BP神经网络在立铣刀结构参数优化中的应用

摘要

钛合金薄壁件的铣削加工过程中,刀具磨损速度快,并且工件容易变形,其主要因素是加工过程中切削力大,切削温度高.文章利用有限元仿真软件AdvantEdge FEM铣削仿真数据,建立整体式立铣刀结构参数与切削力和切削温度的BP神经网络预测模型,并对切削预测模型进行了切削实验验证.在此基础上,利用BP神经网络模型的预测结果对整体式立铣刀的结构参数进行了优化,切削实验证明,优化后的刀具参数可以有效地降低切削力和切削温度,从而有效地改善过程中刀具的切削性能和工件的加工质量.%The cutting force is very large and the cutting temperature is very high in the process of milling a titanium alloy thin-walled part, so the wear rate of milling tool is very fast and the part is deformed easily.In this paper, firstly, the sample data is obtained through simulating the process of milling titanium alloy thin-walled parts by AdvantEdge FEM.Secondly, the BP neural network model between cutting force and temperature with structure parameter of the solid end mill has been built.Thirdly, the credibility of BP neural network model is verified.Finally, the structure parameter of the solid end mill has been optimized through the data from BP neural network prediction model.The cutting force and temperature will be reduced after optimization of the structure parameter of milling tool.The performance of cutting tool and the quality of parts all have been improved.