The research of fault diagnosis method based on weighted Q contribution plot and SDG

Aiming at the disadvantage of the traditional contribution plot of PCA that could not be directly used to locate the fault, a fault diagnosis method based on weighted Q contribution plot and SDG is proposed. First, weighted Q contribution plot is used to locate the fault. Considering weighted Q contribution plot is a pure data analysis method which still exists the risk of leading a incorrect diagnosis result, in order to improve the accuracy of the SDG model we construct the SDG model based on the transfer entropy. Finally, the SDG fault reasoning method is adopted to verify and correct the diagnosis result from weighted Q contribution so as to improve the accuracy rate of the fault diagnosis result. The simulation results based on the Tennessee Eastman process show the effectiveness of the proposed method.