Design of an Artificial Neural Network to Improve the Stability of Smart Power Grids

The problem addressed in this paper is related to improving the stability of the operation of power grids that comprise distributed nodes that could be perturbed from one or more than one location. The proposed approach is to design an artificial neural network (ANN) that estimates the state of stability of the terminal voltages of regulators and converters connected to experimental power grids. The ANN learns from the terminal voltage values of individual nodes in the power grid and classifies their operation as disturbed or not disturbed based on a decision chart. The approach could reveal the presence of both single and multiple disturbances in the power grid using real-time communication with distributed sensors. The results could be used by system operators to adjust the variables of voltage regulators and converters such as the gains of such devices to improve the stability of smart grids.