An optimization approach for improving turbomachinery performance was proposed based on a three-dimensional inverse design method, a Computational Fluid Dynamics (CDF) and optimization algorithm. By combining the three-dimensional inverse design method and CFD predictions, the blade loading parameters which is the major inputs for the three-dimensional inverse design method were treated as design variables and the impeller performance predicted by CFD was treated as an objective function of the optimization problem.
Firstly, to clarify the effects of optimization algorithm, mixed-flow pump impellers (Ns400), with a specific speed of 400 (m3/min,m,min−1) or 0.155 (non-dimensional), were optimized to improve the impeller efficiency by using several optimization algorithm. From these results, it was confirmed that turbomachinery optimization using the three-dimensional inverse design method is a multi-peak problem and it is essential to use exploratory techniques such as Simulated Annealing. Then, a mixed-flow pump impeller (Ns1350), with a specific speed of 1350 (m3/min,m,min−1) or 0.523 (non-dimensional), was optimized to improve the impeller efficiency with constraints for suction performance by Simulated Annealing. Reasonably high efficiency and high suction performance were confirmed by comparing the CFD results with those for the previous design which employed manual optimization.