Microgrid fault diagnosis model based on Weighted Fuzzy Neural Petri Net

With the extensive application of renewable energy generation, microgrid has become a new focus in power system research. Efficient fault diagnosis is important to ensure the safe operation of the microgrid. Because of the characteristics of high complexity and non-linearity of the microgrid topology, it is difficult to build a universal fault diagnosis model under different operating states. To solve this problem, we partition the microgrid topology into the proximal area and the remote area based on distance from the grid. The fault alarm information is classified into three types of protection information sets: primary protection set, near-backup protection set and far-backup protection set. And more complete diagnostic models are built with adding in the interaction between different protections. At the same time, this paper proposes Weighted Fuzzy Neural Petri Net (WFNPN), which can make diagnostic models by Petri net, and train specific parameters by fuzzy neural network without excessive reliance on artificial experience. And the method can reduce computational complexity and improve the model accuracy. Example analysis results verify the versatility and fault tolerance of the model.