DISTRIBUTION GRID FAULT ANALYSIS UNDER LOAD AND RENEWABLE ENERGY UNCERTAINTIES

A versatile intelligent fault diagnosis (IFD) method for a distribution grid integrating renewable energy resources is described. Advanced signal processing techniques extract useful features from recorded three-phase current signals, which are input to a multilayer perceptron neural networks (MLP-NN) to diagnose i.e., to detect, classify, identify the feeder branch, and locate the faults. Once a fault is detected, classified and located, a grid operator may adjust grid parameters or dispatch a technician to correct the fault. The IFD method is independent of load demand, renewable energy generation, and fault information (resistance and inception angle) uncertainties, as well as measurement noise.