A novel hybrid downsampling and optimized random forest approach for islanding detection and non‐islanding power quality events classification in distributed generation integrated system

The quality of power in modern-day system is polluted with increased penetration converter-based distributed generations such as wind farm, solar PV system. In scenarios detection islanding and disturbances well the removal these from quite crucial for equipment maintenance personnel safety. Here, a down-sampling empirical mode decomposition (DEMD) optimized random forest (RF) machine learning approach are hybridized to detect conditions reduced non-detection zone (NDZ) classify non-islanding events highly energy penetrated distribution generation DEMD has special ability filter out fundamental signal an unbiased non-linear approach. Moreover, improved grey wolf optimization technique proposed optimize parameter RF. simulated MATLAB/Simulink IEEE 13-Bus test grid. efficacy method evaluated through comparative analysis existing techniques under normal noisy environments validated narrow NDZ lesser time.