六维力传感器动态误差建模与分析

Thermal Error Modeling and Analysis of CNC Machine Tool

针对线性解耦算法不能有效解决六维传感器动态耦合问题,对六维力传感器动态误差研究进行建模与分析。建立PNN-APSO数学动态解耦模型,通过引入自适应惯性权重来更新粒子速度,进一步增强粒子的整体搜索能力,有效的修正网络连接权值函数等参数,网络被不断优化,有效提高网络的鲁棒性和有效性。结果表明,模型损失值在0.3左右,准确率在0.92~0.98内,该算法可降低传感器实际动态测量环境下的维间耦合误差,满足行业对六维力传感器高精度的检测需求。

In view of the fact that the linear decoupling algorithm cannot effectively solve the dynamic coupling prob-lem of six-dimensional sensors,the dynamic error research of six-dimensional force sensors is modeled and analyzed.The PNN-APSO mathematical dynamic decoupling model is established,the particle velocity is updated by introducing adaptive inertia weights,the overall search ability of particles is further enhanced,and the parameters such as the net-work connection weight function are effectively corrected,and the network is continuously optimized,which effectively improves the robustness and effectiveness of the network.The results show that the loss value of the model is about 0.3 and the accuracy is within 0.92~0.98,which can reduce the interdimensional coupling error in the actual dynamic mea-surement environment of the sensor and meet the industry’s high-precision detection requirements for six-dimensional force sensors.