一种数控折弯加工多参量非线性补偿新方法

摘要

对复变量移动最小二乘法及其基函数、紧支域半径等重要参数进行研究，针对它在非均匀采样点集应用上的不足，提出一种待拟合点紧支域半径随采样点区域稀疏程度动态调整的改进型复变最小二乘法，并将其应用到数控折弯机的多参量非线性补偿中，对多组拟合数据进行对比性曲线拟合以及曲面拟合。结果表明：与传统的复变量移动最小二乘法拟合曲线相比，改进型复变最小二乘法在计算复杂度、平滑性以及精度尤其是点密集区域的精度等方面都取得了较大的改进。%Complex variable moving least squares (MLS) and the important parameters such as basic functions and support do-main radius were discussed. For the lack of its applications on non-uniform sampling point set,an improved complex variable least squares (DRMLS) was presented which support domain radius at fitting point was dynamically adjusted with the thickness of sampling point set. The arithmetic was applied to multi-parameter nonlinear compensation of CNC bending machine with comparative experi-ments between MLS and DRMLS on multiple sets of data. The results show that DRMLS has achieved a larger improvment in computa-tional complexity,smoothness and accuracy,especially accuracy in thick point sets region compared with traditional MLS.