from the mean section velocity triangles. Air angles ‘The above relations for a free vortex stage can give air angles shown in Fig, 11.1b. Here cy, = cg) and cy» = Cg Rotor entry G tan 4 = jolie le Cth Tex From Eq, (11.11) 4" tan By, = es ~ tan Oy, aM tan By, e Tits Similarly, for the tip section « q ‘ay tan oy, = 8 - oat s tan By, = Bux Rotor exit Con _ C, tan Oy, = 2% = —2 Ce From Eq. (11-11) y tan yy = ~~ tan Ory tan Boy, tan Oy tan Ba, = (11.82a) (11.82b) (11.83a) (11.83b) (1L.84a) (11.846) (11.85a) (11.856)484 Turbines, Compressors and Specific work The work done per kilogram of the flow at a given radial section (r= r) is given by Eq. (11.13a) w= hoa ho, = U (cg, ~ ¢9,) Equation (11.81) gives w=or (2 . $) ror W = Igy ~ hg, = @ (Cy ~ C;) = const. (11.86) Degree of reaction Since the air angles in the stage are varying along the blade height, the degree of reaction must also vary. Substituting from Eqs. (11.82b) and (11.84b) in Eq. (11.440) R=1fa)\iu_G,u G 2lu le re, Te, re, At, 2er? (11.87) (11.88) (1.89) This shows that the stage reaction in a free vortex design increases along the blade height, Figure 11.13 shows the variation of air angles and degree of reaction along the blade height in a free vortex stage 11.6.3 Forced Vortex Flow Im a forced vortex flow through the stage the tangential velocity component is directly proportional to the radius. 2 = const. = C (11.90) ‘his when used in the radial equilibrium flow equation (11.79) gives 44 eprp, 4 ope (crf + 4 (F=0 This on simplification yields dc, =-4C? r dr [Link]__Asial Compressor Stages 485 50), 40 30 = 20 - 19 10 083 os 8 0 Jos g 402 8 & 10 0.0 th Tm Te Distance along the blade height Fig. 11.13 Variation of air angles and degree of reaction along the blade height in a free-vortex stage ‘After integration and applying it at the rotor entry 2 =Ky-2Ci" (11.91) Constants C, and K; are known from conditions at the mean radius (7,)- Cy _ Soh Sou = On =¢, (11.92) Tm (11.93), Thus, with known distribution of ‘the quantities u, c, and cg at the rotor entry, the velocity triangles and hence the air angles a and B, along the blade height are known. At the rotor exit, fm 7a 2 ee ‘Therefore, the specific work done along the blade height is given by cogil ee estar c esl Gy (11.94) w= hp ho, = @ (Co, ~ ey) Ww = gy — hoy = (C2 - C1) OF (11.95) Now an expression (similar to (Eq. 11.91) for the axial velocity distribution at the rotor exit can be derived assuming /o, = constant.ines, Compressors and Fans For radial equilibrium conditions, fides en cone dint cg 2 +e, Se a a te This can be written as lid 21d 2 (dh Tap Mot SS GPa (2) (11.96) From Eg. (11.95) a (3), =2(G-C) or (11.97) 4 b (re9) = ae (C3 r')=4c3r (11.98) Equations (11.97) and (11.98) when substituted in Eq. (11.96) give (6) = 4 (Cy C) or- C3 r] dr After integration hy = Ky +2 (G.-C) a -2 C3? (11.99) The new constant (K;) of integration can again be determined from the value of ¢,, at the mean radius. Ky = Chay = 2 (Cy C,) or? +2 C32, (11.100) Equation (11.44c) for the degree of reaction cannot be used here because at a given section c,, # ¢,,. 11.6.4 General Swirl Distribution Compressor blade rows have also been designed by prescribing a general distribution of the tangential (swirl or whirl) velocity component along the blade height. The general expression is b g= ars © com ar © At the rotor entry, 9, = ar" — 6 (11.101) At the rotor exit, co, = ar" + & (11.102) , The specific work is given by w=u Cy ~ co) [Link]- ; ___ Arial Commnesor Stages 487 Substituting from Eqs. (11.101) and (11.102) Ww = Igy — figy = 206 (41.103) This shows that the specific work remians constant along the blade height. Along with the above swirl distribution the axial velocity is assumed constant for writing down the expression for the degree of reaction. From Eq. (11.44b) and Fig, 11.1b, = 5-52 (or = Wea) or + Won) 2u (Con ~ 01) 2u (co ~ Ce) From Eq. (11.12a) or Fig. 11.1b), We — We = Cor — Cai Therefore, re “atte (11.104a) (11.1046) (11.105) This equation is only approximately valid because it will be seen in the following sections that the axial velocity across the rotor at a given section does not remain constant. ‘Two cases are considered here: @ n=0 The degree of reaction from Eq. (11.105) is a R aan or (11.106) and increases along the blade height. The axial velocity distribution is obtained by applying Eq. (11.79). At the rotor entry, (reg =P (0-4) = @P —2abr+ # ao Ff 4 (rep)? = 2a°r - 2ab [Link]488 Turbines, Compressors and Fans Substituting this in Bq. (11.79c) L (2a? r ~ 2ab) dr =~ dle)? ; 2, = const, + 2a? J (& ar 2, = const. ~ 20° (S+inr) (1.07) At the rotor exit a (remy? =? («+2) = a'r? + Qabr + B , 4 (egy =2a°r + Se Wea) = 2a’r + 2ab Substituting this in Eq. (11.79) -E Qer + 2ab) dr =~ dle)? cz iy = const, — 2a? (inr- (11.108) ii) n=1 Equation (11.105) for this condition gives a =1-4 10! R-1-4 (11.109) Thus a stage with constant reaction is obtained. However, this is only approximately true because here also cy) # cy. The axial velocity distribution is given in the following sections: At the rotor entry Pe? (arb) aA roby? + Be (req) ar = ar — abr? + BF (rea? = ar 4% eh = cons, 42 J (r-4 ar Lob 2, = const, — 4a? G r ~P nr) (11.110) [Link]Axial Compressor Stages 489 At the rotor exit By (ep =P (+2) = dr! + Dabr? + (req) = 4a’ +4 2 ia 2 dr deg? = 4a (+2) ar rs 2, = const. — 4a (Fr+2mnr) (ut) ‘The values of the constants a and can be determined by known values of Com and Com at the mean radius. Similarly, the constants in Eqs. (11.107), (11.108), (11.110) and (11.111) are determined from known values of Cyiq and C,_ With known values of 1 cous Cais Cop and cx, the air angles 0, By, 0, and B, along the blade height can be determined ‘> 11.7 Supersonic and Transonic Stages Recent developments in materials and bearing design have made higher peripheral speeds (u ~ 600 mvs) possible in compressors. Higher peripheral speeds lead to supersonic fluid velocities (relative or absolute) in the blade passages which can then be employed for compression through a shock (normal or oblique) over a small axial distance. Supersonic flow #41 in axial compressor stages can also occur unintentionally due to local acceleration of the flow on the blade surfaces, generally on the suction side. This happens when the inlet Mach number is in the proximity of 0.75. When a compressor stage is intentionally designed as supersonic, the flow is supersonic in some part or parts of the stage and a significant part of the static pressure rise is obtained by compression through shock waves, A shock wave is an irreversibility and leads to an increase in entropy and stagnation pressure loss. Therefore, supersonic compressors can provide a higher pressure ratio (= 4 — 10) in a single stage with a relatively lower efficiency (~ 75%) on account of additional losses due to shocks. Table 11.1 gives a clear picture of the orders of pressure rise and stagnation pressure loss for some representative values of the upstream Mach numbers. The main advantages and disadvantages are obvious from the table. Some of them are given below. [Link]490 Turbines, Compressors and d Fans 12 0.843 1513 0.993 Ls 0.701 2458 0.930 2.0 0.578 4.500 0.721 eet 0.513 7.125 0.500 _ Advantages 1, Very high pressure ratio per stage 2. Low weight-to-power ratio 3. Small size and length of the machine 4. Higher flow rates, Disadvantages 1, Excessive loss due to shocks 2. Early separation of the boundary layer on the suction side leading to increase in the profile and annulus losses 3. Very steep or almost vertical performance characteristic leading to unstable operation 4, Difficulty in starting Excessive vibration due to instability of flow w 6. Serious stress and bearing problems on account of very high peripheral speeds. Figure 11.14 shows a supersonic compressor stage with shock in the rotor. The velocity triangles are shown in terms of the Mach numbers corresponding to the velocities. The flow approaches the stage at subsonic velocity ¢, or Mach number M,,. This with a blade Mach number M, gives a supersonic relative Mach number (M,,1) at the entry of rotor blades, The rotor blade passages are so designed that the entering supersonic flow is first converted to a subsonic flow through a shock and then it is subsonically diffused to a relative Mach number M,». The flow at the entry of the stator row is also subsonic. Such a stage can develop a pressure ratio of about 3.0 Another scheme is shown in Fig, 11.15. The flow approaches both the rotor and stator at supersonic velocities. Therefore, besides subsonic diffusion, compression through shocks in both the rotor and stator is Possible. This arrangement can give a very high pressure ratio (~ 6.0) per stage.Axial Compressor Stages 491 Moy = My V- [Link]492 Turbines, Compressors sand Fans (b) supersonic rotor and subsonic stator. (©) subsonic rotor and supersonic stator. Choking of the flow occurs if the velocities are such that the axial Mach number is unity or higher. Therefore, the velocity triangles (blade geometry) are so chosen that the axial Mach number is always less than unity. Blades in rows receiving supersonic flow must have sharp leading edges to avoid strong detached shocks and excessive losses arising from them. To retain some advantages of high speed compressors without suffering too much from their disadvantages, low Mach number supersonic flow or high Mach number subsonic flow can be employed. Such compressor stages are known as transonic stages": #! with flow Mach numbers varying in the range 0.85 ~ 1.3. Such stages do not suffer from unstable flow and have relatively higher efficiencies. The flow in such stages is generally supersonic towards the tip sections of the blades. ‘> 11.8 Performance Characteristics A brief introduction to compressor performance has been given in Sec. 7.7. The performance characteristics of axial compressors or their stages at Various speeds can be presented in terms of the plots of the following parameters: (a) pressure rise vs. flow rate, 4p = f(Q) Ap =f (mn) (©) pressure ratio vs. non-dimensional flow rate (Fig. 7.5), (*s) 7a. Pt Po. (©) loading coefficient vs. flow coefficient (Fig. 7.6), v=sS@ The actual performance curve based on measured values is always below the ideal curve obtained theoretically on account of losses, This is shown in Fig. 11.16. The surge point and stable and unstable flow regimes have been explained in the following sections, 11.8.1 Gff-design Operation A compressor gives its best performance while operating at its design Point, i.e. at the pressure ratio and flow rate for which it has been [Link]Axial Compressor Stages 493 Stage losses Pressure rise Flow rate Fig. 11.16 Ideal and actual performance curves for an axial com- pressor designed. However, like any other machine or system, it is also expected to operate away from the design point. Therefore, a knowledge about its behaviour at off-design operation is also necessary. Off-design characteristic curves can be obtained theoretically from Eqs. (11.16) and (11.11). y= @ (tan @, — tan 0%) But fan Ob = 5 —tan B,. Therefore, w= 1- 6 (tan B, + tan 0%) (11. 112a) The quantity (tan B, + tan a4) can be assumed constant in a wide range of incidence up to the stalling value i,. This is justified in view of small variations in the air angles at the rotor and stator exits. Therefore, writing % = Os A=tan B, + tan of (11.113) If the design values are identified by the superscript*, Eq. (11.112a) along with (11.113) can be written as yr =1- Agr (11.112b) l-y* A= o* At off-design conditions yw=1-4o [Link]494 Turbines, Compressors and Fans yw=1-(l- yw) < (11.114) This equation also gives the off-design characteristic of an axial-flow compressor. Figure 11.17 depicts theoretical characteristic curves for some values of the constant A. For positive values of 4, the curves are falling, while for negative values rising characteristics are obtained. The actual curves will be modified forms of these curves on account of losses. 1.0 0.75 0.50 Loading coefficient 0.25 m ° 0 0.25 0.50 0.75 1.0 Flow coefficient Fig. 11.17 Off-design characteristic curves for an axial com- pressor stage 11.8.2 Surging Unstable flow in axial compressors can be due to the separation of flow from the blade surfaces or complete breakdown of the steady through flow. The first pehnomenon is known as stalling, whereas the second is termed as surging.*°” “’ Both these phenomena occur due to off- design conditions of operation and are aerodynamically and mechanically undesirable. Sometimes, it is difficult to differentiate between operating conditions leading to stalling and surging. It is possible that the flow in some regions stalls without surging taking place. Surging affects the whole machine while stalling is a local phenomenon. Some typical performance characteristic curves at different speeds (My, Np, etc.) are shown in Fig. 11.18. The surge phenomenon is explained with the aid of one of the curves in this Figure. Let the operation of the [Link]___ Axial Compressor Stages 495 Mg ° A 2 4 s Nw urge e tine a Ne s4+— | | “ | Tg tip tig on Flow rate Fig. 11.18 Surging in compressors compressor at a given instant of time be represented by point A (P4, rit,) on the characteristic curve (speed = constant = N,). If the flow rate through the machine is reduced to rity by closing a valve on the delivery pipe, the static pressure upstream of the valve is increased. This higher pressure (pp) is matched with the increased delivery pressure (at B) developed by the compressor. With further throttling of the flow (to tire and its), the increased pressures in the delivery pipe are matched by the compressor delivery pressures at C and $ on the characteristic curve. ‘The characteristic curve at flow rates ‘below it, provides lower pressure as at D and E. However, the pipe pressures due to further closure of the valve (point D) will be higher than these, This mismatching between the pipe pressure and the compressor delivery pressure can only exist for a very short time. This is because the higher pressure in the pipe will blow the air towards the compressor, thus reversing the flow leading to a complete breakdown of the normal steady flow from the compressor to the pipe. During this very short period the pressure in the pipe falls and the compressor tegains its normal stable operation (say at point B) delivering higher flow rate (ring). However, the valve position still corresponds to the flow rate itp. Therefore, the compressor operating conditions return through points C and S to D. Due to the breakdown of the flow through the compressor, the pressure falls further to pg and the entire phenomenon, i.e. the surge cycle EBCSDE is repeated again and [Link].i496 Turbines, Compressors and Fans again. The frequency and magnitude of this to-and-fro motion of the air (Surging) depend on the relative volumes of the compressor and delivery pipe, and the flow rate below ,. Surging of the compressor leads to vibration of the entire machine which can ultimately lead to mechanical failure. Therefore, the operation of compressors on the left of the peak of the performance curve is injurious to the machine and must be avoided. Surge points (S) on each curve corresponding to different speeds can be located and a surge line is drawn as shown in Fig. 11.18. The stable range of operation of the compressor is on the right-hand side of this line There is also a limit of operation on the extreme right of the chara- cteristics when the mass-flow rate cannot be further increased due to choking. This is obviously a function of the Mach number which itself depends on the fluid velocity and its state. 11.8.3 Stalling As stated earlier, stalling is the separation of flow from the blade surface, At low flow rates (lower axial velocities), the incidence is increased as shown in Fig, 11.11. At large values of the incidence, flow separation occurs on the suction side of the blades which is referred to as Positive stalling. Negative stall is due to the separation of flow occurring on the pressure side of the blade due to large values of negative incidence. However, in a great majority of cases this is not as significant as the Positive stall which is the main subject under consideration in this section. The separation of flow on aerofoil blades has been discussed in Sec. 6.1.18. Losses in blade rows due to separation and stalling have been explained in Sec, 8.4.5 In a high pressure ratio multi-stage compressor the axial velocity is already relatively small in the higher pressure stages on account of higher densities. In such stages a small deviation from the design point causes the incidence to exceed its stalling value and stall cells first appear near the hub and tip regions (see Sec. 11.5). The size and number of these stall cells or patches increase with the decreasing flow rates. At very low flow rates they grow larger and affect the entire blade height. Large-scale stalling of the blades causes a significant drop in the delivery pressure which can lead to the reversal of flow or surge. The stage efficiency also drops considerably on account of higher losses. The axisymmetric nature of the flow is also destroyed in the compressor annulus, Rotating stall Figure 11.19 shows four blades (1, 2, 3 and 4) in a compressor rotor. Owing to some distortion or non-uniformity of flow one of the blades (say [Link]Increased Reduced Propagating Unstalling stall cell Fig. 11.19 Stall propagation in a compressor blade row the third) receives the flow at increased incidence. This causes this blade (aumber three) to stall. On account of this, the passage between the third and fourth blades is blocked causing deflection of flow in the neighbouring blades. As a result, the fourth blade again receives flow at increased incidence and the second blade at decreased incidence. Therefore, stalling also occurs on the fourth blade, This progressive deflection of the flow towards the left clears the blade passages on the right on account of the decreasing incidence and the resulting unstalling. Thus the stall cells or patches move towards the left-hand side at a fraction of the blade speed. In the relative system they appear to move in a direction opposite to that of the rotor blades. However, on account of their (stall cell) lower speed as compared to that of the rotor, they move at a certain speed in the direction of the rotation in the absolute frame of coordinates. Rotating stall cells" develop in a variety of patterns at different off-design conditions as shown in Fig. 11.20. The blades are subjected to forced vibrations on account of their passage through the stall cells at a certain frequency. The frequency and amplitude of vibrations depend on the extent of loading and unloading of the blades, and the number of stall cells. The blades can fail due to resonance. This occurs when the frequency of the passage of stall cells through a blade coincides with its natural frequency. Both the efficiency and delivery pressure drop considerably on account of rotating stall. Tey tM CLak dic ccm a Constant A Constant, area of cross-section [Link]| ‘Ssosseidwod |e|xe UI sija0 yJeIs Bune}oy Oz"LL “B14 [Link]_ Axial Compressor Stages _ 499 Constant Fluid velocity Specific heat at constant pressure Constants Diameter Enthalpy Change in enthalpy Constants Mass-flow rate Mach number Index of r Rotor speed Pressure Static pressure rise Stagnation pressure loss Power Volume-flow rate Radius Degree of reaction, gas constant Reynolds number Change in entropy Absolute temperature ‘Tangential or peripheral speed of the blades Work, relative velocity Pressure loss coefficient Greek Symbols a DeqevmMsrd Subscripts ° 1 Air angles in the absolute system Air angles in the relative system Ratio of specific heats Efficiency Enthalpy loss coefficient Density Flow coefficient Stage loading coefficient Rotational speed in rad/s Work-done factor Stagnation values Rotor entry [Link]500 Turbines, » Compressors and Fans 2 Rotor exit 3 Stator or diffuser exit a Actual 6 Blade © Corresponding to velocity ¢ D Diffuser or stator h Hub m Mean, mechanical r Radial rel Relative R Rotor S85 Isentropic ‘SS Static-to-static st Stage t Tip tt total-to-total w Cortesponding to velocity w x Axial y Tangential 8 Tangential ‘> Solved Examples 11.1 An axial compressor stage has the following data: Temperature and pressure at entry 300K, 1.0 bar Degree of reaction 50% Mean blade ring diameter 36 cm Rotational speed 18000 rpm Blade height at entry 6m Air angles at rotor and stator exit 25° Axial velocity 180m/s Work-done factor 0.88 Stage efficiency 85% Mechanical efficiency 96.1% Determine: (a) Air angles at the rotor and stator entry, (b) the mass- flow rate of air, c) the power required to drive the compressor, (d) the loading coefficient, (e) the pressure ratio developed by the stage and (f) the Mach number at the rotor entry.‘Axial Compressor Stages 501 Solution: ndN _ 2x036%18000 _ 0 60 ae fama (a) Referring to Fig. 11.1, G1 = ee tan oF, = 180 tan 25 = 83.935 mis Wy = 339.292 ~ 83.935 = 255.357 mis tan B= 33357 — ars B, = 54.82° (Ans.) Since the stage has 50% reaction ct, = By = 54.82%; 0 = = B, = 25° 10x 10° ®) A mal = 987x300 ti = py cy (aed I) tn = 1.161 x 180 (7x 0.36 x 0.06) th = 14.18 kgs (Ans.) 339.292 m/s ua = 0.530 = 1.161 kg/m? (0) Specific work w = Qu o, (tan B, ~ tan B;) w = 0.88 x 339,292 x 180 (1.418 ~ 0.466) w = SI164.15 Skg, pa Le jinp= M18x 51164 ™ 0.967 P = 750 kW (Ans.) $1164.15 id) = Ws 2, = 0.444 (Ans. @ ve (339.292) (ns) w _ SU6AIS ~ so.919c @ AN o = T005 Isentropic temperature rise AT, = T, (10% - 1) = Thy AT, = 0.85 x 50.91 300 (p,27 — 1) = 43.273 ome 14 43273 14 Py 1+ 300 1.144 p, = 1.6 (Ans.) [Link]502 1 ines, Compressors and Fans 180 180 wy "e088, 6035482 ~ 0576 ~ 3125 mis ‘The relative Mach number at the rotor blade entry Mo = Me 3125 wVyRt, Jiax287x300 = 3125 _ m1 = 34718 ~ 090 This Mach number will give a shock on the suction side of the rotor blade due to local acceleration and deceleration, The Mach number at the rotor blade tips will be slightly higher than this, To avoid the possibility of shocks, the maximum value of the Mach number must be kept below 0.75, 11.2. The conditions of air at the entry of an axial compressor stage are P= 768 mm Hg and 7; = 314 K. The air angles are B.=5I%, h=9, a = 0, =7° The mean diameter and peripheral speed are 50 cm and 100 m/s, respectively. Mass-flow rate through the stage is 25 kg/s; the work. done factor is 0.95 and mechanical efficiency 92%. Assuming a stage efficiency of 88% determine: (@) air angle at the stator entry, (b) blade height at entry and the hub-tip diameter ratio, (©) stage loading coefficient, (d) stage pressure ratio, and (©) the power required to drive the stage, Solution: ~L= = 1.136 kg/m? 287x314 (@) Equation (11.11) is tan @ + tan B, tan 7 + tan 51 = 100 = 0.1228 + 1.2349 = 1.3577 100 = 13577 > 73.65 mis [Link]7 _ Axial Compressor Stages 503 tan o + tan B= tan Of + tan 9 = tan of + 0.158 = 1.3577 (b) () @ fe) 4199; of = 50.18° (Ans.) tit = expy (Thy) 25 = 73.65 1.136 Xm X05 hy hy = 0.19 m= 19 em (Ans.) d, = 50 + 19 = 69 cm d, = 50-19 = 31 om The hub-tip ratio is a _ 3h ad, w=c, AT, = Que, (tan B, ~ tan B) w = 0.95 * 100 x 73.65 (1.2349 ~ 0.158) w = 7534.8 Ikg = 0.449 (Ans.) _ ow _ 75348 _ 97535 v= = Taoctan 7 07535 Ans) ar, = % = 1348 = 7497 c, 1005 AT, = T, (p98 — 1) = Ty AT, = 0.88 x 7.497 = 6.597 ase a ee OTe D, 1 314 0.021 Py = (1.021)° = 1.075 (Ans.) ‘Alternatively, the pressure ratio can be determined by assuming incompressible flow. From Eq, (11.37b) (Ap)op = Noe PW = 0.88 X 1.136 x 7534.8 (Ap), = 0.0753 x 10° Nim? s (ap), = 4 10" = 767.8 mm WG. 1,024 + 0.0753 _ Pia mC see This is a slight underestimation on account of the assumption, raw _ 25x 75348 Tn 0.92, P = 204.75 kW (Ans.) jg x 10° [Link]504 Turbines, Compressors and Fans 11.3 (a) Prove that the efficiency of a 50% reaction axial compressor stage is given by ee cee ian B, - tan B, ©) m= p= Te” (©) In the stage of Ex. 11.1, if the loss coefficient for the blade rows is 0.09, verify the value of its efficiency. (d) Determine the efficiencies of the rotor and diffuser blade rows. TNs Solution: (@) For a 50% reaction stage C1 = C3 = Woy Wy = Cy, Oy = 0 = By, B, = a Therefore, the cascade losses in the rotor and stator blade rows are the same, i.e. Yp = Yp = 0.09 For 7; ~ 7, Eq. (11.74) is G1, + Yp sec? By Mal 7 6 an, tan fi, Ny = 1 2 2 by aoe & aw Therefore, Eqs. (11.576) and (11.61b) yield Te = N= Ne o $= 053, Yq= 0.09 0 = B, = 54.82°, B, = a, = 25° 053 x 0.09 sec? 5482 ~ 1418-0466 0849 Thy = 84.9% (Ans.) This is very close to the assumed value of 85% in Ex. 11.1. (@ From velocity triangles (Fig. 11.1b), for constant axial velocity, st 2 2 3 a = fut ; + = 0.404 sec? ay sec? 5482 [Link]___ Axial Compressor Stages 505 ‘Therefore, Eq. (11.61b) becomes Yo Mp = Me=1- = ee 1-c/e Tip = N= 84.9% (Ans. In this case this is the same as 7), 11.4 Assuming the data of Ex. 11.2 at the mean blade section ("= 1), compute: (a) rotor blade air angles, (b) the flow coefficient, (©) the degree of reaction, (a) the specific work, and (©) the loading coefficient at the hub, mean and tip sections. Assume free vortex flow. Solution: Refer to Fig. 11.1b and replace the suffix y by @ to denote tangential direction Ym = 100 @= T= 935 7 400 adls r, = 05 X31 = 15.5 cm uy, = @ 7, = 400 x 0.155 = 62.0 m/s 1, = 0.5 x 69 = 34.5 cm u, = @r, = 400 x 0.345 = 138 m/s (a) Air angles The air angles at the mean section are Gin = 7, Bin = 51% Boy = 9s am = 50.18° Coim = Gq ta yy, = 73.65 tan 7 = 9.04 m/s Cy =n Coim = 0.25 X 9.04 = 2.26 CG _ 226 = Ge = 226 = 14.58 mi can = P= Gigs = M458 mis q 2.26 = Gi 228 655 ms cau = SE = Fogg = 655 mis tan a, = SH = ies = 0.1978 oy, = 11.19° [Link]506 Turbines, Compressors and Fans tan oj, = Su & %, = 5.08° tan By, = a = tan Oy, = ee ~ 0.1979 = 0.6439 By, = 32.78° (Ans.) tan By, ~ 2 —tan a, = B89 _ qosgo = 1.7848 By, = 60.74° (Ans.) Coom = C ta Oy, = 73.65 tan 50.18 = 88.33 Cy = Tui Cony = 0.25 X 88.33 = 22.084 = 22.084 _ Can = Sine = 142.47 mis tan Bay = 5555 ~ 0.869 = 1.0047 Bh, = 45.135° (Ans.) ()) Flow coefficients = 1.188 (Ans.) = 0.7365 (Ans,) & 2 BOS _ 9 533 a BBs 0.533 (Ans.) (©) Degrees of reaction 1 R= > oy (tan By, + tan By) x 100 R, = 0.5 x 1.188 (0.6439 — 1.092) x 100 [Link]_ Ail ComrsorSaes_507 Ry = ~ 26.62% (Ans.) Ry, = 05 X 0.7365 (1.2349 + 0.158) x 100 Ry = 51.29% (Ans.) R, = 05 x 0.533 (1.7848 + 1.0047) x 100 R, = 74.34% (Ans.) (a) Specific work In a free-vortex flow the specific work remains constant at all sections. w= @(C,~C) w-= 400 (22.084 - 2.26) x 10° ki/kg w = 7.9296 kikg (e) Loading coefficients _ ow _ 79296 WB axe” 2.063 (Ans.) = DBE _ 9.793 (Ans.) “mie, 100x100 = 2 19296 _ Wa eo Gpesase 7 Ot Am) The results obtained are presented in the following table for compatison: Free-vortex stage Hub 32.78 47.52 26.62 7.93 1.188 2.063 Mean = SLO 9.0 51.29 793 0.7365 0.793 Tip 60.74 45.135 7434793 0.533_—0.416 11.5 A forced vortex flow axial compressor stage has the same data at its mean diameter section as in Ex. 11.2. Determine (a) rotor blade air angles, (b) specific work, (¢) loading coefficients, and (a) degree of reaction at the hub, mean and tip sections. Solution: ‘The air angles are Olt = 7°, Bim = 51° Olam = 50-18%, Bam = 9° The radii and tangential velocities are 1, = 15.5 om, uy = 62.0 mis rq, = 25 CM, Uy, = 100-m/s [Link]508 Turbines, Compressors and Fans Angular speed Com (a) Air angles = 345 om, u, 38 m/s 400 rad/s = Cay, = C= 73.65 mis Com = Cy ta Op, = 73.65 tan 7 = 9.04 mis C= fom _ 9.04 _ i 025 36.16 Can = 74 Cy = 0.155 x 36.16 = 5.605 m/s Cou = Cy = 0.345 x 36.16 = 12.475 mis Com = Ky ~ 2C§ ry 73.65? = Ky ~2 x 36.16? x 0.25? K, = 5587.76 cian = Ky ~ 2Ch Chay = 5587.76 — 2 x 36.16? x 0.155? Can = 74.33 m/s Chae = 5587.76 — 2 x 36.16 x 0.345? Car = 72.64 mis Cou _ 5605 = fan ~ 5605 _ 997 a a Com 7433) gee tan By, =“ — tan a, = £29. _ 0.0754 = 0.7587 sty a Byy = 37.18° (Ans) = fou = WATS _ tan = (8 = EA = 0.1717 a ~ B8 = tan Bye Se tan y= EF 0.1717 = 1.728 By, = 59.94° (Ans.) Ceom = Cy M Oyy, = 73.65 tan $0.18 = 88,33 ~ Sem 5, 8833'> C= Ste = OP = 353.32 Coam = Fy C= 0.155 x 353,32 = 54.76 mis Cony = 7; Cy= 0.345 x 353.32 = 121.89 mis Ky = Gan 2 (Cy ~ Cy) 2, + 263 72, K, = 73.65? ~ 2 (353.32 — 36.16) x 400 x 0.25? +2 x 353.32? x 0.25? [Link]a _Axial Compressor Stages 509 K, = 51709 Gay = Ky + 2 (C~ Cy) OFF + 2G Gy, = 5170.9 + 4058 rf = 5170.9 + 4058 x 0.1557 Cay = 72.58 ws ey, = 5170.9 + 4058 x 0.345? = 5653 Cg = 75.18 mm/s = fo. S476 _ oe nase O74 an By-= = 22 _ = tan Bry 7asg ~ 0-754 = 0.10 Bay = 5.71° (Ans.) = £62, = 12189 _ tan oy = 2 = 751g ~ 1624 ee = 138 = tan By, = co, 7 MO BiB 1.621 = 0.215 By, = 12.19 (Ans.) (b) Specific work Wy = (Cy— Cy) Or} = 126.86 17, Kirkg Ww, = 126.86 x 0.155? Wy = 126.86 x 0.25? = 3.05 kikg (Ans.) 7.93 kikg (Ans.) w, = 126.86 x 0.345? = 15.1 ki/kg (Ans.) (c) Loading coefficients = Ma = _3050_ , Wo iP” Gx 0.793 (Ans.) It can be shown that the loading coefficient for a forced vortex stage remains constant along the blade height. This can be checked here. = Wm = —7930_ _ Ym = TE ~ Tox 100 0.793 (Ans.) w, _ 15100 We [2 138% 138 (@) Degree of reaction = 0.793 (Ans.) Since the axial velocity at a given section is varying, the degree of reaction is obtained from510 Turbines, Compressors and Fans _ 4 sec B, ~c2, sec? By 2w At the hub section p, = 1433" sec? 37.18 - 72.58? sec? 5.71 * 23050 Ry = 55.46% (Ans.) At the tip section p, = 12.64 sec 59.94 ~ 7518? seo? 12.1 7 ‘ 2x15100 x 100 100 R, = 50% (Ans.) The results are summarized in the following table Forced vortex stage (variable reaction) Hub 37.18 S71 55.46 305 - 0,793 Mean 51.0 9.0 51.29 7.93 0.7365 0.793 Tip 59.94 12.4 0.00 15.1 - 0.793 11.6 If the stage in Ex. 11.1 is designed according to the general swirl distribution “ b a1 ; b Seat co-a+ taking the same conditions at the blade mid-height, compute air angles at the rotor entry and exit, specific work, loading coeffi- cients and degrees of reaction at the hub, mean and tip sections. Solution: Coim = C im tN OL yy, ~ 180 tan 25 = $3.93 m/s Co2m = Cc2m tah Op», = 180 tan 54.82 = 255.36 mis a + (Coim + Coan) = 0.5 (83.93 + 255.36) a= 169.65 2b = Co2m — Coim = 255.36 — 83.93 = 171.43 b=05 x O18 x 171.43Axial Compressor Stages 511 Air angles be _ 1543 _ / coup = a~ 2 = 16965 {Gig = 65.78 mls 1543 = 169.65 — — = 96, Is con 69.65 ~ G34 17 mis com = at © = 169.65 + 1943 = 979.52 ms , “015 Cyn = a+ © = 169.65 + 18 = 243.13 m/s hi oat From Eq. (11.107) / ig = K, - 20 | hein) ay Ky = 180? +2 x 169.65? (as +in01s) K, = 37250 Gy = — 37250 — 2 X 169.67 (Coe +n 01s) cap = 192.5 mis 0,09095 y= — 37250 ~ 2 x 169.65" (Br +in021) Gq) = 166.30 mis tan Oy, = pe sore = 0.347 Cap 192 tan By = He — tan Oy 282.75 _ 9 347 = 1.1218 sth By, = 48.28° (Ans.) = Cou . 9617 2 tan hy = Oe = 7663 0.578 4 295.85 =o - = 1.802 tan y= E = tan = “T663 — 0,578 = 1.80: By, = 60.97° (Ans.) From Eq, (11.108) \ Ey Ky 2" (inry = 2} K ay 0.09095 \ = 190? + 57562 (nos OTS : [Link]512 Turbines, Compress K, =~ 95388 i 0.09095 ay = ~ 95388 — 57562 (in S-—W ) ay = 220.65 mis c2,, = — 95388 — 57562 (in o21- or) Ca, = 139.3 mis tan By = oon — tan Oy, = ~ 1.235 = 0.046 Byy = 2.65° (Ans.) we IB ag me f ~ tm a= ee = 1.745 = 1.096 Ba, = 47.64° (Ans.) Specific work From Eq. (11.103), the specific work is constant w=2 ob w= 2 x 1885 x 15.43 x 10° ki/kg w = 58.17 ki/kg (dns.) Loading coefficients w _ 58170 = = Se = 0.727 (Ans, Ye up 282.75? (ns) 58170 = 22) = 0.505 (Ans. Yin 3303? (Ans.) 58170 y= 221M. = 0372 (4 M™ 3osgs? an Degrees of reaction Since the axial velocities at the rotor entry and exit are not the same (except at the mean section), Eqs, (I1.44c) and (11.106) are not valid here. Therefore, the following equation is used: I Ry= 5,5 (un S00” By ay see? ry) [Link]Axial Compressor Stages 513 _ 1925? sec? 48.28 - 220.65" sec? 2.65 258170 29.98% (Ans.) Fo Cy 900% By ~ day see Ba) Ry x 100 P a = 163? sec? 6097 =1393? see? 47.64 19 2x 58170 R, = 12.63% (Ans.) The results are summarized in the following table Hub 48.28 265 29.98 = S817 > 0.727 Mean 54.82 25.0 50.0 58.17 0.530 (0.505 Tip 60.97 47.64 1263 SIT 0372 11.7 The design point data for an axial compressor stage is the same as the mean section data in Ex. 11.1, ie. c, = 180 m/s, u = 339.3 mss, o = B, = 25° Calculate the design point flow and loading coefficients. From these, compute the loading coefficients at ¢ = 0.2, 0.4, 0.6 and 0.8. Solution: tan B, + tan a = 2 tan 25 = 0.933 yf =~ 9* (an By + tan &) yr = 1 - 0.9339" yt = 1 - 0.933 x 0.53 y* = 0.505 (dns.) ‘At off-design points, the loading coefficients are calculated from y= 1-0.933 6514 IL 11.2 4.3 114 us Turbines, Compressors and Fans > Questions and Problems (a) Draw a sketch of the two-stage axial flow compressor with inlet guide vanes. () Draw curves indicating the variation of static pressure, temperature and absolute velocity through this compressor. (©) Why does a compressor stage have a lower efficiency and loading factor compared to an equivalent turbine stage? Draw velocity triangles at the entry and exit for the following axial compressor stages: (a) R= } ()R< 3 () R> ; @R=1()R>1(f)R= negative (a) Why is it necessary to employ multi-stage axial compressors to obtain moderate to high pressure ratios? (6) What are the principal distinguishing features of the low pressure and high pressure stages from aerothermodynamic and material considerations? Derive the following relations for an axial compressor stage with constant axial velocity. (@) tan oy + tan B, = tan ay + tan B= e ©) w= 9 (tan B,~tan B,) ©) GP = 6 an a, ~tan c4) @) my = Op),/2 pu e, (tan a ~ tan 04) Draw the /-s diagram for a complete axial-flow compressor stage with R > } Prove the following relations: (@) R= 4 (an p,+tan By = F U~ p (tan of, ~ tan stage lo: a ao (b) Yp sec? a + Yq sec? B, p Sec" oy + Ey sec” By tan B, tan B, Om=1-b6 State the assumptions used. [Link]11.6 11.7 11.8 1.9 11.10 UAL 11.12 Axial Compressor Stages 515 (a) What is the work-done factor for an axial compressor stage? Why is it not employed for turbine stages? (b) How does it vary with the number of stages? (o) Show the axial velocity profiles along the blade height in the first and eighth stage, (a) Describe four schemes of obtaining supersonic compression in an axial compressor stage. (b) What are the advantages and stages? (c) What is a transonic compressor stage? What is surging in axial-flow compressors? What are its effects? Describe briefly. (a) What is stalling in an axial compressor stage? How is it developed? (b) What is rotating stall? Explain briefly the development of small and large stall cells in an axial compressor stage. ‘An axial compressor stage has a mean diameter of 60 cm and runs at 15000 rpm. If the actual temperature rise and pressure ratio developed are 30°C and 1.4 respectively, determine: sadvantages of supersonic (a) the power required to drive the compressor while delivering 57 kgis of air; assume mechanical efficiency of 86.0% and an initial temperature of 35°C, (b) the stage loading coefficient, (c) the stage efficiency, and (@) the degree of reaction if the temperature at the rotor exit is 55°C. Answer: (a) 1998.5 kW (b) 0.135 (c) 94.19% (d) 66.6% If the loss coefficients of the stage in Ex. 11.2 ( = 0.7365, By = 51°, By — 9%; = 7%, Op = 50.18%) are ¥p = 0.07, Yq = 0.078, determine the efficiencies of the diffuser and rotor blade rows and the stage. Answer: Np = 88.0%, Ne = 86.8%, My, = 87.43% the same data as in Ex. 1.1: ‘An axial compressor stage has ry = 15 em, u, = 282.73 mis = 18 cm, ty, = 339.3 m/s 95.85 m/s Tn r, = 21 om, 1, [Link]516 Turbines, Compressors and Fans Con = Com = Cram = 180 mis Ry, = 50% y Cay = By = 54.82%, Cy = Boy = 25° Determine rotor blade air angles, the degree of reaction, specific work, flow coefficient and loading coefficient at the hub, mean and tip sections for constant reaction. Answer: Forced vortex stage (constant reaction) Hub 48.0 20.08 50.0 40376 ~=—0.677_—0.505 Mean 54.82 25.0 50.0 58.17 0.530 0,505 Tip 60.97 3054 50.0 79.17 0417 0.505 11.13 Compute the loading coefficients for the stage in Problem 11.12 at 9= 0.2, 0.4, 0.6 and 0.8. Indicate the design point values of 6 and Y. Plot the ¢~ w curve. aw 06 0.7365 08 ¥ 0.944 0.887 08st 0.793 075Chapter Centrifugal Compressor Stage ill now only axial-flow machines have been discussed. This chapter deals with an energy absorbing and pressure producing machine of the outward flow radial type—the centrifugal compressor*” 4°. As will be seen in the various sections of this chapter the geometrical con- figuration of the flow and the passages is radically different from those in the axial type. ‘A centrifugal compressor like a pump is a head or pressure producing device. The contribution of the centrifugal energy in the total change in the energy level is significant. Section 1.10 highlights some of the special features of radial machines. From discussions given in Chapters 1 and 7 it is amply clear that a centrifugal type of compressor is suitable for low specific speed, higher pressure ratio and lower mass flow applications. Performance-wise, the centrifugal compressor is less efficient (3-5%) than the axial type. However, a much higher pressure ratio“*°“*? (~ 4.0) per stage, single-piece impeller and a wider range of stable operation are some of the attractive aspects of this type. Besides the evolution of a perfect centrifugal pump, the developments of early supercharged aircraft reciprocating engines and later that of high output large diesel engines gave a great impetus to the development, of centrifugal air compressors. These are used in large refrigeration units“, petrochemical plants and a large variety of other industrial applications. In aircraft applications“*** it is only used for small turbo-prop engines. For large turbo-jet engines, the large frontal area resulting from its application outweighs its advantages. ‘While the design and performance characteristics of axial compressors have been widely studied and supported by a huge mass of data acquired from cascade tests, nothing of this order is available for radial flow machines, particularly centrifugal compressors. Some methods and test facility for testing radial diffusing cascades have been described in sec. 8.7. [Link]518 Turbines, Compressors and Fans > 12.1 Elements of a Centrifugal Compressor Stage Figures 12.1 and 12.2 show the principal elements of a centrifugal compressor stage. Casing Shroud Impeller Impeller Accelerating Inducer nozzle IGV section Fig. 12.1 Elements of a centrifugal compressor stage The flow enters a three-dimensional impeller through an accelerating nozzle and a row of inlet guide vanes (IGVs). The inlet nozzle accelerates the flow from its initial conditions (at station i) to the entry of the inlet guide vanes. The IGVs direct the flow in the desired direction at the entry (station 1) of the impeller. The impeller through its blades transfers the shaft work to the fluid and increases its energy level. It can be made in one piece consisting of both the inducer section and a largely radial portion. The inducer receives the flow between the hub and tip diameters (d,, d,) of the impeller eye and Passes it on to the radial portion of the impeller blades. The flow approaching the impeller may be with or without swirl. The inducer section can be looked upon as an axial compressor rotor placed upstream of the radial impeller. In some designs this is made separately and then mounted on the shaft along with the radial impeller. [Link]Centrifugal Compressor Stage 519 Ina great majority of centrifugal compressors the impeller has straight radial blades after the inducer section. At high speeds, the impeller blades are subjected to high stresses which tend to straighten a curved impeller blade. Therefore, the choice of radial straight blades is more sound for higher peripheral speeds. However, in fan and blower applications (Chapter 15), on account of the relatively lower speeds, backward and forward-swept impeller blades are also used. Unlike axial machines, the hub diameter of the radial impellers varies from the entry to the exit. The tips of the blades can be shrouded to prevent leakage, but manufacturing and other problems of the shrouded impellers have kept them open in most applications. The impeller discharges the flow to the diffuser through a vaneless space (Fig. 12.2). Here the static pressure of the fluid rises further on account of the deceleration of the flow. The diffuser may be merely a vaneless space or may consist of a blade ring as shown in Fig. 12.2. For high performance, the design of the diffuser is as important as that of the impeller. Flow blades Vaneless space Impaler blades Fig. 12.2 A centrifugal compressor stage The flow at the periphery of the diffuser is collected by a spiral casing known as the volute which discharges it through the delivery pipe. Figure 12.3 shows a centrifugal impeller with blades located only in the radial section between diameters d, and d>. To prevent high diffusion rate of the flow, the impeller blades are invariably narrower at a larger diameter (by 12.2 Stage Velocity Triangles The notation used here corresponds to the r, 0 and x coordinate system. As per the convention for radial machines, the angles are measured from the tangential direction at a given point. The absolute and relative air angles at the entry and exit of the impeller are dendted by a, 0 and B,, B, respectively. Since the change in radius between the entry and exit of the impeller is large, unlike in axial machines, the tangential velocities at these stations are different: _ aaN 4 60 Entry velocity triangle Figure 12.4 shows the flow at the entry of the inducer section of the impeller without IGVs. The absolute velocity (c,) of the flow is axial (cy = 90°) and the relative velocity (w,) is at an angle B, from the tangential direction, Thus the swirl or whirl component cq = 0. (12.1)Centrifugal Compressor Stage » 521 Flow Inducer section of the impeller oo =O Fig. 12.4 Flow through the inducer section without inlet guide vanes Figure 12.5 shows the flow through axially straight inducer blades in the presence of IGVs, The air angle (or) at the exit of the IGVs is such that it gives the direction of the relative velocity vector (w,) as axial, i B, = 90°. This configuration offers some manufacturing and aerodynamic advantages, viz., (i) centrifugal impellers with straight blades are much easier and cheaper to manufacture and (ii) the relative velocity (w,) approaching the impeller is considerably reduced. In this case B, = 90° and the positive swirl component is co = (12.2) (2.3) Inlot guide vanes Entry Fig. 12.5 Flow through the inducer section with inlet guide vanes Figure 12.6 shows the entry and exit velocity triangles for impeller blades located only in the radial section. For the sake of generality, the [Link]522 Turbines, Compressors and Fans Fig. 12.6 Entry and exit velocity triangles for impeller blades only in the radial section, backward ‘swept blades, Bo < 90° absolute velocity vector c, is shown to have a swirl component cy). However, if there are no guide vanes, c, will be radial (c, = c,1) and a = 90°, ¢g = 0. This particular condition is expressed by “zero whirl or swirl” at the entry and would be assumed in this chapter unless mentioned otherwise. Exit velocity triangle The impeller blades shown in Fig. 12.6 are backward swept, i.c., B < 90°. The exit velocity triangle for these blades is shown in the figure. The flow leaves the blades at a relative velocity w, and an air angle B,. The absolute velocity of flow leaving the impeller is c, at an air angle a. Its tangential (swirl or whirl) component is cg, and the radial component c,». The following relations are obtained from the velocity triangles at the entry and exit shown in Fig. 12.6: Gn =e, Sit a = wy sin B, (12.4) €a1 = ¢1 COs @% = C, cot O = my ~ cy cot By (12.5) 2 = Cy sin @ = wy sin B, (12.6) Cen = 2 608 Oy = ep COL ~ uy — Gy cot By (12.7) Figure 12.7 shows the velocity triangles at the entry and exit of a radial-tipped impeller with blades extending into the inducer section. The velocity triangle at the entry is’similar to that in Fig. 12.6; here ¢,, replaces the velocity component ¢,1. [Link]Centrifugal Compressor Stage 523 Impelier blade ring Fig. 12.7 Entry and exit velocity triangles for impeller with inducer blades, radial-tipped blades, 6, = 90° ‘The exit velocity triangle here is only a special case of the triangle in Fig. 12.6 with 8, = 90°. This condition when applied in Eqs. (12.6) and (12.7) gives ca sin (12.8) Cay = Cy COS Oy = yy COL Oh = ty (12.9) ‘The mass-flow rate from the continuity equation for Figs. 12.3 and 12.6 can be written as ti = pyc db = P2e,pMdyby (12.10) This for Figs. (12.1) and (12.7) is : ti = Drea % di = Pata, (12.1) s the velocity triangles for forward-swept blades (B, rf at the entry. It may be observed that such blades have large fluid deflection and give cg) > uz. This increases the work capacity of the impeller and the pressure rise across it. This configuration is unsuitable for higher speeds in compressor practice and leads to higher losses. However, for fan applications such blades are used in multivane or drum-type centrifugal blowers (Sec. 15.4). [Link]impeller blade ring Fig. 12.8 Entry and exit velocity triangles for forward swept blades (B, > 90°) with zero swirl at entry 12.2.1 Stage Work In a centrifugal compressor the peripheral velocities at the impeller entry and exit are w, and 1, respectively. Therefore, the specific work or the energy transfer is W = xem ~ U1Co, (12.12) In this equation, if cg, is positive (Figs. 12.5, 12.6 and 12.7), the term 4;Co, is subtractive, Therefore, the work and pressure rise in the stage are relatively lower, These quantities are increased by reducing cg, to zero (Fig. 12.4) or making it negative. In the absence of inlet guide vanes, cy = 0. This condition will be assumed throughout in this chapter unless mentioned otherwise. Therefore, Eq. (12.12) gives W = uxe—p (12.13) Substituting from Eq. (12.7) W= a (tp — 3 ot B.) The flow coefficient at the impeller exit is defined as = a= 2 (12.14) Therefore, w =u} (1 ~ @, cot ,) (12.15) [Link]Centrifugal Compressor Stage 525 If ¢, and f, are the actual values, the work given by Eq. (12.15) is the actual work in the stage. The work is also given by the following form of Euler’s equation: wed Gat 5 w-wh 5 gw) 012.16) For a radial-tipped impeller with zero swirl (whirl) at the entry a% = 90°, B, = 90° and Eqs. (12.13) and (12.15) reduce to Tes (12.17) 12.2.2 Pressure Coefficient The head, pressure or loading coefficient is defined in Sec. 7.4.1. As in earlier chapters, here also it is defined by yo (12.18) uy This gives, in a dimensionless form, a measure of the pressure raising capacities of various types of centrifugal compressor impellers of diffe- rent sizes running at different speeds, Equations (12.13) and (12.15) give = £02 Vaan: (12.19a) V=1- @ cot B, (12.19) This expression gives the theoretical performance characteristics of impellers of different geometries. It may be noted that Eq. (12.196) has been derived assuming zero entry swirl and no slip. Figure 12.9 shows the @— y plots for forward-swept, radial and backward-swept impeller blades. The actual characteristics will be obtained by accounting for stage losses. The backward-swept and radial blade impellers give stable characteri- sties. The forward-swept type gives unstable flow conditions on account of the rising characteristic as explained in Sec. 11.8.2 (Fig. 11.18). Equations (12.19) for radial-tipped blades give ye 12.2.3. Stage Pressure Rise The static pressure rise in a centrifugal compressor stage occurs in the impeller, diffuser and the volute. The transfer of energy by the impeller takes place along with the energy transformation process. The pressure rise across the impeller is due to both the diffusion of the relative velocity vector w; to w, and the change in the centrifugal energy (see Sec. 6.9.2) [Link]526 Turbines, Compressors and Fans 14 90) 2 copa Z — Radial (8, = 90°) § 1.0- ———_________+ 2 Bos Sea 3 * 90%) g é a 0 02 0406 08 4012 Flow coetficient, 6 Fig. 12.9 Performance characteristics of different types of centri- fugal impellers (cy, = 0, 4 = 1) The static pressure rise across the diffuser and volute (if any) occurs simply due to the energy transformation processes accompanied by a significant deceleration of the flow. The initial kinetic energy (at the entry of the diffuser) is supplied by the impeller. In this section the pressure rise (or pressure ratio) across the stage is first determined for an isentropic process. For small values of the stage pressure rise (as in axial stages and cent- rifugal fans), the flow can be assumed to be incompressible. Therefore, 7p BPa ~ Ao = w= ud (=o cot fy Apo = put (1 ~ ¢ cot fa) (12.20) Substituting from Eq. (12.19b) Ano =p wag (12.21) However, the pressure rise in a centrifugal compressor stage is high and the change in the density of the fluid across the stage is considerable. Therefore, in most applications, the flow is not incompressible. The pressure ratio for compressible flow is obtained by the following method: The fluid is assumed to be a perfect gas. Therefore, w = Aly = Cy (Tons — Tor) [Link]Centrifugal Compressor Stage 527 L w= opMy 1) yo w= 6,To (Po? 1) (12.22) Substituting from Eq. (12.15) ¢pTor (Pro F ~ 1) = 15 (I~ $ cot By) This on rearrangement yields jo = 2 = 11+(1~¢, cot By) ui (12.23) 0 Poi ° ep To : 2 Pwo = (+ vd) (12.24) Lo > 12.3 Enthalpy-entropy Diagram Figure 12.10 shows an enthalpy-entropy diagram for a centrifugal comp- ressor stage (Figs. 12.1 and 12.2). Flow process occuring in the ace- elerating nozzle (i-1), impeller (1-2), diffuser (2-3) and the volute (3-4) are depicted with values of static and stagnation pressures and enthalpies. ‘The flow, both in the inlet nozzle and guide vanes is accelerating from static pressure p, On account of the losses and increase in the entropy the stagnation pressure loss is py; — po, but the stagnation enthalpy remains constant: oy = hon (12.25a) nt hdamrdd (12.25b) The isentropic compression is represented by the process 1-2s-4ss. This process does not suffer any stagnation pressure loss: Pons = Posss = Posse (12.26) The stagnation enthalpy remains constant. ons = hoses = hosss (12.27) The energy transfer (and transformation) occurs only in the impeller blade passages. The actual (imeversible adiabatic) process is represented by 1-2. The stagnation enthalpies in the relative system at the impeller entry and exit are Fats (02.28) Foret = 17.29528 Turbines, Compressors and Fans Enthalpy Entropy Fig. 12.10 Enthalpy-entropy diagram for flow through a centtri- fugal compressor stage ‘The corresponding stagnation pressures are pp). and Poet Static pressure rise in the diffuser and the the volute occurs during the processes 2-3 and 3-4 respectively. The stagnation enthalpy remains constant from station 2 to 4 but the stagnation pressure decreases progressively. ign = hos = hag (12.30) Poo > Pos > Pos (12.31) [Link]__ Centrifugal Compressor Stage 529 ‘The actual energy transfer (work) appears as the chage in the stagnation enthalpy. Therefore, ftom Eq. (12.16) n= G- A+ 5 oF- w+ 5 0d Wy = hoy This on rearrangement gives Jy hy + + (3 - wi) - ; @G- 1) =0 1 1 1 L (u +3») a u3= (+407) 7 ue (12.32a) eee ee A (12.326) art — M2 Moa 9 This relation is also shown on the /-s diagram (Fig. 12.10). 12.3.1 Stage Efficiency The actual work input to the stage is Wy = fog — hor = 18 (1 - cot By) (12.33a) For a perfect gas, Wa = Cy (Toa — Tox) = 3 (1 ~ $2 cot B:) (12.33b) The ideal work between the same static pressures p, and p, is We = Rosse ~ hor = Gp (Toss ~ Tor) (12.34a) Ta mn opty {ite} no { To = Wy = Cp Tor (pw? — 1} (12.34b) Here the stagnation pressure ratio = Pus Pos (12.35) 7. The last relation in Eq, (12.35) is valid for incompressible flow assuming s Cass The ideal and actual values of the stage work are shown in Fig. 12.10. The total-to-total efficiency of the stage can now be defined by = We Posse ~ hor te ere (12.36a) = £r loss Ta) = (12.36b) «5 (1-92 cot By) [Link]530 Turbines, Compressors and Fans feemeee rl — Se Tn Po % ~) u5 (1-9, cot B,) This equation yields the pressure ratio of the stage for the given initial state of the gas and values of >, and B,. (12.36e) wh Paes Pro= 71+ My (I~ po} (12.37) cp Tox This is similar to Eq, (12.23) for n,, = 1 (reversible stage). 12.3.2 Degree of Reaction A large proportion of energy in the gas at the impeller exit is in the form of kinetic energy. This is converted into static pressure rise by the energy transformation process in the diffuser and volute casing. The division of static pressure rise in the stage between the impeller and the stationary diffusing passages is determined by the degree of reaction. This can be defined either in terms of pressure changes or enthalpy changes in the impeller and the stationary diffusing passages. The discussions given in Secs. 9.5.2 and 11.2.2 explain various methods of defining the degree of reaction. Expressions for the degree of reaction in this section are derived from the following definition. change in static enthalpy in the impeller change in stagnation enthalpy in the stage hy ~hy R= a 12.38) Fon — hot a From Eq. (12.32a) gy 5 WB=wD+ 3 hud) (12.39) For zero swirl at the entry (cg, = 0) Ion — hor = t3Cep (12.40) Therefore, Eqs. (12.39) and (12.40) when put into Eq, (12.38) give eae, p~ Whew) + On? 12) oy 2utyeo For the constant radial velocity component. Cy = Oy = Oy With inducer blades and zero entry swirl (Fig. 12.4), 61> Oy = ep [Link]_ Centrifugal Compressor Stage 531 With these conditions, the following expressions are obtained from the entry and exit velocity triangles: wha uh y= ey 12.42) Bea Wz = Crp + (ty — Ce) fy + Wh — Wuycm + chy 1h - Wh = Wc - Ay - A, (12.43) Equations (12.42) and (12.43), when used in Eq. (12.41), give (12.43a) Substituting from Eq, (12.7) and rearranging eiyl 2 R=5+ 5 Hoth (12.43) Equation (12.436) is plotted in Fig. 12.11. The degree of reaction of the radial-tipped impeller (, = 90°) remains constant at all values of the flow B= 30° 0.85 a 3 3 8 Degree of reaction, R 0.55 0.45 (er hat ey 0 Ua) Flow coefficient, ¢ Fig. 12.11 Variation of degree of reaction with flow coefficient for various values of impeller exit air angle [Link]532 Turbines, Compressors and Fans coefficient. Reaction increases with flow coefficient for backward-swept impeller blades (B, < 90°) and decreases for forward swept type (B, > 90°) as shown, From Eg, (12.19a) (12.44a) (12.44b) Equation (12.44) shows that the higher the degree of reaction, the lower is the stage pressure coefficient and vice versa. This is depicted in Fig. 12.12. The backward-swept impeller blades give a higher degree of reaction and a lower pressure coefficient compared to the radial and forward-swept blades. 16 14 Pressure coefficient, y & a re 02 04 06 08 10 Degree of reaction, R Fig. 12.12 Variation of pressure coefficient with degree of reaction > 12.4 Nature of Impeller Flow The flow pattern in the three-dimensional flow passage of the impeller of a centrifugal compressor is very complex. Various coordinate systems have been used to describe the flow field in such passages. Section 6.4 describes a natural coordinate system (Fig. 6.6). To simplify the understanding of the flow in a radial turbo-machine the flow field can be separately considered in the radial-axial (meridional) plane (Fig. 12.13) and the vane-to-vane plane (Fig, 12.14). Further [Link]Centrifugal Compressor Stage 533 simplification in the theoretical analysis of such a flow is obtained by assuming it to be inviscid. 12.4.1 Flow in the Meridional Plane ‘An infinitesimal fluid element (at radius r) in the meridional plane between the hub and the shroud is shown in Fig. 12.13. The meridional streamline passing at the centre of the element has a radius of curvature R. The meridional velocity is c,, and the velocity component in the tangential direction cp ssn Infritesimal 7 Merigona fluid element 4 sreamtine Z (a) (b) Fig. 12.13 Flow in the meridional plane It is very convenient to study such a flow in the natural coordinate system. An expression for the meridional velocity distribution in the normal direction (n ~ direction) is derived here under the following assu- mptions: 1. isentropic and incompressible flow 2. axisymmetric flow 3. radial blades. The thickness of the element (normal to the paper) is unity. Therefore, its volume is ds dn.534 Turbines, Compressors and Fans The fluid element is subjected to the centrifugal forces both due to the impeller rotation and the curvature of the meridional streamline as shown in Fig. 12.13). The centrifugal force due to the tangential velocity component cy is 2 ¢ ds dn pds dn and that due to the curvature of the streamline is om pds dn & Equating the forces acting on the element in the normal direction 2 : sdn © = {p42 ; Sn pds + pds dn — cos & (0+ 32am) as pavdn & This on rearrangement gives 1a _@ Sn pan 008 6- R (12.45) From Fig. 12.13(a) or Gr = 008 5 (12.46) For axisymmetric flow and radial blades, e=u=or (12.47) em = Equations (12.46) and (12.47), when applied in Eq. (12.45), yield 19 _ op, Or Ww pan OTS, (12.48) Equation (12.32a) gives the general relation 1 LoL ats al w = const. Differentiating and rearranging dh =u du—w dw (12.49) For isentropic flow da = “. and u du = or dr Therefore, dividing throughout by dn, Sq. (12.49) can be rewritten as Lap _ 9 ar aw pan RS, (12.50) [Link]Centrifugal Compressor Stage 535 Combining Eqs. (12.48) and (12.50), dv dn ak (12.51) This on integration gives In w + In (const) = jf Ps wc, = ke ® (12.52) Equation (12.52) gives the velocity distribution in the meridional plane. The value of the constant k can be determined from the continuity equation. The mass-flow rate through the infinitesimal stream tube of cross- sectional area dn x 1 is rit = p by One dn) Substituting from Bq, (12.52) dth =2 prr dn kexp j¢ The total mass-flow rate is obtained by integrating this from hub to the shroud. mm = 2nk J {pref} dn (12.53) 12.4.2 Fiow in the Vane-to-Vane Plane Figure 12.14 shows an infinitesimally thin slice of the flow between the two backward-swept blades of an impeller. An element of the flow between two streamlines dm apart is subtended by an angle d@. The relative velocity on one side of the element is w and on the other side ane om om The circulation around (anti-clockwise) this element is a (w+ al an) (R+ dm) dO— w Rad a Neglecting the product of two small quantities and substituting dA =R dO dm ar _ aw, w dd am R [Link]536 Turbines, Compressors and Fans Fig. 12.14 Flow in the vane-to-vane plane ‘This is the vorticity in the vane-to-vane plane (see Sec. 6.3.3). w R Helmholtz law states that the change in the absolute vorticity of an inviscid fluid with time is zero. In the present case the fluid is assumed to enter the impeller passage without any vorticity. Therefore, if the absolute vorticity in the impeller é= » + (12.54) passage (rotating with an angular velocity @) is to be zero, the flow inside it must have a rotation of — @. However, Rotation = ; vorticity a-le (12.55) 2 s » Equations (12.54) and (12.55) give aw sw he oe =2@ (12.56) Equation (12.56) gives the velocity distribution in the vane-to-vane plane. This rotational flow in the relative system is referred to as “relative eddy”. It affects the energy transfer in the impeller and hence the pressure ratio developed as discussed in the following section. [Link]Centrifugal Compressor Stage 537 > 12.5 Slip Factor! 4, 400 ‘The actual velocity profiles at the impeller exit due to real flow behaviour are shown in Figs. 12.15 and 12.16. The energy transfer occurring in the Exit velocity profile Hub Meridional plane Fig. 12.15 Meridional velocity distribution at the impeller exit Exit velocity profile Vane to vane plane Fig. 12.16 Vane-to-vane velocity distribution at the impeller exit [Link]538 Turbines, Compressors and Fans impeller corresponding to these velocity profiles is less than the one that would have been obtained with one-dimensional flow. The relative eddy mentioned earlier causes the flow in the impeller passages to deviate (Fig. 12.17) from the blade angle (3) at the exit to an angle the difference being larger for a larger blade pitch or smaller number of impeller blades. Fig. 12.17 Exit velocity triangles with and without slip On account of the aforementioned effects, the apex of the actual velocity triangle at the impeller exit is shifted away (opposite to the direction of rotation) from the apex of the ideal velocity triangle as shown in Fig. 12.17. This phenomenon is known as slip and the shift of the apex is the slip velocity (c,). It may be seen that, on account of the slip, the whirl component is reduced which in tum decreases the energy transfer and the pressure developed. The ratio of the actual and ideal values of the whirl components at the exit is known as slip factor (42) = (12.57) ez Therefore, the slip velocity is given by C= Ca Oy = (1-H) om (12.58) [Link]Cenerifugl Compressor Stage 539 The expressions for the actual work done, pressure ratio and stage efficiency can now be rewritten with the slip factor. From Eqs. (12.13) and (12.15) W = Ml ty Cy = Hu} (1 — @ cot fy) (12.59) Similarly, Eqs. (12.36c) and (12.37) are modified to a pT (Pro 7 — DY Hu; (1-9, cot B,) It (12.60) A Po* fine (-¢, cop tet} (12.61) Tor The methods of determining slip factors have been suggested by various investigators. Some of them are described here briefly. 12.5.1 Stodola’s Theory Figure 12.18 depicts the model of flow with slip as suggested by Stodola”, The relative eddy is assumed to fill the entire exit section of the impeller passage. It is considered equivalent to the rotation of a cylinder of diameter d = 2r at an angular velocity @ which is equal and opposite to that of the impelier (Sec. 12.4.2) as shown in the figure. The diameter, and hence, the tangential velocity of the cylinder, is approximately determined as follows: $= 2a lz Fig. 12.18 Stodola’s model of flow with slip [Link]540 Turbines, Compressors and Fans The blade pitch at the outer radius (r,) of the impeller with z blades is 2nr, : The diameter of the cylinder is 2r=ssin By = aa sin B, (12.62) The slip velocity is assumed to be due to rotation of the cylinder. Therefore, = or Substituting for r from Eq. (12.62) = 0 = sin B, (12.63) However, 1; = ar, Therefore, 7 uy sin B, (12.64) Equation (12.64) when put in Eq. (12.58) gives (= #) ee = © wy sin 8, wa1- Substituting from Eq. (12.7) He (12.65) For a radial-tipped blade impeller ( = 90°) wal-Z (12.66) The above expressions for slip show that for a given geometry of flow the slip factor increases with the number of impeller blades. Along with this the fact that the number of impeller blades is one of the governing parameters for losses should not be lost sight of. 12.5.2 Stanitz’s Method A method based on the solution of potential flow in the impeller passages is suggested by Stanitz*” for B, = 45° - 90°. The slip velociy is found to be independent of the blade exit angle and the compressibility. This is given by (12.67)Centrifugal Compressor Stage 541 198 0-0 co = Sw, _ 198 z Con 198 eT 12.68 BOT” 3-6; cot Ba) ae For B= 90° u-1- 18 (12.69) Equations (12.66) and (12.69) are of identical form. 12.5.3. Balje’s Formula Balje suggests an approximate formula for radial-tipped (B, = 90°) blade impellers: -1 {i + 3, (12.70) impeller tip diameter eye tip diameter > 12.6 Diffuser The static pressure of the gas at the impeller exit is further raised by passing it through a diffuser located around the impeller periphery. The absolute velocity (¢,) of the gas at the impeller exit is high which is reduced to a lower velocity (c3) in the diffuser as shown in the enthalpy- entropy diagram (Fig. 12.10). The amount of deceleration and the static pressure rise (p; — p») in the diffuser depend on the degree of reaction and the efficiency of the diffusion process. An efficient diffuser must have minimum losses (Pq) ~ Pox)» Maximum efficiency and maximum reco- very coefficient. Expressions for the efficiency and pressure recovery coefficient have been derived in Sec. 2.4. A facility for testing the performance of a dec- elerating radial cascade (radial vaned diffuser) is described in Sec. 8.7.1. Diffusers in centrifugal compressors are either of the vaneless or vaned type 12.6.1 Vaneless Diffuser ‘As the name indicates, the gas in a vaneless diffuser is diffused in the vaneless space around the impeller before it leaves the stage through a volute casing. In some applications the volute casing is omitted. [Link]542 Turbines, Compressors and Fans The gas in the vaneless diffuser gains static pressure rise simply due to the diffusion process from a smaller diameter (d;) to a larger diameter (d,). The corresponding areas of cross-sections in the radial direction are Ay = ndyb, = 2myby (12.71a) Ay = ndsb; = 2mrsb, (12.716) Such a flow in the vaneless space is a free-vortex flow in which the angular momentum remains constant. This condition gives T2€02 = "303 (12.72) The continuity equation at the entry and exit sections of the vaneless diffuser gives Prtrada = Pstrsts Presa 2tryby) = pseys (2rb3) Pate ;2by = Parse,sbs (12.73a) For a small pressure rise across the diffuser, p; ~ p,. Therefore, FyCpgb2 = P3C,ab3 (12.73b) For a constant width {parallel wall) diffuser by = by N62 = G3 (12.730) The absolute velocity at the diffuser exit is given by ‘ A= yt c= (2) (+ &)= (12.74) 4 Equations (12.72), (12.73c) and (12.74) yield os co% a2 or ec % (12.75) This relation further gives % = 05 = tan! £2 = (12.76) or It should be remembered that this equation is valid only for incompressi- ble flow through a constant width diffuser. Equation (12.75) clearly shows that the diffusion is directly proport- ional to the diameter ratio (d,/d,). This leads to a relatively largesized diffuser which is a serious disadvantage of the vaneless type. In some cases the overall diameter of the compressor may be impractically large. This is a serious limitation which prohibits the use of vaneless diffusers in aeronautical applications. Besides this the vaneless diffuser has a lower efficiency and can be used only for a small pressure rise. [Link]_ Centrifugal Compressor Stage 543 However, for industrial applications, where large-sized compressors are acceptable, the vaneless diffuser is economical and provides a wider range of operation. Besides this, it does not suffer from blade stalling and shock waves. 12.6.2 Vaned Diffuser For a higher pressure ratio across the radial diffuser, the diffusion process has to be achieved across a relatively shorter radial distance. This requires the application of vanes which provide greater guidance to the flow in the diffusing passages. Diffuser blade rings can be fabricated from sheet metal or cast in cambered and uncambered shapes of uniform thickness (Figs. 12.19 and 12.20). Figure 12.21 shows a diffuser ring made up of cambered aerofoil blades. To avoid separation of flow, the divergence of the diffuser blade passages in the vaned diffuser ring can be kept small by employing a large number of vanes. However, this can lead to higher friction losses. Thus an optimum number of diffuser vanes must be employed. The divergence of the flow passages must not exceed 12 degrees. The flow leaving the impeller has jets and wakes. When such a flow enters a large number of diffuser passages, the quality of flow entering different diffuser blade passages differs widely and some of the blades may experience flow separation, leading to rotating stall and poor performance. To avoid such a possibility, it is safer to provide a smaller number of diffuser blades than that of the impeller. In some designs the number of diffuser blades is kept one-third of the number of impeller blades, This arrangement provides a diffuser passage with flows from a number of impeller blade channels. Thus the nature of flow entering various diffuser passages does not differ significantly. ‘Another method to prevent steep velocity gradients at the diffuser entry is to provide a small (0,05 d, — 0.1 d,) vaneless space between the impeller exit and the diffuser entry as shown in Figs. 12.2 and 12.22. This allows the non-uniform impeller flow to mix out and enter the diffuser with less steep velocity profiles. Besides this the absolute velocity (Mach number) of the flow is reduced at the diffuser entry. This is a great advantage, specially if the absolute Mach number at the impeller exit is greater than unity. The supersonic flow at the impeller exit is decelerated in this vaneless space at constant angular momentum without shock. Every diffuser blade ring is designed for given flow conditions at the entry at which optimum performance is obtained. Therefore, at off-design operations the diffuser will give poor performance on account of, mismatching of the flow. In this respect a vaneless diffuser or a vaned544 Turbines, Compressors and Fans Flow Fig. 12.19 Diffuser ring with cambered blades Fig. 12.21 Diffuser ring with cambered aerofoil blades diffuser with aerofoil blades (Fig. 12.21) is better. For some applications it is possible to provide movable diffuser blades whose directions can be adjusted to suit the changed conditions at the entry, In some designs for industrial applications, a vaneless diffuser supplies the air or gas direct to the scro!l casing, whereas for aeronautical applications, various sectors of the vaned diffuser are connected to separate combustion chambers placed around the main shaft,_ Contig Compresor Stage 545, 12.6.3 Area Ratio aneless diverging wall diffuser. The side walls have 26. The area ratio of such a diffuser in the radial Figure 12.22 shows a vi a divergence angle of direction is (12.77) a) Impeller i 2 Fig. 12.22 Radial diffuser passage with diverging walls ‘The semi-divergence angle is given by byxbs 2(n-%) by — by = 2(ry > M2) tan @= (dy — d,) tan 0 tan 0= Now et | pes (12.78) This when used in Eq, (12.77) gives -4 f 47 G ie (12.79) [Link]546. Turbines, Compressors and Fans For parallel walls (tan 8 = 0), this gives A, = ds/dy (12.79b) Ifthe diverging passage of Fig, 12.22 is fitted with straight flat blades ig. 12.20), the area ratio normal to the direction of flow is further increased, From Fig. 12.20, hh = 2 sin0-40,) ~ sin(@0—a,) _ 608 a, Se eieaat (12.80) The area ratio is given by addy sino, dyby = 005" a Mdyb, sind, ~ doby Singey Substituting from Eq, (12.80) f Ars Sis pi 5 a (12.81) 2b, Sina Y (ds/dy) Substituting further from Eq. (12.78) ate {+(4-1) tan 8 — fl- 5% (19 99) aod, l (a, J b,/d; J sin ery | (dy!d,) " Fauation (12.82) shows that the area ratio ofa diffuser can be increased by (@) increasing the diameter ratio, dy/dyy (6) increasing the width ratio, byb,, (©) decreasing the leading edge vane angle, ay (4) various combinations of a, b and c. Some typical values of these parameters are: dydy = 1.4 0 1.8, 4, = As/Ay = 2.5 to 3.0 & = 10 to 20° dyldy = 0.025 — 0.10 Snax = 5° Figure 12.23 shows the plots of the area ratio against the diameter ratio for some diffuser configurations. It may be observed that, for a given [Link]ee Centrifugal Compressor Stage 547 diameter ratio, very large values of the area ratio can be obtained by employing vaned diffusers with diverging walls. 8 Vaned diverging a= 15°, 0= 3° 7 Vaned parallel 6 a= 15° 5 $4 < 3 Vanless 2 diverging o=3° 5 10 12 74 16 18 20 Zoe Diameter ratio, d3/dy Fig. 12.23 Variation of area ratio in radial diffusers with diameter ratio 12.6.4 Mach Number at Diffuser Entry In the absence of the vaneless space between the impeller tip and the diffuser entry, the Mach number at the diffuser entry is given by 2 é M3-5 The flow of a perfect gas with zero whirl at the entry is considered below, From Fig. 12.6. B= cy + Cy = Oy + (ty - G2 cot BY = Wh (B+ (1 6 cot BY} (12.83) The velocity of sound is given by a3 = YRT, + (Y~ Nc, Tor (3) (12.84) on For zero whirl at the entry, fon = epTau= hy + 5 A= hy 5 OF-HH) (12.85) Equation (12.32a) gives y= ht >548 Turbines Compressors and Fans Substituting from Eq. (12.85) f= han + $B Wd) From Fig. 12.6, this gives hy = ig + ; w (1 ~ & cosec? B,) TL ht 6B 2 cosec? gho= = 14+ 2 (1 G3 cosec? 6) (12.86) Ta fy 1 * Dep ~ #2 cone By Equations (12.84) and (12.86) give (7-1) GT {i Equations (12.83) and (12.87) yield eB 2e, Toy a (1-43 cosec? os} (12.87) 2 Ate 2 My- 72 __, __ #24 dr cota)” ay gg) Dept ad (1-93 cosec? B,) Jeph, O% : The stagnation temperature rise ratio is Af, Wyp Wb Toe = tg cot 12.89) To ep Syke, 1 ~ 920% Ba) : The following relation is obtained from Eqs. (12.88) and (12.89) Th M, =r oR bbs) (12.90) Tr For a given gas and duty (fixed values of yand AT),/Tpy), the impeller exit Mach number depends on > and f,. To avoid the possibility of shocks, the Mach number at the diffuser entry must not be greater than 0.9. The actual value of this Mach number will be lower than that given by Eq, (12.88) due to diffusion in the vaneless space. The deterioration of diffuser performance is significant in the presence of shocks in the flow field. > 12.7 Volute Casing®*-627 The volute or scroll casing collects and guides the flow from the diffuser or the impeller (in the absence of a diffuser). The flow is finally discharged from the volute through the delivery pipe. For high pressure centrifugal compressors or blowers, the gas from the impeller is discharged through 4 vaned diffuser, whereas for low pressure fans and blowers, the impeller flow is invariably collected directly by the volute since a diffuser is not required because of the relatively low pressures. Figures 12.24 and 12.25 [Link]- _Centifugal Combressor Stage 549 show a volute casing along with the impeller, diffuser and vaneless spaces. The volute base circle radius (r;) is a little larger (1.05 to 1.10 times the diffuser or impeller radius) than the impeller or diffuser exit radius. The vaneles space before volute decreases the non-uniformities and turbulence of flow entering the volute as well as noise level. Delivery pipe j—> Exit Tongue Vaneless spaces Diffuser Fig. 12.24 Scroll or volute casing of a centrifugal machine Some degree of diffusion in the volute passage is also achieved in some designs, while others operate at constant static pressure. Different cross-sections are employed for the volute passage as shown in Fig. 12.26. The rectangular section is simple and convenient when the volute casing is fabricated from sheet metal by welding the curved wall to the two parallel side walls. While the rectangular section is very common in centrifugal blowers, the cir-cular section is widely used in compressor practice. While, the volute performance is dependent on the quality of flow passed on to it from the impeller or diffuser, the performance of the impeller or the diffuser also depends on the environment ereated by the volute around them. The non-uniform pressure distribution around the impeller provided by its volute gives rise to the undesirable radial thrust and bearing pressures ‘Two most widely used methods of volute design are discussed below. [Link].i550 Turbines, Compressors and Fans 6 916ue ye uonDe8 ainjon, Buseo aynjoa e yBnouyy Mol Sez By <9 S] <08L =6 Py = eae) ae i ‘did j 0-8 ‘Apnieg 098 =6 [Link]Centrifugal Compressor Stage 551 t | "4 "4 Leesan: Axis (@) Rectangular (@) Circular Fig. 12.26 Different cross-sections of the volute passage 12.7.1 Free Vortex Design Here the flow through the volute passage is assumed to follow Eq. (12.72) for a free vortex flow which is a c= > (12.91) Equation (12.76) further shows that in such a flow for b; = b, the direction of the streamlines remains constant, ie., oe C7] The total volume (Q) of the flow supplied by the impeller is uniformly divided at the volute base circle. Therefore, the flow rate at a section of the volute passage @ degrees away from the section at 6 = 0° is = Q0= 365 2 (12.93) The flow rate through an infinitesimai section (Fig. 12.25) of cross- section (dr x b,) is dQp = cob; dr Substituting from Eq, (12.91), tan @= — = const. (12.92) dQy = Kbs + For the full cross-section of the volute passage, Kp tate % Qn ior 7 7 Kbs in A (12.94) Equations (12.93) and (12.94) give552 Turbines, Compressors and Fans For a rectangular cross-section, it is required to determine the radius (r,) of the volute boundary from @ = 0° to @ = 360°, This can be determined from =r exp (& 4) (12.9) If the cross-section is not rectangular (Fig. 12.25), then the passage area (4g) and the radius (7) of the centre of gravity of the cross-section are to be determined. Here d0,= Kb & , Qo=K foe (12.96) 4 _ 6 2 7 360 K c250 The volume-flow rate (Q) can be determined from the mass-flow rate, assuming the average density of the gas in the volute passage as equal Pa to p= Pe, © Pa Re 12.7.2 Constant Mean Velocity Design For obtaining high efficiency, it is found from experience that it is neces- sary to maintain constant velocity of the fluid in the volute passage at the design point. This would also give uniform static pressure distribution around the impeller. In actual practice, both the velocity and pressure vary across the cross-section of the volute passage at a given section. There- fore, to be more precise, the mean velocity and pressure along the volute Passage are assumed to remain constant. However, this assumption will be violated at the off-design point. For a given value of the mean velocity (c,,), the area distribution is obtained from @ = oy d= =o Qo= en do= 365 2 Therefore, 8 40-365 (12.98) For a rectangular cross-section, Ag= bs (r4- 15) (12.99) [Link]Centrifuge Compressor Stage 553 Thus the volute radius (r,) for given values of r, and b, can be determined. 12.7.3 Volute Tongue Theoretically, the logarithmic curve of the volute casing begins at the impeller exit, but in practice this is not possible. If it is shifted to the ba circle (Fig, 12.25) at @ = 0°, a sharp-edged lip will be formed. This is known as the tongue or cut water of the volute. Its size and geometry have significant effect on the performance of the centrifugal compressors and blowers. In practice the tongue is cut back to a blunt edge and thus actually starts at @= 6, (Fig. 12.24). At this point its inclination (@) must be the same as that of the streamlines. Therefore, referring to Fig. 12.25, the inclination of an elemental length of the volute boundary r 40 is tan or= tan 0 = £ = const. For a given radius ratio (ry/r4), the angle (@, in radians) at which the tongue starts at the base circle is determined. a aee tana) 3 1 iy (2) (12.100) tan 073 Kn Shifting the tongue as shown above improves the performance significantly and the pressure distribution around the impeller is close to a uniform profile. Besides this, the discharge at the maximum efficiency point is also increased and the noise level decreased. ‘The outflow from the volute at the throat is critically affected by the location and the geometry of the tongue. It divides the flow into two streams—one that flows out and the second which renters the volute through the gap at the tongue. If the inclination of the tongue does not conform to the flow direction shock losses and disturbed flow conditions in this area will arise. The gap between the impeller (or diffuser) and the base circle should not be too large because this increases the recirculation of the fluid and leads to additional losses. > 12.8 Stage Losses‘ ‘The power supplied to the centrifugal compressor stage is the power input at the coupling less the mechanical losses on account of the bearing, seal [Link]554 Turbines, Compressors and Fans _ and disc friction. The aerodynamic losses occurring in the stage during the flow processes from its entry to exit are taken into account by the siage efficiency. These losses result from fluid friction, separation, circulatory motion and shock wave formations. They lead to an increase in entropy and a decrease in stagnation pressure. The disc friction loss, though aerodynamic in nature, is considered along with the other shaft losses, The nature of flow and losses occurring in centrifugal compressor stages is considerably different from those in axial compressor stages on account of different configurations of flow passages in the two types. The centrifugal stages, on account of the relatively longer flow passages and greater turning of the flow, suffer higher losses compared to the axial type. This explains the generally lower values of the efficiency of the centrifugal stages compared to the axial type A comparison of axial and radial stages has been given in Secs. 1.9 and 1.10. In this section different losses have been described separately on the basis of their different nature. The components of the stage in which they occur have been mentioned where necessary. 12.8.1 Friction Losses A major portion of the losses is due to fluid friction in stationary and rotating blade passages. The flow, except in the accelerating nozzle and the inlet guide vanes is throughout decelerating, Therefore, the thickening boundary layer (see Sec. 6.1.18) separates where the adverse pressure gradient is too steep. This leads to additional losses on account of stalling and wasteful expenditure of energy in vortices. Secondary vortices develop in diffuser and volute passages. Losses due to friction depend on the friction factor (Sec. 6.1.17), passage length and the square of the fluid velocity. Therefore, a stage with relatively longer impeller, diffuser and volute passages, and higher fluid velocities shows poor performance. The boundary layer on the rotating surfaces is thrown away due to centrifugal force. Therefore, itis more profitable to obtain higher pressure tise by diffusion of flow in the rotating passages. Thus high degree reaction blades, like backward-swept impeller blades, give more efficient stages. Friction losses in the accelerating nozzle and inlet guide vanes are relatively much smaller, On account of high velocities and the dec- clerations that follow at the leading edges of the inducer and the diffuser blades, shock waves (if present) cause additional losses, They can cause separation of the boundary layers leading to higher losses.Centrifugal Compressor Stage 555 12.8.2 Impeller Entry Losses In higher pressure centrifugal compressors, the radial-tipped impeller blades extend into the axial portion (Figs. 12.1 and 12.2). Thus the incom- ing flow is efficiently guided from the axial to the radial direction. However, in centrifugal blowers with relatively lower pressure rise, the impeller blades are located only in the radial portion (Fig. 12.3). Here the flow enters axially and turns radially in the vaneless space before entering the impeller blades. In this process the fluid suffers losses similar to those in a bend. These losses depend on the velocities c, and c, (Fig. 12.10), but are small compared to other losses. 12.8.3 Shock Losses Additional losses that occur in a row of blades in a centrifugal compressor stage on account of incidence are conventionally known as shock losses. The change of incidence itself very frequently results from the operation of the stage away from the design flow conditions. It is unfortunate that this term has come to stay in centrifugal compressors, because in the usual aerodynamic sense, a shock is a discontinuity and arises when a supersonic flow decelerates to subsonic. The shock loss referred to here has nothing of this nature. During the off-design conditions, the flow at the entry of the impeller and diffuser blades approaches them with some degree of incidence. For instance, Fig, 12.27 depicts off-design velocity triangles at the entry of the inducer blades. At the same rotational speed, the reduced flow rate introduces positive incidence whereas negative incidence results from increased flow rate. Large incidences (specially positive), lead to flow separation, stalling and surge. Increased (-/) Fig. 12.27 Entry velocity triangles at off-design operation [Link]556 Turbines, Compressors and Fans Figure 12.28 [(a) and (b)] explains the “shock model” of flow at the impeller entry. Design point conditions are represented by the quantities By, w;, c@ off-design point values are represented by w, and c,. The so- called shock loss results from the sudden change of the velocity vector to correspond to the blade angle (design point air angle) B, through a shock velocity component cy, as shown. The actual axial velocity compo- nent during this change remains unaltered due to continuity considerations. Shock losses are proportional to the square of the shock velocity component. (@) Positive incidence (b) Negative incidence Fig. 12.28 Shock velocity (c,y) (a) due to positive incidence (b) due to negative incidence When shock losses are plotted against incidence (Fig. 12,29), it is found that they increase rapidly at large values of incidence. Losses ~ve<— 0 —>*Vve Incidence Fig. 12.29 Typical variation of shock losses with incidence [Link]ee _ Centrifugal Compressor Stage 557 Shock losses as explained above also occur in the diffuser and volute. 12.8.4. Clearance and Leakage Losses Certain minimum clearances are necessary between the impeller shaft and the casing, and between the outer periphery of the impeller eye and the .g (Fig. 12.1). The leakage of the gas through the shaft clearance is minimized by employing glands. For small shaft diameters with labyrinth glands, the leakage of gas is small. On account of a higher peripheral speed and a large diameter, itis very difficult to provide sealing between the casing and the impeller eye tip. The leakage through this clearance from the impeller exit is recirculated and additional work is done on a portion of the impeller flow which does not reach the stage exit. This loss is governed by the clearance, diameter ratio (do/d,) and the pressure at the impeller tip. It may by noted here that static pressure at the impeller exit is high for a higher degree of reaction. ‘> 12.9 Performance Characteristics’ 47 48° 49° ‘As discussed in Secs. 7.7 and 7.8, the performance characteristic of a centrifugal compressor or a blower at a given speed can be plotted in terms of the following quantities: r= vf) Figure 12.30 shows the theoretical and actual performance chara- teristics (g-y plot) for a centrifugal stage. The actual characteristic is obtained by deducting the stage losses ftom the theoretical head or pressure coefficient, Therefore, the nature of the actual characteristic depends on the manner in which the stage losses vary with the operating parameters. Friction and shock losses effect the performance significantly. ‘As explained in Sec. 11.8, the range of stable operation is restricted by surging and choking which occur at some values of the flow coefficient peculiar to a given stage. The point corresponding to the maximum pressure and efficiency is generally close to the surge point. The basic causes and nature of unstable flow in centrifugal stages are the same as discussed in Secs. 11.8.2 and 11.8.3. However, these stages, particularly those employing a vaneless diffuser, have a wider range of stable operation. This is on account of the absence of stalling of the vaned [Link]558 Turbines, Compressors and Fans ‘Theoretical Pressure coefficient, Actual Flow coefficient, @ Fig. 12.30 Losses and performance characteristic of a centri- fugal compressor stage diffuser. In some centrifugal stages it has been possible to achieve stable operation on the branch of the characteristic with positive slope. Local stalling of some inducer and diffuser blades occurs even at design point operation, Besides this, rotating stall on the lines explained in Sec. 11.8.3, would occur in both the impeller and diffuser. Surging in the centrifugal impeller is generally provoked by large-scale stalling of the diffuser blades. Choking of the centrifugal stage occurs when the Mach number at either the inducer blades or the diffuser throat reaches unity. ETE Ce fore. ESR USA So Velocity of sound Area of cross-section Impeller, diffuser or volute width Fluid velocity Specific heat at constant pressure Diameter Enthalpy Incidence angle Constants Mass-flow, rate, distance in the meridional plane Mach number [Link]Centrifugal Compressor Stage _ 559 Rotational speed Pressure Power Volume-flow rate Radius Gas constant, degree of reaction, radius of curvature Entropy, blade pitch, distance along the streamline Temperature Peripheral speed Relative velocity, work Number of blades s,m Natural coordinates BN EEVE DIOVTVS Greek symbols Air angle in the absolute system Air angle in the relative system Ratio of specific heats ‘Angle shown in Fig. 12.13 Efficiency Diffuser wall angle, angles shown in Figs. 12.4 and 12.25 Slip factor Vorticity, loss coefficient Circulation Flow coefficient Pressure coefficient Rotation, rotational speed in rad/s SES UWE TS SNTR scripts & £ oF Stagnation value Entry to the impeller Exit from the impeller Exit from the diffuser Exit from the volute ‘Actual Blade Hub Entry to the nozzle Meridional Radial, ratio rel Relative s Stip were TR REN560 Turbines, Compressors and Fans Ss, 88 Isentropic sh Shock st Stage t Tip, tongue tt Total-to-total Corresponding to velocity w Axial Design values IGVs Inlet guide vanes Tangential, corresponding to angular position @ *> Solved Examples 12.1 Air enters the inducer blades of a centrifugal compressor at po, = 1.02 bar, Ty, = 335 K. The hub and tip diameters of the impeller eye are 10 and 25 cm respectively. If the compressor runs at 7200 xpm and delivers 5.0 kg/s of air, determine the air angle at the inducer blade entry and the relative Mach number. If IGVs are used to obiain a straight inducer section, determine the air angle at the IGVs exit and the new value of the relative Mach number. Solution: .0412 m? y 1 AIA (diy = 7 0.25? - 0.108) Both the density and the axial velocity component at the entry of the inducer are unknown, Therefore, these are determined by trial and error. Po _ 102x10° L = Po _ 10210 Ket PL Rt ~ 287x333 In the absence of IGVs, = 1,0609 kg/m* 2 =o,2% -__5 eA BA 10809x 00412 ~ 439 mvs 2 2 di _ 11430? _ 2c, ~ 2x1005 ~ 51K 2 T= % - = = 335 — 6.51 = 328.49 K 2ey — 32849\"5 _ = 102 (238 } = 0.952 bar_ _ Centrifugal Compressor Stage 561 0.952 x 10° Pt 397% 32849 ‘The assumed value of c,, can now be checked. cq = [Link] x 0.0412) = 120.16 mvs = 1.01 kg/m? Since the diffencree is large, another trial is made. I Let ¢ =e) = 123 ms 2 2 = 12 _ 2c, 2x 1005 7527K T, = 335 - 7.527 = 327.473 K _ (32743)* 2 1 (@ aa ) x 1,02 = 0.942 py = 0.942 x 109/287 x 327.473) = 1.0023 kg/m? For a check, ¢,; is recalculated q = 5/(1.0023 x 0.0412) = 121 mis This value (compared to the assumed value of 123 mvs) is acceptable, The difference is only about 1%. d= 5 (d, + d) = 0.5 (0.1 + 0.25) = 0.175 m md,N _ 2x0175x 7200 _ en 65.97 mis From Fig. 12.4, = oa = 121 tan By = SE = Geog = 184 B, = 61.49 (Ans.) w, = 121 _ = 137.8 mis sinB, sin 6597 a, = fyRT, = J14%287x 327473 = 362.737 mis wm. 1378 M,, = “1 = 2 = 0: Mos =~ 362737 ~ 38 Ans) The axial entry of the air into the inducer can be obtained by employing IGVs (Fig. 12.5). In this case on, = tan! i = 61.49 (Ans.) j By = 90° Gq =) = 121 mis [Link]562 Turbines, Compressors and Fans 7 The new value of the relative Mach number is different on account of the changed values of w, and ay. 121 137.8 ms 2 2 ef _ 1378" = ee 94 2x 1005 ~ 447K T, = 335 ~ 9.447 = 325,553 K a, = {14287325553 = 361.67 m/s = 121 wl 36167 12.2 Determine the pressure ratio developed and the power required to drive a centrifugal air compressor (impeller diameter = 45 cm) run- ning at 7200 rpm. Assume zero swirl at the entry and 7, = 288 K. = 0.334 (Ans.) Solution: tdyN _ ex 045% 7200 _ . gp = 169.65 mis 2 Fa 2 3s 142 _ Jf, , 169.65 care 1005 x 288 Pro = 1.393 (Ans.) 169.657 = 199.65" _ 99 78 Ky 1 ee Eg P = 28.78 kWi{kg/s) (Ans.) 12.3 A centrifugal air compressor stage has the following data: = Pro w type of impeller radial-tipped speed 17000 rpm impeller tip diameter 48 cm eye tip diameter 24 cm eye hub diameter 12 om mass-flow rate 8 kgs slip factor 0.92 stage efficiency 0.17 entry conditions Po, = 1.05 bar, Ty, = 306 K Determine: (a) The air angles at the hub, mean and tip sections of the inducer, maximum Mach number at the inducer entry, total pressure ratio [Link]_ Centrifugal Compressor Stage 563 developed and power required to drive the compressor without IGVs. (b) The air angles at the hub, mean and tip sections of the IGVs at exit for axial entry to the inducer, total pressure ratio developed and the power required. Solution: Ae A (é- d= a (0.24? ~ 0.12%) = 0.0339 m? 1.05 10° — = 1.195 kg/m? aT oan Poi = (a) Without IGVs (Fig. 12.4) Prey = Mh C1) = Cy = 8(1.195 x 0.0339) = 197.48 m/s Since the actual density will be lower than pp, the axial velocity will be higher. Therefore, as a first trial, a value of ¢,, = 205 m/s is assumed. 2 2 205 °C 2c, 2x 1005 ee T, = 306 — 20.908 = 285.092 K Fa 35 a-(Z) pam (82) x 1.05 = 0.8196 bar To. 306 08196 105 _ 3 = 97x 285095 — 1° Kem As a check C4 = 841.0 x 0.0339) = 235 m/s This is much higher than the assumed value. Therefore, another trial is required. Tl Let cy, = 240 mis 2 2 d 240 A iS = 28.65 2Cy 2x 1005 oo T, = 306 ~ 28.656 = 277.344 K _ (277344V* - p= ( on ) x 1.05 = 0.7442 bar 0.7442 x 105 3 = —— = 0).! kgit PL aerxca7r34a ~ 0935 kom [Link]564 Turbines, Compressors and Fans As a check a = 80.935 x 0.0339) = 252.39 mis The difference is still large. IIL. Let ey, = 262 m/s This gives T, = 271.85, p, = 0.694 bar, P; = 0.8894 kg/m? A cross-check gives ¢,, = 265.33 mis This is an acceptable value. a= 4 012+ 024) =0.18 m iy = aa - BX 018% 17000 = 160.22 m/s = 2 X 160.22 = 106.8 m/s 4, y= s X 160,22 = 427.253 mis Assuming the inducer blades to have free vortex flow, Cat = Cth = Cctm = Cry Therefore, the air angles are Cu _ 26533 tan By, = Th 1068 2484 Buy = 68.07° (Ans.) fee 28 = 58.87° (Ans.) By = tan 26535 ~ 51.169 (Ans) 7 ean 5 pet = 340.65 m/s The value of the temperature 7, corresponding to c,, = 265.33 m/s is 270.975 K. Therefore, the acoustic velocity at the inducer entry is a, = J14x287x 270975 = 329.967 m/s The tip Mach number is = Mu. 34065 _ Mum 7 = s3pog7 7 1032 (ans. [Link]Conta Compresr Sige 505 { + Ssaxarr eee 1005 x 306 Pro = 3-416 (Ans.) P= ri 2 = 8 x 0.92 x 427.2537/1000 P = 1343.5 KW (dns.) (b) With IGVs (Fig. 12.5) For axial entry throughout the inducer blades the air angles at the IGVs exit are: yy = 68.07° jp, = 58.87° (Ans.) , = 51.16° Oat WIN ee ime ‘The absolute velocities and the static temperatures along the height at the inducer entry will vary. At its tip. Wi sin cy, eu 7 340.647 _ 11, = 306 ~ gps 7 24827 K % 287 x 248.27 = 315.84 m/s The relative Mach number at the tip is ay, = 26533 My = Seay = 0-840 Ans un = Feagg ~ 0-840 (Ans.) 7 f+2e quid ah pt i} a3 -16022°)| oo 1005 x 306 : | 0.77 (092 427.258" Py = 2.905 (Ans.) P= tin ugh) P= 8 (0.92 x 427? — 160.22°1000 P = 1137 kW (das) [Link]566 Turbines, Compressors and Fans 12.4(a) Derive an expression for the flow Mach number (M,) at the impeller exit of a centrifugal compressor in terms of the following parameters: My -1(BMantebs} : (b) In a radial-tipped blade impeller the flow coefficient 4, is 0.268 and the diameter ratio (dy/dy) is 2.667. The mean diameter at impeller entry is 18 cm, and speed 8000 rpm. The entry conditions of air are po, = 1.0 bar and Ty, = 293 K. Determine the blade Mach number at entry and the flow Mach number at the impeller exit. Solution: (a) Equation (12.88) is me Bo 03 + (1, cot B,)? Desh w ? cosec? lye m U9 cosec? B) w= “) 2 q SeTo = yy RT = 2 2 2 B d,)" up 2 4 -o-(4 ie erlang raked (a) Ba a Substituting this in the expression for M3, (FM tel rao, cot a m= es [y-N(a,) a We 4) Mi, (1-63 cosec? B,) (b) For radial-tipped blades B, = 90° ; - (2) MA (+93) MB —— (4) 2 42 TES) wae > (Z) aa-8 nd,N 8000 = TaN 0 ( m= SB = x 0.18 x S00 ~ 95398 mis = 0.2197 (Ans) 4% 287 x 293Centrifugal Compressor Stage _ 567 Therefore, w= — 2 02197)" 140268") _ 1424 (2.667 x 02197)? (1- 0.268") M3, = 0.3459 My = 0.588 (Ans.) 12.5 The tangential velocity component of air at the volute base circle (r = 25 cm) is 177.5 m/s, Determine its shape and throat-to- diameter ratio for a constant width of 12 cm and discharge 5.4 m°/s assuming (@) free vortex flow and (b) constant mean velocity of 145 mis. Solution: (a) Free vortex flow K = ryC93 = 0.25 x 177.5 = 44.375 m/s 2-54 = 101 0 1016 14 = 75 OXD {28} = 25 exp { on ‘The volute radii at eight angular positions are given in the follow- ing table: ry 28.38 47.12 5338 60.71 68. The length of the throat L = 68.92 — 25 = 43.92 cm L _ 4392 30 0.88 (Ans.) ay (6) Constant mnean velocity ¢y ~ 145 10s Ay= £ x eA x 108 = 372.41 {et cm? b (rg 13) = Ae 12 (4 - 25) = 372.41 {et 6) ry = 25 + 31.03 {2h ona [Link]