摘要
Abstract
Ultra precision machining technology is a key technology in the high-end manufacturing field.Mastering the key technology of ultra precision machining error control,ensuring and improving the machining accuracy of CNC machine tools,has become a research hotspot for improving the level of machining and manufacturing.The error caused by thermal deformation is one of the main factors affecting the accuracy of CNC machine tools.In order to further improve the accuracy of thermal error prediction for machine tools,this paper proposes a numerical control machine tool thermal error prediction model based on differential fusion long short-term memory neural network(DF-LSTM).This model introduces differential prediction and combines it with direct prediction,aiming to effectively balance the trend and volatility of the data to be predicted.At the same time,in response to the problem that there are many neural network algorithms but there is still a gap in compensation effectiveness,this paper designs a thermal error optimization model(GA-BPNN)combining genetic algorithm(GA)and BP neural network(BPNN),which utilizes GA's global search ability to optimize the initial weights and thresholds of BPNN,thereby significantly improving compensation accuracy and model convergence speed.The simulation results show that DF-LSTM can improve the accuracy of error prediction;GA-BPNN can reduce the thermal error of molds.关键词
电子器件/微结构模具/超精密加工/误差补偿/动态控制Key words
electronic device/microstructure mold/ultra precision machining/error compensation/dynamic control分类
矿业与冶金
