| Summary: | 碩士 === 國立成功大學 === 航空太空工程學系碩博士班 === 95 === The purpose of the present study is to investigate the viscous and tip clearance effects on the transonic rotor 67 under acoustic excitations. A high-resolution aeroacoustic Navier-Stokes flow solver that can capture shock wave and resolve acoustic waves was first developed and validated. This numerical procedure employs the modified Osher-Chakravarthy upwind MUSCL type Total Variation Diminishing (TVD) scheme for acoustic and discontinuity capturing. Time-accuracy is accomplished by using implicit Approximate Lower/Upper Factorization (ALU) together with Newton subiterations. Sound source modeling, characteristic farfiled and inflow/outflow boundary condition treatments were carefully implemented to result in an accurate aeroacoustic solver. Algebraic turbulent model was employed. Comparing the results of Euler simulation without tip clearance and Navier-Stokes simulation with tip clearance, significant discrepancy in shock position and airloads was observed. The viscous simulation results indicate that the tip leakage flow forms a well-defined vortex emanating from the blade tip leading edge, which moves toward the pressure side of the blade passage. A distortion of the passage shock was seen arising from the interaction between leakage vortex and shock. Large pressure rise behind the shock causes a substantial diffusion of the vortex, resulting in a region of low speed zone occurring behind the interaction spot as well as a high blockage found in the blade passage. For low-frequency acoustic excitation, the results of inviscid quasi-steady and inviscid unsteady simulations are similar. In order to render the viscous simulation be realized within a tolerable time frame, quasi-steady approximation for the Navier-Stokes simulations was adopted. It was found in the acoustically excited flows that the change of airloads of the Navier-Stokes simulation, which is caused mainly by the shock excursion, is greater than that of the Euler simulation. Shock root structure is in general weaker in viscous flow owing to the presence of boundary layers, which causes shock in viscous flow be easier to move by acoustic excitation. The present excited flow field is not sensitive to the forcing locations. In general, both shock and primary vortex move upstream in the blowing phase and move downstream in suction phase. The effectiveness of acoustic excitation was shown to be more prominent in the viscous simulation, implying that the feasibility of acoustic excitation is much more possible than that was predicted in the inviscid analysis.
|
|---|
