基于ICA-PCA和Lasso的过程故障诊断

Process fault detection and diagnosis based on ICA-PCA and Lasso

为了解决复杂工业过程中变量多,难以判断引起故障的主要异常变量的问题,提出一种基于ICA-PCA(独立成分分析和主成分分析)算法和Lasso(最小绝对收缩和选择算子)回归算法的过程故障检测与诊断的集成模型.首先,建立ICA-PCA模型提取数据的高斯信号和非高斯信号,构造相关统计量实现在线故障检测;然后,基于ICA-PCA模型获得的过程状态及故障信息,进一步构造基于Lasso回归算法的故障诊断模型,实现故障发生时的主要异常变量的定位和选择;最后,利用Matlab进行了TE(田纳西-伊斯曼)过程的数值仿真实验,并与已有故障诊断方法分布式PCA贡献图法进行比较,结果表明所提出的方法是有效的.

In order to solve the complex industrial process variables, it is difficult to judge caused by failure of the main abnormal variables, based on ICA-PCA （independent component analysis and prin cipal component analysis） algorithm and Lasso（least absolute shrinkage and selection operator） regression algorithm of fault detection and diagnosis of integrated model was proposed. First, ICA-PCA model was established to extract the data of the Gaussian signal and the non Gaussian signal, struc- ture related statistics for online fault detection; then, based on ICA PCA model the process state and fault information were obtained, further structure based on Lasso regression algorithm was estab- lished for fault diagnosis model and the orientation and choice of the fault occurs were realized when the main abnormal variables. Finally, the numerical simulation experiment of TE （Eastman Tennes see） process was carried out by using the simulation software Matlab, and the results were compared with that of the existing fault diagnosis method for distributed PCA with diagram method. The results show that the proposed method is effective.