数据驱动下基于元动作的数控车削能耗预测方法

Data-driven Energy Consumption Prediction Method of CNC Turning Based on Meta-action

针对传统的能耗预测模型预测精度较低、泛化性差的问题,提出一种数据驱动下基于元动作的数控车削能耗预测方法。通过分析数控车床的能耗特性,揭示各时段元动作与能耗的关联关系;提出车削元动作及能耗数据的获取方法,并通过元动作状态对数控车床状态进行判断;使用高斯过程回归构建以元动作状态作为输入的数控车削能耗预测模型,实现数据驱动的数控车削加工能耗动态预测。最后,对比3种拟合算法,验证高斯过程回归算法的优越性,并通过比较理论模型和基于元动作的数据模型,以加工某轴类工件为例,验证该方法的有效性。

Aiming at the problems of low accuracy and poor generalization of conventional energy consumption prediction model,a data-driven energy consumption prediction method of CNC turning was proposed based on meta-action.The effects of meta-action on energy consumption were explored under each period of turning via analyzing the energy consumption characteristics of CND turning.A data collection method of meta-action and energy consumption was proposed and the states of turning were judged by the meta-action data.Then,based on the data collected in the processes of CNC machining,energy consumption prediction model was built using Gaussian process regression,which was used to predict dynamic energy consumption of CNC turning.Finally,the superiority of the proposed model was verified by comparing three fitting methods,taking the machining of an axial workpiece as an example,a case study was conducted to validate the practicality of the proposed energy consumption prediction method through comparison with theoretical model.