WSEAS Transactions on Systems
Print ISSN: 1109-2777, E-ISSN: 2224-2678
Volume 11, 2012
A Neural-based Gradient Optimization in Large Power Systems
Author:
Abstract: An artificial neural network (ANN) is commonly used as a universal function approximator which is very useful in several problems. The main ability of an ANN is to become similar to the behavior of the system being analyzed through examples acquired in past situations or through experiments. This paper presents a neural-based gradient optimization (NGO) method that applies an ANN in optimization problems in order to approximate the objective function to be minimized. According to this approach, the inputs of the ANN are the decision variables and the output is the objective function. Thus, an NGO uses the ANN topology to adjust the decision variables to find the optimal solution. The NGO is used to optimize a large power system without using an analytical model, and instead only use its historical record of behavior. The NGO was tested with the standard well-known power systems, the IEEE-14 and IEEE-118 bus. The performance of the NGO was compared with that of the traditional gradient-based optimization method.
Search Articles
Keywords: System Optimization, System Identification, Optimal Reactive Power Flow, Artificial Neural Network.