Developing neural networks library in RSCAD for real-time power system simulation

Abstract

Real-time digital simulators (RTDS) are used for real-time simulation of power systems. As the complexity of the electric power grid and associated control increases in the future, modeling and simulation of the power network as well as the control becomes essential. This requirement will be even more prominent in the context of smart grid. As enabling technology, intelligent methods of monitoring and control that utilize computational intelligence techniques are expected to be an integral part of smart grids. Therefore integration of computational intelligence based tools in power system simulation tools is an important aspect of smart grid research. Most past and current applications of neural networks in power systems are carried out offline in non-real-time platforms. In this study, neural networks libraries are developed to run on RTDS. The neural networks component is then used to predict the speed deviation of a generator in a multi-machine power system. These neural networks components can be trained in real-time and hence can be useful tool for smart grid applications.