GrC-based real-time fault predictive diagnosis method

Traditional diagnostic systems generally are off-line and confirmatory. When abnormal variables occur, they cannot provide predictive diagnosis and maintenance suggestions in time. So the anomaly may be further accumulated and lead to more serious faults. This paper proposes a real-time fault predictive diagnosis method, which is based on granular reduction algorithm and granular similarity calculation. By using this method, the predictive fault diagnosis of a real-time system can be changed into a set of simple mathematical operations. Finally, the validity and feasibility of the method are illustrated with a real example.