Experimental And Numerical Investigations On Complex Flow Structures Of Multistage Axial Flow Compressors

Viscous Endwall complex flow will seriously reduce the aerodynamic performance of multistage axial compressor. Previous investigations show that the endwall losses, including tip clearance and secondary flow, usually account for approximately 50-70% of total losses in compressors. So it is essential to study the flow structure and main flow mechanisms of multistage compressor in depth, especially the endwall flow, which has an important pratical significance to the aerodynamic performance improvement of high pressure compressors and aero-engines. In consideration of its large size and high rotational speed, detailed flowfield measurement is very difficult. Furthermore, it is high-cost, high-risk and time-consuming, so low-speed model testing received substantial applications. But till now the experimental research carried out around this field in China is still scarce, which largely restricts the improvement of the aero-engine design ability in our country.Compressor performance and detailed internal flowfield structures of two 4-stage compressors used for low-speed model testing with two rotors which are tip-critical and non-tip-critical respectively, were studied with comprehensive experimental measurements in this dissertation, supplemented with steady/unsteady CFD numerical method. The main flow structures in the compressors were measured, including: flow characteristics of approximate-repeating stage, the forward movement of tip leakage flow(TLF) and the interface between TLF and mainflow, the development of blade boundary layer and the emergence of two corner vortexes which largely reduces the compressor performance in the endwall-suction corners, the initial flow seperation occurance near mid-span for stators with larger bowed-shape while reducing the flowrate, et al. The physical mechanisms of interactions for various steady/unsteady flow and the originations of high flow loss in multistage axial compressor were revealed. With a certain advanced design technologies hub flow was improved and compressor aerodynamic performance was increased correspondingly. This research established a good foundation for the design of high pressure compressor of modern advance commercial engine. This dissertation mainly consisted the following third parts:The first part briefly discussed the aerodynamic design of low-speed datum 4-approximate-repeating-stage, based on modeling principles validated with previous experimental results, and detailed experimental studies of its internal flow structure. The design consists of setup of high-speed modeling target, global parameter design of low-speed model stage, throughflow design of both repeating-stage and approximate-repeating-stage. Both of approximate-repeating-stage design and consideration of blockage in low-speed compressor design have not been presented in previous published paper, which complete the design system of low-speed model testing, and they can effectively improve the modeling accuracy and the reliability of low-speed compressor design.The uncertainty and main influence factor for all important parameters were analyzed: the uncertainty of flowrate coefficient and total-total pressure rise wass 0.42% and 0.4% respectively. Detailed flow structure in the modeling stage were comprehensively described through various measurements, including: flow characteristics of approximate-repeating stage, the forward movement of tip leakage flow(TLF) and the interface between TLF and mainflow, the development of blade boundary layer and the emergence of two corner vortexes which largely reduces the compressor performance in the endwall-suction corners, et al.. Datum experimental results showed that the blade loading near the hub of the rotor for the 3rd stage was relatively large and a certain flow seperation occured, which would induce larger flow blockage, larger flow seperation and higher total pressure loss.In consideration that a certain flow seperation occurred near the hub of the modeling stage, the second part of this dissertation mainly focused on the revised designs of datum compressor blades of the 3rd stage and the correspondingly low-speed compressor experimental results of its performance and detailed internal flow structures. Parametric investigations of circumferential stacking principles of rotors and stators for the modeling stage were carried out. The choices of endwall dihedral angle were determined, which was effective in reducing flow loss near those regions. The optimization factors for the final revised program consists of “J” type stacking for the rotor, increased stator inlet metal angle, leading-edge loaded technology for the stator and larger bowed shape. Results showed that compressor torque efficiency increased nearly 1 point, total-total pressure rise improved 1.4% and the stall flowrate remains unchanged. The optimization factors were basically revealed from the detailed flowfield structures. With the experimental and CFD results, basic flow mechanisms of the revisions that improve the aero-performance were analyzed. Unsteady flowfield travese by dynamic total pressure probe were carried out to deeepen the knowledge of unsteady interactions in the embedded stages. With spectrum analysis and ensemble-average methods, it can be seen that the shedding of upstream boundary layer caused higher energy and higher root mean square values of total pressure(PtRMS) for the blade passing frequency(fBPF) at both sides of the stator wake region. In addition, the interactions of boundary layer for the stator and upstream stators strengthed the high-order harmonics, which is more obvious near the tip and stator outlet, and the energy magnitude of 2fBPF is 8 times of that for fBPF near the tip region at stage 3 outlet. This finding is considered a value for aero-noise reduction and vibration control in multistage compressors.Detailed numerical investigations of two compressor rotors A and B usde for low speed modeling testing were discussed in the third part. Main flow structure of tip leakage vortex(TLV) was revealed by studies. For non-tip-critical rotor B, flow leaked from tip clearance near the leading edge entrains into TLV below 62.5% height position in the tip gap, and it occupies outer space in the TLV as the height position increased. Conversely, flow above 62.5% height position of tip gap won’t entrain into TLV and it mainly performs as double leakage. For tip-critical rotor A, self-unsteadiness occurs at its near-stall conditions. The occurance of this phenomenon relied heavily on tip clearance size, while upstream boundary layer condition has little effect. Furthermore, two blockage region occurs in the tip region, the larger of which caused by the strong interactions of tip secondary vortex(TSV) and TLV breakdown leads to large flow loss, and it directly determines the aero-stability of the rotor. It deems that to effectively remove or at least reduce the strength of TSV with a certain flow control strategy is helpful to expand the stability range of rotor A.