Answers are grounded in this paper and cite their sources.
Model-based methods need to accurately model the sy-stem, which can be difficult with the nonlinear nature of motors. Insulation faults in induction motors (IMs) may result in significant production losses and disruption in an industrial application, so they are important. The aim of this research is to introduce a Fault Detection and Diagnosis (FDD) method based on the Chien-Physics-Informed Neural Networks (C-PINN) approach. The input signals of the Data Challenge 2021 data are pre-processed with Derivative-free Extended Kalman Filtering (DF-EKF) to eliminate noise and normalise the data. The High-Order Synchro-Extracting Transform (HOSET) is used to extract statistical features for improving diagnostics. The model is also enhanced by the Fractional Pelican African Vulture Optimisation (FPAVO) algorithm to overcome the problem of parameter optimisation in the C-PINN. The simulation results in MATLAB show that the proposed method was found to be more accurate (99.3%) and attained a higher true positive rate (0.993) and a lower false positive rate (0.007) than existing methods, and also reduced the computational complexity. Overall, this comprehensive system provides a more reliable and efficient way to detect faults, further improving the safety and durability of induction motors in industrial applications.
No takes yet. Share an insight, caveat, or question.
S et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: