An Entropy-Based Hybrid Conjugate Gradient Scheme for Locating High Impedance Faults in DC Microgrids

Abstract

This Paper proposes an efficient inversion scheme based on Shannon entropy to determine the location of high impedance faults (HIFs) in DC microgrids in the presence of noise. The proposed scheme adopts a phenomenological approach that uses a physical model to iteratively solve the inverse problem. The problem is posed as an optimisation problem that determines the most likely location of a fault by minimising an appropriate objective function; it denotes the disparity between the entropy of the current waveforms observed by the protection relays at the time of fault detection and their model-predicted counterparts. A hybrid optimisation technique based on the conjugate gradient method is adopted in the inversion process that secures a robust and fast convergence of the proposed method. The primary advantage of this scheme is its accuracy in noisy environments with no limitation on the type of fault and value of fault impedance. To examine the performance of the proposed scheme, a mesh-type DC microgrid is simulated and the results are compared with one of the state of-the-art methods. It has been shown that the proposed method can accurately locate HIFs with fault resistance as high as 100 Ω and signal-to-noise ratio as low as 20 dB.