论文部分内容阅读
提出一种基于支持向量数据描述(support vector data description,SVDD)的生产过程监控、诊断与优化方法.首先,利用正常样本建立SVDD监控模型,获得控制限;然后,利用贡献图对超过控制限的异常点进行诊断,分析引起异常的主要原因;最后,利用邻近点替换法对异常的生产样本进行工艺参数优化.将新方法应用于热轧薄板的生产过程中,结果表明:新方法比传统的监控方法 T2 PCA具有更高的检出率,且可以实现对异常点的工艺参数优化,使之回到受控状态.
A method of monitoring, diagnosis and optimization of production process based on support vector data description (SVDD) is proposed.Firstly, the SVDD monitoring model is established by using normal samples to obtain the control limit. Then, using the contribution graph, The abnormal point was diagnosed to analyze the main cause of the abnormality.Finally, the process parameters were optimized by using the method of adjacent point replacement.The new method was applied to the production process of hot-rolled sheet.The results show that the new method is more efficient than the traditional method The monitoring method T2 PCA has a higher detection rate and can optimize process parameters for outliers and bring it back to a controlled state.