Cellular Multilayer Perceptron for Prediction of Voltages in a Power System

With the increase in renewable energy sources connected to the power grid, better identification tools are needed for power system voltage profile prediction. Current deregulation trends have led to voltages operating close to stability limits which increases the need for quick estimation tools for system security and contingency analysis. This paper presents a cellular multilayer perceptron (CMLP) architecture for fast identification and prediction of bus voltages. The CMLP method is compared with a standard MLP neural network for bus voltage prediction on the 12-bus three-area test power system. CMLPs can represent a direct mapping of any power system simplifying the equations and allowing for easy scalability to large power systems.