A combinatorial optimization design method applied to S-shaped compressor transition duct design

Abstract

This paper presents a combinatorial optimization method based on uniform design in combination with response surface methodology and genetic algorithm. Uniform design is used to obtain experimental points and response surface methodology to establish a mathematical regression model. Subsequently, genetic algorithm is employed to acquire optimal solution of the objective function. The optimization method has been applied to a two-dimensional S-shaped transition duct design. The process is performed with two design variables. One defines the drop height ratio which describes wall profile, and the other depicts the length ratio between the axial length of the S-shaped transition duct and the duct inlet height. Total pressure loss coefficient as an aerodynamic performance parameter is selected as the objective function for optimization. The objective function is numerically assessed at design points sampled by uniform design in the experimental domain. The initial transition duct was designed with a radius-change to length ratio 11.6% larger than current engine design limits, and the optimization yields a decrease of 36.9% in total pressure loss and more uniform distributions of parameters at the outlet. The paper shows that the described optimization method can be applied to turbofan engines to increase the radial offset and decrease the axial design space between the fans and cores without jeopardizing performance.