Nu-support vector regression model implementation for distributed generation siting and sizing

One of the most important things in improving the performance of an electric grid is the placement and sizing of distributed generation (DG) units. Installing the ideal DG size at the ideal locations has been shown to minimise power loss in an electrical network in addition to improving the voltage stability index. This paper evaluates the performances of four simple machine learning algorithms in determining the optimal size and location of a distributed generator (DG) for a test system. An altered version of the IEEE-30 bus test network serves as the test system under consideration. Close evaluation of the results show that the performance of nu-support vector regression (nu-SVR) closely matches the manually obtained output using MATLAB PSAT.