论文部分内容阅读
针对传统PCA故障检测算法的结果有定论不明确的缺陷,提出一种基于Q统计量分离的故障检测新方法,把Q统计量分为PVR和CVR,前者代表显著与主元有关的变量信息,后者代表与主元无明显关系的变量信息,再配合T~2统计量共同用于监测过程,检测效果更细致。将此方法结合基于累积方差贡献率(CPV)和复相关系数(MCC)确定过程监测模型主元数的新方法,监测β-甘露聚糖酶发酵工业的过程,与传统的PCA故障检测方法比较,仿真研究结果表明该算法能够确保主元空间(PCS)中的信息存量,充分刻画过程变化,有效识别正常工况变化与故障,正确检测微弱故障,提高过程监控的准确性。
Aiming at the shortcomings of the traditional PCA fault detection algorithm, a new fault detection method based on Q statistics separation is proposed. The Q statistics are divided into PVR and CVR, the former represents significant variable information related to the principal component, The latter represents the variable information which has no obvious relationship with the main element, and then used together with T ~ 2 statistics to monitor the process, the detection effect is more detailed. This method was combined with a new method to determine the main elements of the process monitoring model based on the cumulative variance contribution (CPV) and the complex correlation coefficient (MCC) to monitor the β-mannanase fermentation industry process, compared with the traditional PCA fault detection method The simulation results show that the algorithm can ensure the information inventory in the principal component space (PCS), fully characterize the process changes, effectively identify the changes and failures of normal working conditions, correctly detect weak faults and improve the accuracy of process monitoring.