INFLUENCE OF VANELESS REGION DESIGN GEOMETRY ON THE PERFORMANCE OF CENTRIFUGAL COMPRESSOR
Ahmed S. Hassan
Mechanical Engineering Department, Assiut, University, Assiut, Egypt.
Abstract:- Influence of the vaneless region design geometry on the compressor performance and wide of stable operation have been experientially investigated. Two main geometries of the vaneless region design were investigated. The first design geometry is the clearance between the rotor exit and diffuser vanes inlet. This clearance was changed from minimum to maximum in five steps as C = 0.02, 0.1, 0.18, 0.26 and 0.34 relative to the distance between the rotor exit and the main diffuser vane inlet (original one). The second design geometry is considered by installing small vanes at the vaneless region between the rotor exit and the main diffuser vane inlet with different lengths and circumferential positions. The small vanes lengths were changed as, S = 0.9, 0.75, 0.5 and 0.25 relative to the clearance between the rotor exit and diffuser vane inlet. The circumferential positions of these small vanes were changed from the extension of the main diffuser vanes to 0.25 near pressure side of the main diffuser vanes to the middle or 0.5 of the diffuser channel and to 0.25 near suction side of the main diffuser vanes. The effects of all parameters on the limit of stable operation due to stall and surge at reduced flow conditions as well as on the compressor pressure coefficient were demonstrated. The pressure distributions at one of the diffuser passages at its inlet and exit were obtained. Simultaneously with the above conventional measurements, two pressure transducers with high sensitivity response were installed at the inlet and the exit of the diffuser to measure the time variation of static pressure. The data were processed using the Fast Fourier Transformation analysis (FFT) to estimate the Power Spectrum Density (PSD) for detecting the initiation of rotating stall and surge. The results show 13% improvements in stable flow range with clearance ratio of 0.18 and 5% in pressure coefficient and efficiency. Installing small vanes at 0.25 from the main diffuser vane pressure side gives about 36% improvement in stable flow range relative to the diffuser without small vanes. Comparisons between the experimental results and the available previous experimental and theoretical work show acceptable agreements.
Keywords: Pressure transducers, Vaneless region, Range of stable operation, Time variation of static pressure, Power Spectrum Density.
Nomenclature
b2 Width at rotor exit, m
C Clearance ratio between the rotor exit and diffuser vane inlet, C = c/r2
c Clearance between the rotor exit and diffuser vane inlet, m
c1 Clearance between rotor exit and main diffuser vane inlet, m
Specific heat at constant pressure, KJ/kg K
S length ratio of small vanes, S =l/c1
p Static pressure, N/m2
Pr Compressor pressure ratio
Pressure difference
r2 Impeller outer radius, m
U Impeller tip speed, m/s
r Density, kg/m3
Efficiency,=/
F Flow coefficient = Q/2b2r2U
Y Pressure coefficient =2/
, Angles between two impeller and
diffuser blades respectively, deg
Angles from blades pressure side, deg
1 Introduction
Many compressors used for industrial applications and for refrigeration units require operating over a wide flow range corresponding to load conditions. While, as the flow rate reduces, the compressors have a limited operation range, due to occurrence of self-exciting phenomena that result in machine fracture, like rotating stall and surge [1]. Therefore, the limiting working conditions for stable operation at reduced flow rates have dragged more attention because, it usually occurs very close to the point of the highest outlet pressure and is most often acts a trigger of system surge [2]. Moreover, the flow leaving the impeller has jet and wake and when this flow enters a large number of diffuser passages, the behavior of flow entering different diffuser vanes differs significantly and some of the vanes will experience flow separation leading to rotating stall and poor performance. Hayami et al [3] relates beginning of rotating stall to triggering surge at the area between the rotor tip and the diffuser vane leading edge. Therefore, the performance characteristics of the compressor are complicated functions of diffuser geometry, and its inlet and exit flow conditions. However, many investigations investigate effects of diffuser geometrical parameters and configuration modification on the flow behavior and unsteadiness in both vaneless diffusers [4-8], and in vaned diffusers [9-11]. Engeda [12] made some efforts to propose a general design procedure for the vaneless region and stated that, a change in the radius ratio of the vaneless region for the same impeller radius will change the radius ratio of the vaned portion too. Consequently, there will be mutual and combined effects on the characteristics of the flow within all passages due to changing of both radius ratios. In addition, some investigators studied effects of installing of small vanes either in the vaneless region near the diffuser inlet [8, 9] or inside the vaned diffuser passages [10] on the characteristics of the compressors. In addition, there has been a common consensus among the studies related to the effect of installing small vanes in the vaneless space, concerning the favorable effect of the installation of such vanes on the overall compressor performance. Nakagawa et al. [9] stated that the performance of the compressor will improved with installing small vanes near the diffuser inlet, but it was not possible to investigate the effect of small vanes geometry and position in such experimental investigations because of the complication of the required test rig as well as the tedious procedures needed. Drtina et al. [10] numerically studied effects of geometry and positions of small vanes within the straight blades of the diffuser on compressor performance. They concluded that the flow is sensitive to the position and geometry of the small vanes. To the author’s knowledge, these effects of the geometry (length, width, and shape) and positions of small vane installed in the vaneless region have not fully investigated. Abdel-Hafez et al. [13] numerically investigated the effect of installing small vanes in the vaneless region with diffuser has straight blades on performance of a centrifugal compressor. They concluded that installing small vanes at the vaneless region with specific design increase both the stable flow range by about 28.6% of the compressor maximum flow rates and pressure rise coefficient by about 35% compared with the diffuser without small vanes.
In the present work, the effect of different clearances between the diffuser vane leading edge and rotor exit (c) relative to the rotor tip radius (r2), C = c/r2 = 0.02, 0.1, 0.18, 0.26 and 0.34 on the compressor limit of stability (stall and surge as well as maximum flow rate) and characteristics were experimentally investigated. In addition, the effect of installing small vanes in the vaneless region or upstream the leading edge of the diffuser vanes on the compressor performance characteristics were experimentally investigated. The effect of the length and position of these small vanes on the flow characteristics was investigated. Therefore, the lengths of these small vanes relative to the clearance between the rotor exit and the leading edge of the main diffuser vanes were changed from 0.9 to 0.25. Also, the circumferential positions of these small vanes were changed from pressure side of the main diffuser vanes relative to its channel width from 0 to 1. In additional to the conventional measurements of pressures and temperatures those necessaries for compressor performances characteristics and the pressure distribution at one of the diffuser passages at inlet and exit, the time variations of static pressure as well as power spectrum density at different operating conditions and tests were investigated. The effects of the above two main different parameters on the limit of stable operation due to stall and surge at reduced flow conditions are demonstrated.
2 Compressor Test Facility
The experiments were performed in an open loop centrifugal compressor test facility. This compressor consists of radial blade rotor and parabolic vanes diffuser. The inlet of the compressor is at atmospheric conditions. At the outlet of the diffuser, the flow is collected by scroll and fed into a large tank, which followed by a normalized tube that allows for flow control and measurements. The flow rate is measured using an orifice flow meter and controlled by a valve that is located at the end of the discharge duct, which is used to control the flow rate. The compressor is driven by 5 kW variable speeds D-C motor with manually adjusted speed from 0 to 6000 rpm. Conventional wall pressure taps were installed along the front casing of the whole compressor passage as shown in Fig.1, to detect the pressure rise along the diffuser passage. Those taps were located in front of the impeller passage (points 1 to 9), in the vaneless region between the rotor exit and diffuser vanes inlet (points 10 to 17), along the diffuser passages at different three radial sections (points 18 to 26) and at diffuser exit (points 27 to 39). Also, pressure and temperature tapes were located at outlet pipe of the compressor for securing the overall compressor performance.
In addition to securing conventional compressor characteristics, the time variations of static pressures wee measured at two points at the vaneless region or diffuser inlet and at the diffuser exit using two pressure transducers. The specifications of these pressure transducers are; omega type, PX-236-100GV silicon diaphragm with full bridge design for high sensitivity. The pressure signals from these pressure transducers were amplified through a DC amplifier (SENSOTEC'S SA-BII). The amplified pressure signals were simultaneously sampled for one second at a sampling rate of 1 kHz and digitalized through a 16 bit A/D converter board (high performance, high speed, multi-function acquisition card, PCL-812PG) installed in a personal computer. The board is supported by PC-SCOPE software, which turns the computer to oscilloscope and stores the pressure waveforms in ASCII file. Subsequently the data in the file were processed using the Fast Fourier Transformation Analysis (FFT) to estimate the Power Spectrum Density (PSD) by Welch’s averaged, the modified periodogram method for discrete-time signal vector. The arrangements of the pressure signals from the two pressure transducers and the recording system, which were used for studying the steady and the unsteady flow phenomena, are shown in Fig. 2. Using this advanced measuring technique, the limit of flow stability or the critical flow rate for initiation of rotating stall and surge was detected precisely.
Fig.1 Locations of pressure taps on the compressor casing
Fig.2 Pressure recording system
To achieve the aim of the present work, different clearances as shown in Fig.3a between the rotor exit and the diffuser vane inlet (c), relative to the rotor tip radius (r2), denoted C (= c/r2): = 0.02, 0.1, 0.18, 0.26 and 0.34 were considered. Also, small vanes with different lengths as shown in Fig. 3b were located on the vaneless region at different circumferential positions. At 0.25 from diffuser main vane pressure side with length of 0.9 relative to the total clearance between the rotor exit and the main diffuser vane inlet denoted Sp = 0.9 and at length 0.5 denoted Sp=0.5. And, Sm = 0.9 and Sm = 0.5 denote that the small vanes are located at the middle channel of the main diffuser vanes with length 0.9 and 0.5, respectively. At near the main vane suction side denoted Ss=0.9 and 0.5 depends on their lengths. At along the main diffuser vane denoted So = 0.9 and 0.5. However, the lengths of these small vanes relative to the distance between the rotor exit and diffuser vanes inlet were changed as S = 0.9, 0.75, 0.5, and 0.25.
(a ) (b)
Fig.3 Splitter with different positions
3 Experimental Results and Discussions
3.1 Effect of the clearance between the rotor exit and the diffuser vanes inlet on the compressor characteristics
Figure 4 shows effects of the different clearances between the rotor exit and the diffuser vane inlet relative to the rotor tip radius, C= 0.02, 0.1, 0.18, 0.26 and 0.34 on the compressor performance. This figure indicates that at clearance between the rotor exit and diffuser vane inlet, C = 0.18 the compressor gives maximum pressure coefficient as shown in the left side of the graph at low flow rates. This gain in compressor pressure coefficient at the higher efficiency region is due to good compatible design for the diffuser vane leading edge position relative to the rotor exit that leads to the improvements in compressor pressure coefficient relative to the other clearances. At the right side of the graph or maximum flow rates for lower efficiency of the compressor, low-pressure coefficient was observed with C = 0.18 relative to the large clearances. This is may be due to the blockage at the leading edge of the diffuser vanes. At clearance C, of 0.1, low values of pressure coefficients are observed. This is due to increase the blockage at the diffuser leading edge, hence increasing the acceleration of the flow in the diffuser with the small clearance, and consequently decreasing the pressure coefficient.
Figure 5 shows effects of the clearances between the rotor exit and the diffuser vane inlet on the compressor efficiency. Improvement in the compressor efficiency with clearance, C = 0.18, at the lower flow rates was observed. At compressor operating point of F = 0.08, the compressor gives about 3.5 % improving in efficiency relative to that of clearance C = 0.34 or with the original compressor. Figure 6 shows effects of the clearances between the rotor exit and diffuser vanes inlet on the distribution of the pressure rise at the vaneless region and at the diffuser exit. Using clearance ratio of, C = 0.18 gives uniform distribution of the pressure rise at the vaneless region (Fig.6-a), while with the same test at C=0.18 maximum pressure rise is observed at the diffuser exit (Fig.6-b). That is, designing the compressor at clearance, C = 0.18 gives the maximum pressure at the diffuser exit. This is due to uniform flow distribution at the vaneless region with best position of the diffuser relative to rotor.