quotOPTIMAL POWER FLOW OF AN INTERCONNECTED POWER SYSTEM USING LINEAR PROGRAMMING AND ARTIFICIAL NEURAL NETWORKquot

Abstract

quotThe Optimal Power Flow (OPF) problem was first defined in the early sixties. Over several decades, numerous formulations and solution strategies have been evolved. Various models using Jacobian, P-Q decoupling and constant matrices have been tried. Numerous solution tools such as several nonlinear programming (NLP) techniques, Successive Linear Programming (SLP) technique, Artificial Intelligence (AI) based techniques like Genetic Algorithm and others have been used. Advances in efficient computing methods have been utilized. Efficient matrix storage and computational methods like sparse matrix technique have been employed to develop better OPF methods. newlineIn this thesis, a method is proposed to solve the OPF problem considering multiple objectives of generation cost minimization, transmission loss minimization and voltage stability margin maximization. The OPF is obtained by two ways, the first method is used to optimize the reactive power dispatch in the system by employing Quadratic programming with linearized constraints technique and the second method optimizes the scheduling of generators using back propagation neural network. Economic consideration has been taken into account in deriving the objective function, which is based on the objective function of the optimal power flow with losses expressed in terms of reactive power. Optimizing the reactive power will minimize the system s loss newlineand consequently reduce the total generation cost. These losses are expressed in terms of quadratic function of active and reactive powers. newlineTo improve the performance of the proposed ORPD, two approaches are proposed, firstly, to model all the system s parameters related directly to reactive power as an equivalent var source like the bus voltage constraints or as a fictitious var source like the OLTC taps. Secondly, since the quadratic programming based algorithms are used to solve the ORPD, the bus voltage constraints and line flow constraints are linearized. newlineThis method is coded and tested on standard test systems. The te