Three different optimization methods using genetic algorithms have been developed, aiming to achieve better aggressive intermediate turbine duct (ITD) performance. To overcome defects of simple genetic algorithms, a niche genetic algorithm is used, for its better adaptability to multi-peak function. These three methods are two-dimensional optimization for pursuing the highest static pressure coefficient; two-dimensional optimization via controlling static pressure coefficient; and a further three-dimensional optimization. The second method introduces a restrainer to make sure the maximum value of static pressure coefficient gradient less than a limitation. A simulation case, an ITD with eight struts, was implemented to demonstrate the capabilities of the presented optimization methods. Compared to the baseline ITD, the results show as follow. The first method obtains a best static pressure coefficient but a severe separation. The second optimization method with static pressure coefficient gradient control can definitely suppress separation in the ITD, the second method also obtains better static pressure coefficient and the lowest total pressure loss coefficient. The third method, a further three-dimensional optimization can obtain better ITD overall performance because of a more realistic simulation. Nevertheless, the second two-dimensional optimization method can get good enough results while it is apparently much more time-saving compared to three-dimensional one, which make it more suitable for engineering applications.
Parameter Estimation in a Three-dimensional Wind Field Model Using Genetic Algorithms
