基于贝叶斯算法的铣削工艺参数配置优化方法  

Optimization Method of Milling Process Parameter Configuration Based on Bayesian Algorithm

作　　者：张志强 张为民[1,2] 薛峰 陈圣华 马瑞 赵东方 Zhang Zhiqiang;Zhang Weimin;Xue Feng

机构地区：[1]同济大学机械与能源工程学院,上海201804 [2]同济大学先进制造技术中心,上海201804 [3]台达电子工业股份有限公司,中国台湾省台北114501 [4]台达智能科技(北京)有限公司,北京100026

出　　处：《机械制造》2022年第7期19-25,50,共8页Machinery

基　　金：台达智能科技(北京)有限公司委托同济大学科研合作项目。

摘　　要：随着个性化制造需求的增加和产品生命周期的缩短,制造商需要能够针对新产品进行工艺参数快速适配,以响应市场变化。在金属铣削加工中,新产品的工艺参数寻优效果和效率决定了加工质量和交付时间。针对现有铣削工艺参数寻优方法试验次数多且调优效果欠佳的问题,提出基于贝叶斯算法的铣削工艺参数配置优化方法。建立以高斯过程回归为概率代理模型,以置信下界策略为采集函数的贝叶斯算法,以多个加工质量的归一质量评价指标为优化目标,以多个工艺参数为优化变量,在实际生产中对铣削工艺参数配置问题进行验证。试验结果表明,与田口法相比,基于贝叶斯算法的铣削工艺参数配置优化方法寻优次数减少58.8%,优化效果提升9.3%,由此验证这一方法在提升铣削工艺参数配置优化效果的同时可以极大缩短配置周期,降低寻优成本。With the increasing demand for personalized manufacturing and the shortening of product life cycle, manufacturer needs to be able to quickly adapt process parameter for new product to respond to market change.In metal milling, the optimization effect and efficiency of process parameter of new product determines the machining quality and delivery time.Aiming at the problem that the existing optimization method of milling process parameter has many tests and poor tuning effect, an optimization method of milling process parameter configuration based on Bayesian algorithm was proposed.A Bayesian algorithm with Gaussian process regression as the probability surrogate model and with the confidence lower bound strategy as the acquisition function was established.Taking normalized quality evaluation index of multiple machining qualities as the optimization goal and multiple process parameters as optimization variables, the configuration of milling process parameter was verified in real production.The test results show that, compared with Taguchi method, the optimization method of milling process parameter configuration based on Bayesian algorithm reduces the number of optimizations by 58.8%, and the optimization effect increases by 9.3%.Therefore, it is verified that this method can greatly shorten the configuration cycle and reduce the optimization cost while improving the optimization effect of milling process parameter configuration.

关 键 词：贝叶斯算法 铣削 参数 配置 优化 

分 类 号：TG54[金属学及工艺—金属切削加工及机床]