Abstract:

Based on the genetic optimization theory, a design method for turbomachine blades is presented. The method, taking the minimal flow loss in a cascade of the candidate blades as its objective, searches for the optimal blade profile by a genetic algorithm process. A population of candidate blades represented in terms of the so called “chromosomes” is operated by genetic algorithm operators so as to evolve iteratively to the minimization of the cascade′s flow loss, and finally to generate an optimal blade profile. For the easy use of the genetic algorithm the blade profiles are parameterized using a blade thickness distribution superimposed on a camber line which is a cubic polynomial modified by four Hicks Henne functions. A CFD solver, developed by the authors based on the boundary element method, is employed in the method to provide the flow field analysis for a given blade profile. The method is verified by a practical case of a centrifugal compressor diffuser blade design. The numerical simulations show that the new approach successfully searched out the diffuser blade that has minimal flow losses.

Key words: genet ic algo rithm, turbomach ine, blade, design, diffuser