Numerical Simulation Of Wake Flow Characteristics In High Pressure Turbine Cascade

In turbine engines, the upstream wake would have a great impact influence on the downstream blade, thus it’s vital important to understand the characteristc in the wake flow. coupling with the trailing edge shock and trailing ejection, making wake flow-field extremely sophisticated. In this essay, we are attempting to investigated the wake flows characteristic in the high pressure turbine based on unsteady RANS calculaiton.Calculations were carried out using an self-development 3D CFD code, named 3D-Fluid, the unsteady Favre-averaged three dimensional compressible Navier-Stokes equa-tions written in curvilinear coordinated is solved in 3D-Fluid, Both FDM(Finite Differ-ence Method) and FVM(Finite Volume Method) can be used in 3D-Fluid to discretize the equation. Various Upwind scheme are used for convective flux while second central dif-ferencing is used for viscous terms. Time advancement can be explicit, like Runge Kutta method, or be implicit, like LU-SGS. Dual time stepping is used to model unsteady flows. Multigrid and residual smoothing are available for acceleration, many turbulence models are implemented in 3D-Fluid, including B-L model, Wilcox k-ω model, also a hybrid model PANS based on k-ω is also included. The solver has been validated vary a lot of classical problems, the results showed that 3D-Fluid can be used in a wide scope of application and can be used to solve complex fluid problems.The experiment data for VKI turbine blade is used to validate the code’s ability to model the complex wake flow characteristic in the turbine cascade, both RANS and uRANS calculaiton is made, compared with experiments, it’s find that althouth RANS can predict the mean flow quantities well, but it failed to predic the trailing edge flow characteristics, the uRANS results showed that vortex street shed from blunt trailing edge, wich has a great influence one trailling edge pressure distribution and the influence is not limited in the trailing reagion.The numerical issue of turbulence model, spatial discretization scheme and time-step is examined. Although all the choosing turbulence model can predict boundary layer shape very well, none of them can predict a right trailing edge pressure distribution, which should be caused by the inability of RANS to model the turbulence characteristic in the wake flow. Different order of spatial difference scheme have similiar results, the result got by a seventh order WENO shceme is slightly beeter. the choose of time-step has vital influence on the resutls, a smaller value tends to get a better result. but futher decreasing time-step won’t improve the simulaiton all the way, but can only lead to a time-step independence state.The influence of trailing edge geometry is studied. It’s find that a elliptic trailing edge tends to increase the base pressure and thus reducing the mixing loss greatly. A thinner trailing edge has a much less loss than thicker edge. Also, it’s noticed that the boundary layer state has a close relationship with the formation of vortex street. The loss caused by wake vortex in different exit Mach number is studied, after diving the profile loss into boundary layer loss, shock loss and mixing loss, it’s found and the boundary layer loss and shock loss are very different between results predicted by RANS and uRANS, it can be easily understand cause RANS can’t predict the trailing edge pressure well.