基于改进IDW法的铣削机器人模态参数预测  

Modal parameter prediction of milling robot based on improved IDW method

作　　者：段玉斌 赵一阳 刘金玉 曹家彬 闫维新[1] DUAN Yubin;ZHAO Yiyang;LIU Jinyu;CAO Jiabin;YAN Weixin(School of Mechanical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区：[1]上海交通大学机械与动力工程学院,上海200240

出　　处：《中国测试》2025年第1期159-165,180,共8页China Measurement & Test

基　　金：国家重点研发计划(2017YFB1300603)。

摘　　要：铣削加工工业机器人模态参数识别是分析其加工时振动现象的基础,模态参数与机器人的姿态高度相关,因此存在加工过程中其随时间改变的现象。传统的有限元法以及试验模态分析法难以快速准确地预测铣削机器人加工空间内所有姿态下的模态参数,对此文章提出一种改进的距离反比加权(IDW)频响曲线预测法,实现机器人任意姿态下的频响曲线预测,并采用正交多项式曲线拟合法准确辨识机器人的模态参数。首先对铣削加工系统进行有限元模态仿真,在此基础上确定传感器布放位置;采用力锤冲击试验得到加工平面内有限位姿点的频响曲线,之后利用改进IDW法对部分位姿点的频响曲线进行预测;最后利用正交多项式曲线拟合法基于预测频响曲线实现模态参数的辨识并与实测模态参数进行对比,分析结果表明:1)改进IDW法可有效提高频响曲线的预测精度;2)基于预测频响曲线的模态参数识别结果与基于实测频响曲线的参数识别结果接近,低阶固有频率的预测误差在8%以内。The modal parameter identification of milling industrial robots is the basis for analyzing their vibration phenomena during machining.The modal parameters of the robot are highly correlated with its posture,so there is a phenomenon that their modal parameters will change over time during the machining process.Traditional finite element methods and experimental modal analysis methods are difficult to quickly and accurately predict the modal parameters of all postures in the machining space of milling robots.Therefore,this paper proposes an improved inverse distance weighted(IDW)frequency response curve prediction method,which realizes the frequency response curve prediction of robots under any posture.The orthogonal polynomial curve fitting method is used to accurately identify the modal parameters of the robot.Firstly,finite element modal simulation was conducted on the milling system,and based on this,the sensor placement position was determined.Secondly,hammer impact test is carried out to obtain frequency response curves at different positions in the machined plane.Then the improved IDW method is used to predict the frequency response curves of some position points.Finally,based on the predicted frequency response curve,the modal parameters are identified by the orthogonal polynomial curve fitting method and compared with the experimental modal parameters.The analysis results show that:1)The improved IDW method effectively improves the prediction accuracy of frequency response curve;2)The results of modal parameter identification based on predicted frequency response curve are close to those based on experimental frequency response curve,and the prediction error of low-order natural frequency is within 8%.

关 键 词：铣削机器人 频响函数预测 模态参数辨识 改进距离反比加权法 

分 类 号：TB9[一般工业技术—计量学]