基于非线性时变Profile的复杂薄壁零件多阶段加工过程波动建模与控制

This project will take the complex thin-walled part multistage machining process fluctuation nonlinear time-varying profile monitor as the project research object，aiming to address the issues such as the volatile multistage machining process, difficult quality control, high probability of occurrence of quality problems, and complex manufacturing processes and so on, and the multistage process fluctuation modeling, monitoring and error traceability will be taken as the project research content, which will be dealt with theoretical derivation, simulation analysis and experimental verification in the multistage fluctuation modeling, monitoring and traceability process. In this project, the relationship between the fluctuation of multi-process process and the parameters of process system, and the evolution mechanism of process fluctuation will be revealed. The non-linear and uncertain influence of thin-walled machining processes on quality fluctuation will be studied, and the multistage machining process fluctuation nonlinear time-varying profile model based on the extend variations of stream will be built, which will describe the fluctuation development and evolution of the machining process. Based on this, the multistage nonlinear time-varying fluctuation profile monitoring model will be designed. At the same time, the variable point theory will be adopted to identify and diagnose process fluctuation change points. And the quality fluctuation change points, the rules of the process change with time and the process out of control can be obtained by the change point diagnosing and recognizing. And the self-adaptive robust control strategy will be proposed. Finally, the process fluctuation traceability model based on causal association analysis and gray correlation analysis will be built, and the joint effect between equipment operation status and process fluctuations will be analyzed. This project can provide theoretical guidance and technical support for the quality fluctuation control of the complex thin-walled part machining process, and the result also can be applied to the scientific research and engineering experiments of the other complex mechanical product machining process error control.

本项目以复杂薄壁零件多阶段加工过程波动非线性时变Profile监控为研究对象，针对多阶段加工过程易波动、质量控制难、精度要求高、制造过程复杂，导致出现质量问题的概率高、质量控制难度大、过程波动不易监控等问题，以过程波动建模、监控及溯源等为研究内容进行理论推导、仿真分析和实验验证，形成以过程监控为手段的复杂薄壁零件加工过程波动的定量预测与精确控制能力，大幅度提高制造质量。通过本项目研究揭示多阶段加工过程波动与工艺系统参数的关联关系以及工序过程波动演化机理，探索多阶段加工过程非线性和不确定性因素对质量波动的影响规律，建立基于扩展误差流的多阶段加工过程波动非线性时变Profile模型；在此基础上，设计基于非线性高维时变Profile的多阶段加工过程波动监控方法，建立多阶段加工过程非线性时变Profile波动监控机制，构建自启动自适应鲁棒控制策略，通过变点诊断与识别得到质量波动的突变点、加工过程随时间发生变化的规则以及过程的失控情况；最终结合误差预测模型、工程领域知识和历史数据，建立因果关联分析和灰关联熵的过程波动溯源模型。本课题可为复杂薄壁零件加工过程质量波动控制提供理论指导和技术支持。

本项目以复杂薄壁零件多阶段加工过程为研究对象，通过对加工过程易波动、质量控制难，导致质量问题发生概率高、过程波动监控和溯源不准确等问题进行剖析和分析，以过程波动建模、监控及溯源等为研究内容进行了产品质量预测的理论推导、数据分析和实验验证。构建了多阶段加工过程波动与工艺系统参数的关联关系，采用扩展的误差流和机器学习模型研究了多误差源因素对复杂薄壁零件加工过程质量波动的影响规律，构建了多阶段加工过程波动演变的时变函数。采用动态滑动时间窗口和聚类算法结合统计过程控制设计了多阶段加工过程监控模型进行质量波动监控，采用多工位和零件批量的时空传递分析，有效识别了产品质量波动的突变点以及过程的失控情况。结合误差时序预测模型、监控模型，采用因果关联分析实现了过程波动的溯源。研究方法在航空发动机叶片、天线薄板零件等对象上进行了深入的验证。研究成果在制造工程领域和工业工程领域的国际知名期刊发表了多篇高水平论文，其中1篇论文入选本领域近十年ESI TOP 1%高被引论文。