基于试验设计及支持向量机的向心叶轮结构优化设计方法

针对向心叶轮结构优化设计耗时长的难点，研究并采用了2类关键技术：其一是利用试验设计技术从大量的轮盘结构尺寸中挑选出重要尺寸作为优化设计变量，以减小优化问题规模;其二是利用支持向量机技术建立向心叶轮强度及振动分析的代理模型以缩短分析时间.在以上研究工作的基础上建立了1种快速高效的向心叶轮结构优化设计方法.随后的1个算例表明：优化后的叶轮寿命提高4倍，优化设计时间为传统方法的1/27，具有良好的工程应用前景.

Structure optimization method of radial inflow impeller based on design of experiment and support vector machine

The investigation of two critical techniques was conducted to shorten the structure optimization time of a radial inflow impeller. Firstly, in order to reduce optimization scale, the technique of design of experiment (DOE) was used to choose the critical ones from a large number of disk structure dimensions as the optimization design variables. Secondly, in order to lessen analysis time, the technique of support vector machine (SVM) was used to establish surrogate model for stress and vibration analysis of the radial inflow impeller. Based on the investigation above, a structure optimization method of radial inflow impeller was presented, which was low time-consuming. It was demonstrated by one example that the life of optimized impeller increased almost four times and optimization time was one twenty-seventh of that of traditional optimization method.