基于Kriging插值代理模型的轴流式止回阀多目标优化

Multi-objective Optimization of Axial Flow Check Valve Based on Kriging Interpolation Model

基于Kriging插值代理模型,对轴流式止回阀的各项性能进行综合优化,具体的优化目标为:降低正向流阻;减小阀芯振动;提升止回性能。通过试验设计方案找到影响轴流式止回阀正向流阻的主要结构参数因素,并以其作为优化对象,通过Kriging插值法拟合主要结构参数对应的性能样本点,在保证拟合精度的情况下,使用NSGA-II遗传算法同时对3个主要结构参数进行多目标优化,得到了三目标Pareto前沿,经过权衡不同因素对性能指标的影响大小,最终得到了综合性能最优的结构参数为:阀瓣丰满度系数α_(1)=2.199,阀芯长度L=386.322 mm,阀芯半径R=113.997 mm。其对应的止回阀性能为:正向出口流量Q=161.839 kg/s,正向出口流速v=3.30235 m/s,止回阀进口压力p=97632.2287 Pa。相比优化之前,阀门正向流阻减小了1.2%,阀芯振动降低了0.24%,止回性能提升了2.3%。

Based on the Kriging interpolation agent model,the performance of axial flow check valve was comprehensively opti⁃mized.The specific optimization objectives were as follows:to reduce forward flow resistance;to reduce vibration of the spool;to im⁃prove check performance.The main structure parameters affecting the forward flow resistance of the axial check valve were found through experimental design scheme.Taking them as optimization objects,the Kriging interpolation method was used to fit the sample points cor⁃responding to the main structure parameters.In the case of ensuring fitting precision,NSGA-II genetic algorithm was used for the multiobjective optimization of the three main structure parameters.Three Pareto fronts were obtained.After weighing the influence of different factors on performance indexes,the structural parameters with the optimal comprehensive performance are finally obtained as follows:disc fullness coefficientα_(1)=2.199,spool length L=386.322 mm,spool radius R=113.997 mm.The corresponding check valve perform⁃ance is:forward outlet flow Q=161.839 kg/s,forward outlet flow rate v=3.30235 m/s,check valve inlet pressure p=97632.2287 Pa.Compared with pre-optimization,the valve forward flow resistance is reduced by 1.2%,spool vibration is reduced by 0.24%,and check performance is improved by 2.3%.