Intelligent Fault Diagnosis Methodology Based on Improved Hybrid Kernel Extreme Learning Machine

In recent years, smart grids' rapid development and application have attracted widespread attention. Although cyberphysical power systems (CPPS) have been operating successfully in several geographic regions, there is still room for improvement in ensuring system stability, reliability, and security. In power systems, many mechanical devices such as generators, transformers, and motors depend on the proper functioning of bearings. To monitor the health status of bearings, this paper presents a fault diagnosis model that combines an Improved Pelican Optimization Algorithm (IPOA) with Hybrid Kernel Extreme Learning Machine (HKELM) for the Variational Mode Decomposition (VMD) and incorporates Refined Composite Multi-scale Fuzzy Entropy (RCMFE) preprocessing based on the XJTU-SY bearing dataset. The proposed method is experimentally compared with different models, and the results demonstrate that the VMD-RCMFE-IPOA-HKELM bearing fault classification model in this paper outperforms the other models in terms of accuracy (98 %), KAPPA coefficient (97.5 %), and MAF1 value (98.0 %), showing more accurate and reliable performance.