数据驱动的离散制造系统性能退化机理建模与预测控制方法研究  

Research on Data-driven Performance Degradation Mechanism Modelling and Predictive Control Method for Discrete Manufacturing System

作　　者：王文波 张映锋[2] 顾寄南[1] 张耿[2] 完严[1] WANG Wenbo;ZHANG Yingfeng;GU Jinan;ZHANG Geng;WAN Yan(School of Mechanical Engineering,Jiangsu University,Zhenjiang 212013;Key Laboratory of Industrial Engineering and Intelligent Manufacturing,Ministry of Industry and Information Technology,Northwestern Polytechnical University,Xi'an 710072)

机构地区：[1]江苏大学机械工程学院,镇江212013 [2]西北工业大学工业工程与智能制造工信部重点实验室,西安710072

出　　处：《机械工程学报》2024年第16期400-411,共12页Journal of Mechanical Engineering

基　　金：国家自然科学基金资助项目(52105516)。

摘　　要：异常事件的主动感知与优化控制是制造系统可靠运行的关键。针对离散制造系统中生产环境不确定性强、关键性能退化规律复杂和数据云端融合时间长等带来的异常识别滞后与优化响应困难的问题,提出一种数据驱动的离散制造系统性能退化机理建模与预测控制方法体系。通过对基于边云协同的离散制造系统智能环境配置与性能提取、基于跨时空状态数据的制造系统关键性能退化机理建模、融合运维知识与预测性能的制造异常预警与主动优化控制等关键技术的深入分析和研究,建立一种具有边云协同交互决策能力的智能制造系统,并实现异常事件事前预警与主动自适应优化决策。所提体系架构和关键技术可以为下一代“智能工厂”的预测性控制方案的落地应用提供重要基础理论和技术支撑。The active sensing and optimal control of abnormal events are the core issues to ensure the reliable operation of discrete manufacturing system.For a discrete manufacturing system,many uncertain factors exist in the production system,the degradation patterns of production performance are hard to describe,and the cloud-based data fusion methods consume a long time,as a result,the detection of abnormal events often delays and the optimal decisions are hard to find.To meet these requirements,it is provided a data-driven production performance degradation mechanism and predictive control method for discrete manufacturing system.Firstly,the industrial Internet of Things and Cyber Physical System technologies are combined to establish a cloud-edge cooperation environment and extract the key production performance information.Secondly,the degradation mechanism of production performance is modelled based on the fusion of spatial-temporal manufacturing data.Thirdly,the operational knowledge and predicted performance of manufacturing system will be combined to present an exception early-warning and proactively optimal control method.Then,the capacity of cloud-edge cooperation,early-warning of production exceptions and predictive and optimal control of smart discrete manufacturing system can be achieved.The proposed strategy,method and model provide the important support and technical reference for the predictive control of next generation smart factory.

关 键 词：智能制造 自适应优化 离散制造系统 性能退化 预测控制 

分 类 号：TP39[自动化与计算机技术—计算机应用技术] TP301[自动化与计算机技术—计算机科学与技术]