Adjoint-Based Optimization Method with Linearized Sst Turbulence Model and A Frozen Gamma-Theta Transition Model Approach for Turbomachinery Design
Pengfei Zhang,Juan Lu,Zhiduo Wang,Liming Song,Zhenping Feng
DOI: https://doi.org/10.1115/gt2015-42582
2015-01-01
Abstract:In this paper, based on the grid node coordinates variation and Jacobian Matrices, the turbulent continuous adjoint method with linearized turbulence model is studied and developed to fully account for the variation of turbulent eddy viscosity. The corresponding adjoint equations, boundary conditions and the final sensitivities are formulated with a general expression. To implement the adjoint optimization of the transition flow, a flow solver combining the transition model with the turbulence model is employed, and an adjoint optimization framework with linearized SST turbulence model and a frozen Gamma-Theta transition model is established. In order to choose an appropriate objective for the transition flow optimization, four objectives are studied, including the entropy generation, the total pressure loss coefficient, the field integral of turbulent kinetic energy, the area ratio of transition and turbulent regions to the suciton side. And finally the entropy generation is adopted as the objective. Then, the derivation of the adjoint system for the entropy generation optimization is presented. To demonstrate the validity of the adjoint system for transition flow, four shape optimizations for the bypass transitions and the separation-induced transition are implemented. A 2D isentropic case for bypass transitions is conducted to compares the performances of the fully turbulent adjoint system and the frozen Gamma-Theta transition adjoint system, while the other isothermal case is performed to take the aerodynamic and heat transfer issues into account together. The case of separation-induced transition is performed and also consistent well with its flow mechanism. The four optimization results illustrate the effectiveness of the adjoint system for the transition flow optimization, which can improves the performance of overall cascades and the transition region.