Unsteady fault diagnosis method for chemical process based on SVM

Support Vector Machines (SVM) have met with significant success in numerous real-world learning tasks. This paper reports our evaluation of SVM on unsteady fault diagnosis for chemical processes such as the CSTR model. We use fixed time series data as the input space, and the target is to classify various pre-determined fault types with high accuracy and high efficiency. The adopted SVM tool is J.C.Platt's (2000) SVM 0.54 (Matlab Toolbox). Experimental results of the CSTR model shows its effectiveness over traditional unsteady fault diagnosis methods.