机床电主轴数字孪生模型构建及更新策略研究  

Research on the Construction and Update Strategy of Machine Tool Electric Spindle Digital Twin Model

作　　者：周正飞 张凤丽 牛晓彤 王金江[1] ZHOU Zheng-fei;ZHANG Feng-li;NIU Xiao-tong;WANG Jin-jiang(College of Mechanical and Transportation Engineering,China University of Petroleum,Beijing 102249)

机构地区：[1]中国石油大学(北京)机械与储运工程学院,北京102249

出　　处：《制造业自动化》2025年第3期100-109,共10页Manufacturing Automation

摘　　要：数字孪生模型的高保真性对于机床在智能应用中的表现至关重要。电主轴在数字孪生建模时存在复杂耦合关系难描述以及敏感参数易变的问题,模型的响应与真实物理系统出现一定误差,导致基于孪生模型的各阶段仿真分析存在局限性。提出一种基于贝叶斯推断的孪生模型更新方法。首先综合各学科知识构建基于多领域统一建模语言的电主轴孪生模型;其次分析运行过程中参数变化机理,引入贝叶斯推断方法构建孪生模型更新策略,结合实际运行数据与先验知识,通过变分推断求解贝叶斯后验分布;最后通过机床电主轴实例验证了模型更新方法的有效性,更新后模型仿真误差达到5%以下,有效提高了孪生模型保真度。The high fidelity of digital twin models is crucial for the performance of machine tools in intelligent applications.The electric spindle encounters problems of complex coupling relationships that are difficult to describe and sensitive parameters that are susceptible to change during digital twin modeling,which causes certain errors between the model's response and the real physical system,resulting in limitations in simulation analysis of various stages based on twin models.This article proposes a twin model updating method based on Bayesian inference.Firstly,a twin model of electric spindle based on multi domain unified modeling language is constructed by integrating knowledge from various disciplines;Secondly,the mechanism of parameter changes is anylized during operation,while Bayesian inference method is introduced to construct a twin model update strategy,and solve Bayesian posterior distribution is solved through variational inference by combining actual operation data and prior knowledge;Finally,the effectiveness of the model update method is verified through an example of the machine tool electric spindle.After the update,the simulation error of the model reaches below 5%,effectively improving the fidelity of the twin model of the electric spindle.

关 键 词：电主轴 数字孪生 更新策略 贝叶斯推断 

分 类 号：TH166[机械工程—机械制造及自动化] TG659[金属学及工艺—金属切削加工及机床]