A fuzzy type-2 fault detection methodology to minimize false alarm rate in induction motor monitoring applications

Automatic routines for Fault Detection and Diagnosis (FDD) are very important in industrial monitoring systems. However, false alarms potentially occur. High false alarm rates may lead to outages and consequent losses in the production process. To address such problem, a new FDD strategy based on type-2 fuzzy systems is proposed herein to minimize the false alarm rate. By applying system identification techniques, parametric models are estimated in order to represent the operation of the system under several levels of fault severity. The test system is a detailed dynamic nonlinear model of induction motor drive. The faults considered were partial short-circuit in stator winding coils. A performance comparison was made by implementing the monitoring system with both a type-2 fuzzy system interval and a type-1 fuzzy system. The results obtained thereby showed the improved performance and robustness of type-2 fuzzy system-based monitoring system, which outdoes the performance obtained by a type-1 fuzzy system. Furthermore, the performance of the proposed type-2 fuzzy system-based monitoring system may be further improved by using a Genetic Algorithm for tuning the parameters of the fuzzy type-2 system. •A type-2 fuzzy model-based strategy to fault detection and diagnosis in industrial systems is presented.•This novel approach aims at minimizing false alarm rate, a serious drawback in current methodologies.•By merging type-1 and type-2 fuzzy supervisors, the proposed approach makes a better use of available uncertainties a priori information.•A case study is presented on fault detection and diagnosis in an induction motor system.