Research on a Semi-Quantitative Fault Prediction Method of Systems with Uncertainty

The problem formulation of semi-quantitative fault prediction is given firstly in this paper.Based on one step and multi-step Fuzzy Kalman predictors and the rule of fault prediction,a semi-quantitative fault prediction method is proposed to predict the soft faults of a class of systems with uncertainty.The accurate analytic models and the statistical characteristics of noises,which are necessary in the traditional methods,are not used in this method.In addition,the noises are depicted as the fuzzy qualitative interval, so this method has some robustness.Finally,the simulation results demonstrate the effectiveness of the proposed approach,and the possible research direction in the future is given.