Evolutionary programming based optimal power flow and its validation for deregulated power system analysis Yog Raj Sood * Department of Electrical Engineering, National Institute of Technology, Hamirpur (H.P.) 177 005, India Received 16 April 2004; received in revised form 16 December 2005; accepted 23 March 2006 Abstract Optimal power flow (OPF) has been widely used in power system operation and planning. In deregulated environment of power sec- tor, it is of increasing importance, for determination of electricity prices and also for congestion management. The classical methods are usually confirmed to specific cases of the OPF and do not offer great freedom in objective functions or the types of constraints that may be used. With a non-monotonic solution surface, classical methods are highly sensitive to starting points and frequently converge to local optimal solution or diverge altogether. This paper describes an efficient evolutionary programming based optimal power flow and com- pares its results with well known classical methods, in order to prove its validity for present deregulated power system analysis. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Optimal power flow; Evolutionary programming; Deregulation; Steepest descent method; Genetic algorithm 1. Introduction The OPF optimizes a power system operating objective function, while satisfying a set of system constraints. In general, OPF problem is a large dimension nonlinear, non-convex and highly constrained optimization problem. It is non-convex due to existence of nonlinear AC power flow equality constraints, non-convex unit operating cost functions and units with prohibited operating zones. This non-convexity is further increased when valve point load- ing effects of the thermal generators have to be included [16] or FACTS devices are included in the network. Many classical techniques have been reported in the liter- ature [9–12], such as nonlinear programming (NLP), qua- dratic programming (QP) and linear programming (LP). The gradient based methods [5,12] and Newton methods [15] suffer from the difficulty in handling inequality con- straints. Moreover, these NLP and QP methods rely on convexity to obtain the global optimum solution and as such are forced to simplify relationships in order to ensure convexity. To apply linear programming [2], input–output function is to be expressed as a set of linear functions, which may lead to loss of accuracy. Moreover they are not guar- anteed to converge to the global optimum of the general non-convex OPF problem. These days, genetic algorithm (GA) [3,7,8,16,19] and evolutionary programming tech- niques (EP) [6,17,18,20] has been suggested to overcome the above-mentioned difficulties of classical methods. In these days, an evolutionary programming approach has been used to solve OPF for the analysis of deregulated model [13,14]. So it is necessary to validate the proposed approach with the help of well known basic classical tech- nique likes gradient steepest descent method. In this paper, OPF algorithm of three approaches, steepest descent method, GA and EP have been developed and applied to IEEE-30 bus test system and their results are compared. In order to, further confirm the validity; the results of EP are also compared with results obtained using matlab opti- mization toolbox. 2. Optimal power flow problem Let the objective function to be minimized, is given below 0142-0615/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijepes.2006.03.024 * Tel.: +91 1972 254522; fax: +91 1972 223834. E-mail address: yrsood@gmail.com www.elsevier.com/locate/ijepes Electrical Power and Energy Systems 29 (2007) 65–75