Fault Detection Of Process Industry Based On PCA

With the rapid development of science and technology,the complexity of process industry is getting higher and higher.Once a fault occurs,it is likely to cause a chain reaction,resulting in significant losses.Therefore,it is of great value to carry out real-time monitoring to detect faults quickly and accurately.According to the process industry features such as time-varying,nonlinear and large amount of data,an improved algorithm is proposed based on analyzing the ability of several existing fault detection algorithms.Moreover,simulation and verification on TE platform are carried out:(1)The fault diagnosis effect of PCA in the TE process is analyzed,and the simulation results show that PCA can not detect faults well because of the nonlinear and time-varying characteristics.(2)DKPCA is suitable to deal with dynamic,nonlinear problems,but for large sample data set,it occupancy a lot of computer memory and large computation,in order to solve these problems,this paper proposed an improved DKPCA based on effective feature subspace(EFS-DKPCA),which based on a orthonormal basis of the sub-space spanned by the training samples mapped onto the smaller feature space to optimize the structural representation of the sample in the feature space.The experimental results show that the algorithm can effectively improve the efficiency of fault detection,and the correct rate is almost unaffected.(3)Considering the time-varying characteristic of parameters in the process industry,an improved DKPCA method based on block(BDKPCA)is proposed,it using a time-varying kernel matrix instead of a fixed kernel matrix to establish the principal component model,which was suitable for online model updating.Experimental results show that the proposed algorithm has better detection performance than DKPCA.(4)Considering the effect of noise and detection of aging problems,we first use wavelet to denoise the data before the fault detection,and then combine the EFS-DKPCA and BDKPCA algorithm to detect the faults of TE process.The simulation results show that the algorithm is superior to PCA?KPCA?DKPC A?EFS-DKPCA?BDKPCA.