Optimal Power Pooling for a Multiple Area Power System through PSO

Abstract

In a multiple area power system power can be transferred from one area to other to improve the load factor, reliability, security and economics of the power system. The generation cost of each unit in an area and the cost of power transfer from other area are given. For a given load in an area the problem is how much power to be generated internally by all the generating units in the area (unit wise) and how power has to be transferred (pooled) from other area for a given total load demand in the area (for economic operation of the power system). This problem has been solved using Lagrange multiplier method recently. The limitation of this method is that it is applicable only if the generation cost of each generating unit is quadratic. Sometimes the power generation cost of each unit is not quadratic but is other type of nonlinear function e.g. the valve-point effect problem it is addition of quadratic and sinusoidal function. The main objective of this paper is to overcome this limitation by solving this general type of nonlinear optimal problem using a Meta heuristic method, PSO (Particle Swarm Optimization) and its variants. It is to point out that for implementation of PSO it is required to select a population size. Generally the population size is selected on ad- hoc basis. This effects the computation time as well as number of iteration for solution. A method has been suggested to select population size on the basis of optimal computation time as second objective. The method is explained by an example. The results obtained by PSO and its variants are compared among themselves and with the results obtained by analytical method. MATLAB 7 software is used for computation.