基于灵敏度分析的可调静叶调节机构运动精度优化

为提升高负荷多级轴流压气机静叶调节机构的运动精度，针对5级联调构型，梳理了37个对调节精度可能存在影响的 参数，提出了一种基于局部灵敏度分析的参数降维预处理优化方法。通过局部灵敏度分析认识各参数对调节精度的影响规律，甄 选重要参数，依靠自主优化程序与多体运动学软件联合的技术途径并采用SQP算法对机构进行优化：仅采用相对灵敏度影响最大 的9个参数进行优化时，调节精度达到0.4020°；采用剔除13个绝对灵敏度最小参数后的参数进行优化时，调节精度优化达到 0.2000°；而采用全部37个参数进行优化后，调节精度依然为0.2000°。结果表明：所提出的降维预处理优化方法能够在保证运动 精度的同时有效减小计算规模，联合优化的方式较传统虚拟样机优化具有更高的精度。

Motion Accuracy Optimization of Variable Stator Vane Adjusting Mechanism based on Sensitivity Analysis

To improve the accuracy of the variable stator vane adjusting mechanism of a highly loaded multistage axial compressor, by sorting out 37 parameters that may influence the adjusting accuracy of the five-cascade modulation configuration, a parameter dimension reduction preprocessing optimization method based on local sensitivity analysis was proposed. Through local sensitivity analysis, the influ? ence of each parameter on the adjusting accuracy was understood, and important parameters were selected. Based on the SQP algorithm, the mechanism was optimized by the joint technical approach of an in-house optimization program and Multibody Dynamics (MBD) software. The results show that if the 9 parameters with the greatest influence on the relative sensitivity are used for optimization, the adjusting accu? racy is 0.4020°; if the 13 parameters with the smallest absolute sensitivity are removed for optimization, the adjusting accuracy reaches 0.2000°; and if all 37 parameters are used, the accuracy is still 0.2000°. The proposed dimension reduction preprocessing optimization method can ensure motion accuracy and effectively reduce the calculation scale, the joint optimization method also has higher accuracy than traditional virtual prototype optimization.