多目标QPSO-Elman网络下的数控切削参数优化  

Numerical Control Cutting Parameter Optimization Based on Multi-Objective QPSO-Elman Deep Network

作　　者：韩辉辉[1] 付辉 HAN Hui-hui;FU Hui(Chongqing Industry Polytechnic College,Chongqing 401120,China;College of Electrical and Information Engineering,Lanzhou University of Technology,Lanzhou 730050,China)

机构地区：[1]重庆工业职业技术学院,重庆401120 [2]兰州理工大学电气工程与信息工程学院,兰州730050

出　　处：《组合机床与自动化加工技术》2022年第9期116-120,125,共6页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金(62001198);甘肃省青年科技基金计划(21JR7RA247)。

摘　　要：为提高数控切削加工制造水平,科学地配置参数,以实现高效率与低成本的优化目标,提出一种多目标QPSO-Elman网络下的参数优化。首先,在考虑参量设定与条件约束的基础上,给出数控切削多目标参量模型;其次,由于粒子群方法兼容性强,引入量子行为使粒子逐步进化,并设置外部存储集获取各阶段较优的粒子和非支配解,进而结合双承接层Elman网络,提升数据处理能力,从而解决复杂的多目标参数优化问题。采用MATLAB R2022b完成仿真对比实验,结果表明,该方法在定切深与变切深条件下均能够实现较低的加工成本,较高的切削速率与较优的工件表面粗糙度,切削参数设置更为科学。In order to improve the manufacturing level of CNC cutting,scientifically configure parameters,and achieve the optimization goal of high efficiency and low cost,a multi-objective parameter optimization based on QPSO-Elman network was proposed.Firstly,considering parameter setting and condition constraint,a multi-objective parameter model of NC cutting is presented.Secondly,due to the strong compatibility of particle swarm optimization method,quantum behavior is introduced to gradually evolve particles,and external memory set is set to obtain the optimal particle and non-dominant solution at each stage,and then combined with the double-layer Elman network to improve the data processing capacity,so as to solve the complex multi-objective parameter optimization problem.MATLAB R2022 b was used to complete the simulation comparison experiment.The results show that the method can achieve lower machining cost,higher cutting rate and better workpiece surface roughness under the conditions of constant cutting depth and variable cutting depth,and the cutting parameter setting is more scientific.

关 键 词：数控切削 多目标 非支配解 双承接层 数据处理 

分 类 号：TH161[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]