Modified Complete Ensemble Empirical Mode Decomposition based HIF detection approach for microgrid system

Abstract

Detection of high impedance fault (HIF) in an active distribution system is a challenging task. It is learned from the fault characteristics that detection and discrimination of HIF during different critical conditions is impossible using the fault current magnitude. Under dependable situations such as energization of a transformer, nonlinear load and capacitor bank detection and discrimination of HIF is challenging. In a distribution system with an inverter based distributed generator (IBDG), the current contribution during islanding mode is very low and also for HIF condition. To mitigate these issues, an intelligent approach applying Modified Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (MCEEMDAN) on residual current signal is developed. The second intrinsic mode function (IMF2) is extracted using MCEEMDAN and its Teager Kaiser Energy Operator (TKEO) is computed to detect and discriminate the HIF against other physical events. The novelty of MCEEMDAN approach lies with its noise free output and faster response as compared to other time–frequency approaches. The method is tested for several fault and non-fault cases including, presence of noise, harmonics, and unbalance loadings. The comparison with recently reported techniques for very HIF and ungrounded system proofs the efficacy of the method.