SCIENCE CHINA Technological Sciences Progress of Projects Supported by NSFC• RESEARCH PAPER •
October 2013 Vol.56 No.10: 2531–2539 doi: 10.1007/s11431-013-5317-z
Effects of blade bowing on the performance of a high pressure-ratio turbocharger centrifugal compressor with self-recirculation casing treatment ZHENG XinQian* & LAN ChuanJie State Key Laboratory of Automotive Safety and Energy, Tsinghua University, Beijing 100084, China Received March 25, 2013; accepted July 24, 2013; published online August 27, 2013
The effects of blade bowing on the performance of a high pressure-ratio turbocharger centrifugal compressor were studied by experiments and numerical simulation. The results showed that the negative bowing was capable of increasing the choke mass rate and the efficiency but decreased the surge mass flow rate, while the positive bowing had the opposite effects. When coupling with the self-recirculation casing treatment, the surge mass flow rate of the compressor with negative bowing blade was almost identical with that of the prototype, while the choke mass flow rate was still larger, and the total effect contributed to an increase of the stable flow range by 5.85% at design speed. Besides, the flow mechanism of the coupling effects of blade bowing and self-recirculation casing treatment was discussed. blade bowing, self-recirculation casing treatment, high pressure ratio, turbocharger, centrifugal compressor Citation:
Zheng X Q, Lan C J. Effects of blade bowing on the performance of a high pressure-ratio turbocharger centrifugal compressor with self-recirculation casing treatment. Sci China Tech Sci, 2013, 56: 25312539, doi: 10.1007/s11431-013-5317-z
Nomenclature a attachment line width of the horizontal groove of SRCT bb width of front groove of SRCT bf width of rear groove of SRCT br DSFR difference of SFR between bowing blade and radial blade height of groove of SRCT hb Nb negative bowed blade bow proportion H-S distance Pd Pb positive bowed blade Rb radial blade R Theta offset R s separation line sp saddle point
*Corresponding author (email:
[email protected]) © Science China Press and Springer-Verlag Berlin Heidelberg 2013
Sf
distance between main blade leading edge and front groove of SRCT distance between main blade leading edge and rear Sr groove of SRCT SFR stable flow range SRCT self-recirculation casing treatment
1 Introduction With increasing low-end torque and downsizing requirements, turbocharger centrifugal compressors need wider stable flow range especially at high pressure ratio. When the boost pressure exceeds approximately 3.5, the flow field of the compressor inducer is expected to be transonic with complicated interactions of shock wave, clearance flow and boundary layer on both blades and endwalls, which seriously narrows the stable flow range as well as decreases the tech.scichina.com
www.springerlink.com
2532
Zheng X Q, et al.
Sci China Tech Sci
efficiency [1, 2]. Thus, widening the stable flow range as well as increasing the efficiency is the key issue for advanced turbocharger design and development. Extensive flow control methods for compressor stability enhancement have been proposed [3‒6]. Among these methods, self-recirculation casing treatment (SRCT) is known as one of the most effective ways to delay the compressor surge, and has been also widely used in the turbocharger. Experimental research on SRCT was firstly performed on the vehicle turbocharger centrifugal compressor by Fisher [7], and since then extensive researches have proven the considerable effect of SRCT on the stability of compressors [8, 9]. It was concluded that the pressure gradient between the slots drove the flow recirculation through the casing. When the compressor operated near the stall condition, the low-momentum flow near the shroud of the impeller was sucked into the casing and forced to move toward the upstream of the main flow due to the inverse pressure gradient. As a result, the incidence of the impeller was reduced and the stability was enhanced. The work on SRCT of Hunziker et al. [10] indicated that near surge the low energy fluid sucked from the top of the blades had a velocity component in the rotating direction, which helped to pre-whirl the inlet flow. Ishida [11, 12] suggested that SRCT retarded the development of leakage vortex and thus stabilized the flow. Although SRCT is capable of extending the stable flow range mainly by decreasing the surge mass flow of compressor, it lacks of positive effects on the stage efficiency [10, 13]. Non-radial stacking blade technologies (i.e. bowed, sweep, and leaned blades) have been developed to improve the compressor performance. Extensive studies indicated that non-radial stacking blade is an effective way to control the three-dimensional flow field in blade passages [14‒16]. A detailed literature survey conducted by Vad [17] suggested that the non-radial stacking blades in axial compressor rotors improved the efficiency and extended the stall-free operating range by offering the increased capability for reduction of near-wall and tip clearance losses as well as control of secondary flows and radial migration of high-loss fluid. Fischer et al. [18] applied the bowed stator vanes in a four-stage axial compressor, and succeeded in alleviating the corner stall by shifting the flow to mid-span of the vanes and thus lowering loading near the hub. This led to an increase in the overall total pressure rise as well as the efficiency. The experiments and numerical simulation conducted by Gummer et al. [19] indicated that the positive sweep and dihedral of an axial compressor stator were able to reduce the end-wall losses and extended the stable flow range. Most applications of non-radial stacking blades were focusing on axial compressors. Research of non-radial stacking blades in centrifugal compressors is seldom. For the turbocharger centrifugal compressor, the state of art blade design method incorporates the radial stacking (namely the
October (2013) Vol.56 No.10
gravity centers of the impeller blade are stacked on a radial straight line). As the blade of centrifugal compressors shares similar axial section (inducer) as that of axial compressor, and the design of inducer is believed to be closely related to the performance of the centrifugal compressor, it is surmised that the non-radial stacking may help to control the flow field in inducer and improves the compressor performance. In this work, one form of the non-radial stacking blade, i.e. bowed blade, was employed in a high pressure-ratio turbocharger centrifugal compressor with self-recirculation casing treatment, and both experiments and numerical simulations were conducted to explore the coupling effects of SRCT and the bowed blades.
2 Facility and methods 2.1
Experiment facilities
Main parameters of the centrifugal compressor under study are listed in Table 1. The configuration of SRCT was studied by the authors in ref. [8], and six parameters were used to define its geometry, as shown in Figure 1. Among these parameters, Sr (i.e. the distance between the main blade leading edge and the rear groove) is the key geometric parameter dominating the backflow rate of SRCT. In this work, Sr was set to be 5.25 mm. The design SRCT geometry parameters were listed in Table 2. An experimental rig for testing turbochargers was used to measure the performance of the centrifugal compressor. The centrifugal compressor was propelled by a turbine. The rotational speed of the compressor was controlled by two valves in front of the turbine to allow for control of the flow rate in the turbine. The flow rate in the compressor was controlled by two valves at its outlet with a sensitivity of less than 0.002 kg s1. The major parameters measured during performance testing included total/static pressure and temperature at the inlet/outlet of the compressor, mass flow rate, rotational speed, ambient pressure, and ambient temperature. The temperature was measured by thermocouple with an error of Table 1
Detailed parameters of compressor
Structural parameters Design pressure ratio Design speed Impeller inlet radius Impeller outlet radius Inlet hub blade angle Inlet tip blade angle Number of blades Outlet backswept angle Vaneless diffuser outlet radius Vaneless diffuser outlet width
Value and unit 4.0 136000 r min1 27.5 mm 37.5 mm 65.8º 42.4º 7/7 main/splitter 35º 55.24 mm 3.75 mm
Zheng X Q, et al.
Figure 1
Sci China Tech Sci
Sketch of SRCT in a centrifugal compressor.
Table 2 Parameters of SRCT Parameters bf Sf Sr br hb bb
Value (mm) 5 7.5 5.25 2.4 4 6.5
less than ±1.75°C; pressure was measured by diaphragm pressure sensors with an error of less than ±1.25 kPa; mass flow rate was measured using a vortex flow meter with a relative error within ±0.5%; rotating speed was measured by an electromagnetic transducer with a relative error within ±0.25%. The surge point was judged by monitoring the acoustic noise during the surge process. A distinct cyclic noise akin to a “sewing machine whistle” could be heard at the onset of unstable operating conditions in the compressor, known as compressor surge. The operating conditions just before the onset of “sewing machine whistle” were determined to be the surge point on the performance map. 2.2
2533
October (2013) Vol.56 No.10
a higher resolution of flow quantities. The grid exhibited an acceptable quality as defined by measures of orthogonality (minimum 15°), relative grid spacing in boundary layers (expansion ratio<10), and grid skewness (aspect ratio <1000). The non-dimensional distance of the first grid point from the wall y+ was kept under 10. A constant tip clearance of 0.25 mm was assumed. Figure 3 shows the comparison of CFD and experimental results. It can be seen that the predicted total pressure ratio and efficiency by CFD are higher than those obtained from experiments. Only one passage was modeled and calculated in CFD, which neglected the influence of the volute on the performance and caused the deviation between the results from CFD and experiments. The influence of the volute on the performance of a high pressure-ratio compressor can be referred to ref. [20]. For more information about the validation of the simulation, please refer to ref. [21], in which the performance characteristics predicted by the simulation were compared with the experimental results.
Figure 2
Grids of single passage.
Numerical methods
Numerical simulation was performed using NUMECA EURANUS solver, which is based on a three-dimensional, compressible finite volume scheme to solve steady-state Reynolds-averaged Navier-Stokes equations in conservative formulation. The Spalart-Allmaras (S-A) model was chosen for turbulence closure. A central scheme was applied to spatial discretization and a fourth-order Runge-Kutta scheme was used for temporal discretization. A single impeller passage was modeled. The H&I type grid was adopted to obtain a qualified grid for the blades with different bowing parameters. SRCT was modeled using the Z_R effect function in AutoGrid5, as shown in Figure 2. Grid independency assessment was carried out and the final model consisted of 450000 nodes in order to attain
Figure 3 Comparison of CFD and experimental results. (a) Pressure characteristics; (b) efficiency characteristics at design speed.
2534
Zheng X Q, et al.
Sci China Tech Sci
3 Effect of blade bowing 3.1
Bowed blade designs
The bowed main blades were designed based on the software Concept NREC. Blade bowing is defined by two parameters named “bow proportion H-S distance” (Pd) and “R Theta offset” (R), as shown in Figure 4(a). Pd represents the percentage of span height where the circumferential offset of the blade is the maximum; R is the maximum circumferential offset of the main blade leading edge. If the offset is in the direction of the rotation, then it is referred to as “positive bowing” (Pb), and the opposite direction as “negative bowing” (Nb). The prototype with radial stacking blade was named “Rb”. The definition of “positive” and “negative” of bowing can be referred to Figure 4(b). For the bowed blade, Pd was kept to be 50% from blade inlet to outlet, while R decreased from blade inlet to outlet (referring to Figure 5). Four cases of bowed blades with different values of Rwere designed, including two positive bowing cases (R equaled +4, +2 mm, respectively named as “Pb+4” and “Pb+2”) and three negative cases (R equaled ‒2, ‒1 mm, respectively named as “Nb‒2”, and “Nb‒1”).
Figure 4
Figure 5
October (2013) Vol.56 No.10
3.2
Experiment results
The compressors with bowed blades and without SRCT were tested. Figure 6 shows the characteristics of the cases. As the experiment results show, bowing has a considerable influence on compressor performance. The impact of bowing increases with the rotational speed. At low speed, from 60000 r min1 to 100000 r min1, rotating stall is expected to occur in the vaneless diffuser first, and bowing of the impeller has little impact on compressor’s surge mass flow. But when the rotation speed is higher (100000 r min1‒ 1360000 r min1), rotating stall is expected to happen near the inlet first, and bowing of the impeller has a significant effect on the surge mass flow rate. The difference between the bowing’s effect in the high speed and low speed areas leads to the map of the compressor remaining unchanged under low speed and rotating around the surge point in 100000 r min1 at high speed. That means positive bowing will make the map rotate anticlockwise, while negative bowing will make the map rotate clockwise. Meanwhile, positive bowing will decrease the pressure ratio of com-
Bowed blade configuration.
Distribution of R (R2 m).
Figure 6 Compressor performance. (a) Pressure characteristics; (b) efficiency characteristics at 136000 r min1.
Zheng X Q, et al.
Sci China Tech Sci
pressor, while negative bowing will increase it. In conclusion, negative bowing increases the choke mass flow rate, the surge mass flow rate and the pressure ratio, while positive bowing decreases them. At design speed (136000 r min-1), negative bowing of 2 mm (Pb‒2) increases the choke mass flow rate by 3.41%, and the surge mass flow rate by
October (2013) Vol.56 No.10
3.44%. Positive bowing of 4 mm (Pb+4) decreases choke mass flow by 10.2%, but decreases the surge mass flow by 10.27%. Therefore, the stable flow ranges (SFR) of the three cases are nearly the same. m choke m surge SFR 100%. m choke N const
3.3
Figure 7 sections.
Meridianal view of the impeller and locations of the seven
Figure 8
Figure 9
2535
(1)
CFD results
Figure 7 shows the meridianal view of the impeller. The impeller passages were divided into seven sections, and on each section, the secondary flow on near surge condition was plotted, as shown in Figure 8. It can be seen that for negative bowing, the clearance vortex appears on section 2, while this vortex appears on section 3 and is not so severe. On the contrary, the positive bowing alleviates the tip clearance vortex compared with both negative bowing and radial blade. As the tip clearance flow is believed to be a dominant factor for the compressor stability, the negative bowing tends to trigger the impeller stall and causes the compressor surge as shown in Figure 6. Figure 9 shows the limited streamline at suction surface under design condition. It is found that negative bowing ad-
Flow vortex in impeller passages in near surge condition.
Limited streamline at suction surface in the designed condition.
2536
Zheng X Q, et al.
Sci China Tech Sci
October (2013) Vol.56 No.10
vances the boundary layer flow separate near the middle part of the span, where a saddle point is formed as shown in Figure 9(b). The separation flow then is brought out by the main flow in the passage, instead of going upward and accumulating in the blade tip. By this way the low energy fluid along the span is redistributed and improves the compressor efficiency, as shown in Figure 6(b).
4 Coupling effect of bowing and SRCT on compressor’s performance 4.1
Experiment results
Figure 10 shows the performance characteristics of Cases Nb‒2, Pb+4 with SRCT. Still it can be seen that negative bowing increases the choke mass flow rate compared with the prototype (marked as “Rb+SRCT”), while positive blade bowing decreases the choke and surge mass flow rate, as well as the pressure ratio. The effects become server as the rotating speed increases. At design speed (136000 r mim-1), negative bowing plus SRCT (marked as “Nb‒2+SRCT”) increases choke mass flow by 3.47%, compared to the prototype plus SRTC (marked as “Rb+SRCT”). In addition, when SRCT is available, the negative bowing case (Nb‒2) and the prototype (without blade bowing) have almost the same surge mass flow rates. The tests results indicated that negative bowing and SRCT increase SFR by 5.85% compared with the prototype (with SRCT), while positive bowing plus SRCT decreases SFR by 2.5%. 4.2
CFD results
4.2.1 Near choke conditions The choke mass flow rate of the compressor is usually decided by the area of impeller throat. Since the R decreases from the leading edge to the trailing edge for the negative bowing blade design, the throat area is increased. On the contrast, positive bowing tends to decrease this area. This explains the opposite effect of the positive and the negative bowing on the choke mass flow rate. SRCT affects the flow field in two ways. On one hand, SRCT represents another passage for the fluid to move from upstream of throat to downstream. On the other hand, SRCT links all passages of the impeller, thereby enforcing uniformity of the flow field and weakening the shock downstream of the throat. Bowing of the blades and SRCT affect the choking mass flow rate independently from each other. Figure 11 shows the relative Mach number contour near choke. When SRCT is applied, the impact of blade bowing on choking is slightly weakened. Negative bowing can still increase the choking mass flow rate by enlarging the size of the throat, but the effect on the supersonic flow downstream of the throat is not as obvious as when there is not SRCT. The reason is that SRCT has homogenized the flow field
Figure 10
Performance characteristics of the compressors.
parameters across the passages downstream of the throat, thereby weakening the supersonic flow field significantly. 4.2.2 Near surge conditions In near surge conditions, the fluid is driven to upstream from the slot of SRCT, as shown in Figure 12. There exists a vortex near the shroud, which decreases the effective flow area of the compressor passage. Due to this vortex, the flow rate near the blade tip is increased, and radial migrating of low energy fluid to the blade tips is weakened significantly. Since the effective flow area of impeller passage is decreased, the axial component of the fluid velocity will increase to allow for a stronger resistance of the migration of the secondary flow. As shown in Figure 13, compared to the compressor without SRCT, the case with SRCT has a better flow field featuring a weaker upward migration of the secondary flow and less flow separation near the tip at the pressure surface. When SRCT and negative bowing are available, the flow field is improved significantly, as shown in Figure 14. The jet flow induced by SRCT in the impeller inlet increases the axial velocity of the flow near the blade tip, thus weakening radial migration. The saddle point “sp1” at the blade tip shifts down, and the relevant separation line “s12” extends, which prevents secondary flow from migrating upward, thereby decreases the source of low energy fluid at the blade tip suction surface. Leakage flow separates at “s2” and re-attaches at a2, with parts of it flowing down to “s12”, which means low energy fluid is driven from the top to the middle of the blade. The separation bubble at the top therefore shrinks and the flow field gets better. The coupling effects of SRCT and positive bowing are
Zheng X Q, et al.
Sci China Tech Sci
similar to the case with radial bowing and SRCT. However, there is no saddle point near the tip because the secondary flow exhibits a stronger migration tendency at the suction
Figure 11
Figure 12
October (2013) Vol.56 No.10
2537
surface when the blade is positively bowed. In comparison with the streamline distribution in Figures 13‒15, it can be concluded that the improvement brought by that SRCT is
Relative Mach number contour near choke.
Streamline and entropy contour in meridian view of impeller passage.
2538
Zheng X Q, et al.
Sci China Tech Sci
Figure 13
October (2013) Vol.56 No.10
Limited streamline of radial blade.
Figure 14
Limited streamline of negative bowing blade.
Figure 15
Limited streamline of positive bowing blade.
Zheng X Q, et al.
Sci China Tech Sci
slight when the blade is positively bowed.
5 Conclusions In this work, the coupling effects of blade bowing and SRCT on the performance of a high pressure-ratio turbocharger centrifugal compressor were studied by experiments and numerical simulations. Several conclusions can be drawn as follows: 1) Negative bowing slightly decreases the surge mass flow rate but clearly increases the choke mass flow rate. Besides, negative bowing increases the pressure ratio and the efficiency. Positive bowing decreases the surge mass flow rate, the choke mass flow rate, pressure ratio and the efficiency. 2) Negative blade bowing together with SRCT can improve SFR of the compressor. At design speed, SFR is increased by 5.85%. 3) SRCT increases the effective flow area of impeller passage and introduces jet into the flow field near the blade tip, thereby increases the axial velocity of the fluid near the shroud. The saddle point shifts down to the middle of blade. Low energy fluid separates at this point and mixes with the main flow. This reduces the accumulation of low energy fluid in the blade tip thus delaying the impeller stall. This work was supported by the National Natural Science Foundation of China (Grant No. 51176087). 1
2
3
4
5
Krain H, Karpinski G, Beversdorff M. Flow analysis in a transonic centrifugal compressor rotor using 3-component laser velocimetry. ASME Paper No. 2001-GT-0315, 2001 Ibaraki S, Furukawa M, Iwakiri K, et al. Vortical flow structure and loss generation process in a transonic centrifugal compressor impeller. ASME Paper No. GT2007-27791, 2007 Arnulfi G L, Blanchini F, Giannattasio P, et al. Extensive study on the control of centrifugal compressor surge. Proc IMechE Part A: J Power Energy, 2006, 220(3): 289‒304 Arnulfi G L, Giannattasio P, Giusto C, et al. Multistage centrifugal compressor surge analysis: part ii-numeical simulation and dynamic control parameters evaluation. J Turbomach, 1999, 121(2): 312‒320 Saha S L, Kurokawa J, Matsui J, et al. Passive control of rotating stall in a parallel-wall vaned diffuser by J-grooves. J Fluid Eng—T ASME,
October (2013) Vol.56 No.10
6
7 8
9
10
11
12
13
14
15 16
17
18
19
20
21
2539
2001, 123(3): 507‒515 Asaga Y, Mizuki S, Tsujita H, et al. Surge control of a centrifugal compression system by using a bouncing ball. J Therm Sci, 2008, 17 (2): 141‒146 Fisher F B. Application of map width enhancement devices to turbocharger compressor stages. SAE Trans, 1989, 97(6): 1303‒1310 Zheng X Q, Zhang Y J, Yang M Y, et al. Stability improvement of high-pressure-ratio turbocharger centrifugal compressor by asymmetric flow control: Part II—non-axisymmetric self recirculation casing treatment. ASME Paper No. GT2010-22582, 2010 Tamaki H, Masaru U, Kawakubo T, et al. Aerodynamic design to increase pressure ratio of centrifugal compressors for turbochargers. ASME Paper No. GT2009-59160, 2009 Hunziker R, Dickmann H P, Emmrich R. Numerical and experimental investigation of a centrifugal compressor with an inducer casing bleed system. Proc IMechE Part A: J Power Energy, 2001, 215 (6): 783‒791 Ishida M, Sakaguchi D, Ueki H. Optimization of inlet ring groove arrangement for suppression of unstable flow in a centrifugal impeller. ASME Paper No. GT2005-68675, 2005 Ishida M, Surana D, Ueki H, et al. Suppression of unstable flow at small flow rates in a centrifugal blower by controlling tip leakage flow and reverse flow. J Turbomach, 2005, 127(1): 76‒83 Khalfallah S, Ghenaiet A. Analyses of radial compressor stability improvement by casing treatment. Proc IMechE Part A: J Power Energy, 2012, 226(7): 807‒821 Weigold H D, Neubert R J, Behlke R F, et al. Bowed stators: an example of CFD applied to improve multistage compressor efficiency. J Turbomach, 1997, 199 (2): 161‒168 Benini E. Three-dimensional multi-objective design optimization of a transonic compressor rotor. J Propul Power, 2004, 20(4): 559‒565 Gallimore S J, Bolger J J, Cumpsty N A, et al. The use of sweep and dihedral in multistage axial flow compressor blading — Part I: University research and methods development. J Turbomach, 2002, 124(4): 521‒532 Vad J. Aerodynamic effects of blade sweep and skew in low-speed axial flow rotors at the design flow rate: an overview. Proc IMechE Part A: J Power Energy, 2008, 222: 69‒85 Fischer A, Riess W, Seume R J. Performance of strongly bowed stators in a 4-stage high speed compressor. ASME Paper GT2003-38392, 2003 Gummer V, Wenger U, Kau H P. Using sweep and dihedral to control three-dimensional flow in transonic stators of axial compressors. J Turbomach, 2001, 123(1): 40‒48 Lin Y, Zheng X Q, Jin L, et al. A novel experimental method to evaluate the impact of the volute’s asymmetry on the performance of a high pressure-ratio turbocharger compressor. Sci China Tech Sci, 2012, 55(6): 1695‒1700 Zheng X Q, Huenteler J, Yang M Y, et al. Influence of the volute on the flow in a centrifugal compressor of a high-pressure-ratio turbocharger. Proc IMechE Part A: J Power Energy, 2010, 224(8): 1157‒ 1169