指数加权主元分析法及其在故障诊断中的应用

摘要

在研究主元分析法( PCA)理论的基础上，提出指数加权主元分析( EWPCA)算法。这种算法通过不断更新相关矩阵来实时监视动态生产过程中的超时趋势和设定点改变等状态。实验结果表明，该方法可以较好地反映生产过程中的实时信息，并能有效检测出系统的异常状况，具有广阔的实际应用前景。%Exponential weighted PCA ( EWPCA ) theory have been further studied based on PCA theory in this paper. This theory is applied to the monitoring processes where drift occurred over time and changes of setting point took place by improving the correlation matrix. The simulations show that this way can better describe the real-time information in production process, detect anomalies effectively, and it has a wide application in industrial processes.