机理驱动的数控机床进给轴时变误差建模和补偿方法

Mechanism-driven Method for Time-varying Error Modeling and Compensation of CNC Machine Tool’s Feed Axes

目前数控机床时变误差实时补偿常采用的数据驱动模型,存在模型鲁棒性差、所需传感器数量多、工程实际应用受限等问题。为此,基于机理驱动建模方法建立了数控机床进给轴的时变误差预测模型。该模型可预测出丝杠在任意运动中的实时温度场。仅需在进给轴丝杠旁的床身上布置一个温度传感器,用于采集丝杠的初始温度以及加工时的实时环境温度。补偿器基于TCP/IP协议与数控系统进行通讯,实时读取机床坐标并写入补偿量。基于国家科技重大专项的“换脑工程项目”,在某航发企业对使用十年以上的立式加工中心HAAS VF-2进行了进给轴时变误差测试和补偿。结果表明,在丝杠存在磨损老化和热变形的情况下,提出的时变误差补偿模型仍然取得了很好的效果,机床的精度稳定性得到明显提升。

At present,the data-driven model used for real-time time-varying error compensation of CNC machine tool has some problems,such as poor robustness of the model,a large number of sensors required and limited practical application in engineering.For this reason,a time-varying error prediction model for feed axes of CNC machine tool was established based on mechanism-driven modeling method.The real-time temperature field of screw in any motion is predicted by this model.Only one temperature sensor is placed on the body near the screw to acquire the initial temperature of screw and the real-time environmental temperature in machining.The compensator communicates with CNC system based on TCP/IP protocol,which can read the coordinates of the machine tool from CNC system and write the value of compensation into CNC system in real time.The test and real-time compensation of time-varying was realized for feed axes of a vertical machining center(HAAS VF-2),which has been used for more than 10 years,based on“Brain Replacement Engineering”of National Science and Technology Major Projects.The results show that the time-varying error compensation model proposed in this paper still acquired good results when the screw was worn and aging with thermal error,and the accuracy and stability of the machine tool had been significantly improved.