MULTI-OBJECTIVE AERODYNAMIC OPTIMIZATION OF GAS TURBINE EXHAUST DIFFUSER: EXPERIMENTAL AND NUMERICAL VALIDATIONS

To improve the aerodynamic performance of a typical exhaust diffuser, a multi-objective optimization system for the shell profile of the exhaust diffuser is constructed under the condition of considering the inlet swirl. Exhaust diffuser performance is evaluated using two objectives: static pressure recovery coefficient and outlet velocity uniformity coefficient. The shell profile is defined by five control points, and the aerodynamic performance of the exhaust diffuser model is calculated using the realizable-k-epsilon turbulence model. The multiobjective optimization is carried out using the NSGA-II algorithm combined with the Kriging agent model, and two optimal design schemes, optimal design 1 with the maximum static pressure recovery coefficient and optimal design 2 with the best outlet velocity uniformity, are obtained. Numerical calculation and experimental measurement were carried out to evaluate the aerodynamic performance of the baseline and the optimized exhaust diffuser. The results show that the separation vortices in the two optimized exhaust diffusers are significantly weakened. The reverse pressure gradient and flow loss near the strut of optimal design 1 decrease significantly and the outflow uniformity of optimal design 2 is improved significantly. The static pressure recovery coefficient and the outlet velocity uniformity coefficient of the optimal design 1 exhaust diffuser are increased by 11.5% and 3.2%, respectively, and the static pressure recovery coefficient and the outlet velocity uniformity coefficient of the optimal design 2 are increased by 5.7% and 10.9%, respectively.