基于WOA-BP算法的氟金云母钻削工艺参数优化  

Modeling and Optimization of Drilling Process Parameters of Fluorophlogopite Ceramics Basedon Improved BP Neural Network with WOA Algorithm

作　　者：戴春雨 马廉洁[1,2] 孙德谦 李红双 陶其赫 DAI Chunyu;MA Lianjie;SUN Deqian;LI Hongshuang;TAO Qihe(School of Mechanical Engineering and Automation,Northeastern University,Shenyang 110819,China;School of Control Engineering,Northeastern University at Qinhuangdao,Qinhuangdao 066004,China)

机构地区：[1]东北大学机械工程与自动化学院,沈阳110819 [2]东北大学秦皇岛分校控制工程学院,秦皇岛066004

出　　处：《组合机床与自动化加工技术》2024年第1期135-139,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金项目(51975113)。

摘　　要：通过氟金云母陶瓷钻削实验,测试了在不同加工参数下的材料去除量和刀具磨损量。利用WOA算法优化BP神经网络,并基于单因素实验值和WOA-BP网络预测值,利用最小二乘法拟合,建立了材料去除率和刀具磨损率关于工艺参数的一元模型,以相关系数检验了模型的精确度。在一元模型的基础上提出了多元模型,基于正交实验值和WOA算法对多元模型进行求解,模型误差在合理范围内。以材料去除率最大和刀具磨损率最小为优化目标,基于WOA算法进行了工艺参数双目标优化,得到了一组最优参数。基于最优工艺参数进行验证实验,实验结果表明得到的最优参数是合理的。The material removal and tool wear under different processing parameters were measured by drilling experiments of fluorophlogopite ceramics The WOA algorithm is used to optimize the BP neural network.Based on the single factor experimental value and the prediction value of the WOA-BP network,a unitary model of the material removal rate and tool wear rate with respect to the process parameters is established by using the least square fitting method,and the accuracy of the model is checked by the correlation coefficient a multivariate model is proposed on the basis of the unitary model.The multivariate model is solved based on orthogonal experimental values and WOA algorithm.The model error is within a reasonable range taking the maximum material removal rate and the minimum tool wear rate as the optimization objectives,the WOA algorithm was used to optimize the process parameters and a set of optimal parameters were obtained Based on the optimal process parameters,the validation experiment results show that the optimal parameters are reasonable.

关 键 词：钻削加工 工艺参数 WOA算法 BP神经网络 双目标优化 

分 类 号：TH16[机械工程—机械制造及自动化] TG52[金属学及工艺—金属切削加工及机床]