基于HSA-SVR的压电式车削测力仪多维力解耦
Decoupling for Multidimensional Force of Piezoelectric Turning Dynamometer Based on HSA-SVR
张军 1蔡佳乐 1王郁赫 1滕玄德 1张鹏 1王尊豪1
作者信息
- 1. 大连理工大学,高性能精密制造全国重点实验室
- 折叠
摘要
文中针对压电式多维力测力仪向间干扰大,制约测量精度的问题,分析了向间干扰对测力仪测量精度的影响,提出了一种基于支持向量回归机(SVR)的非线性解耦算法.利用混合模拟退火算法(HSA)对SVR进行参数寻优,对比并分析了HSA-SVR和线性最小二乘解耦法(LS)的解耦性能,证明经该方法解耦后向间干扰最大为0.526%,非线性误差最大为0.214%,HSA-SVR具有更好的非线性解耦效果.
Abstract
Addressing the issue of significant inter-directional interference in piezoelectric multidimensional force measuring instruments,which constrains measurement accuracy,this paper analyzed the impact of inter-directional interference on the preci-sion of force measuring instruments and proposed a nonlinear decoupling algorithm based on support vector regression(SVR)ma-chine learning.The optimization of SVR parameters was achieved using a hybrid simulated annealing algorithm(HSA).The 0decoupling performance of HSA-SVR was compared and analyzed with linear least squares decoupling(LS).The results demonstratethat after decoupling,the maximum inter-directional interference is only 0.526%,and the maximum nonlinear error is 0.214%.HSA-SVR exhibits superior performance in decouplingnonlinearities.
关键词
压电测力仪/多维力测量/支持向量回归机/非线性解耦方法/融合算法Key words
piezoelectric dynamometers/multidimensional force measurement/support vector regression machine/nonlinear decoupling method/fusion algorithm引用本文复制引用
基金项目
国家自然科学基金(52075079)
国家科技重大专项(J2019-V-0011-0106)
出版年
2024
NETL
NSTL
万方数据