Single phase and three phase power flow analysis using hybrid differential evolution and particle swarm optimization

Abstract

The first and foremost step in solving any power system problem is to conduct power flow analysis for the power system. It is necessary for planning, operation, economic scheduling, stability analysis, contingency analysis and exchange of power between utilities. Conventionally, the load flow or power flow problem is solved using numerical techniques such as Newton-Raphson (NR) and Fast Decoupled Method. Solutions obtained using these techniques depend on getting the inverse of the Jacobian matrix of the system. The evolutionary algorithms such as Genetic Algorithm, Evolutionary Programming, Particle Swarm Optimization and Differential Evolution have been proposed to solve the single-phase power flow problem and to determine the maximum loadability limit of power system. Differential Evolution is one of the powerful algorithms of evolutionary computation since it has excellent convergence characteristics and only a few control parameters. The computational algorithm of DE is simple to understand and implement. PSO shows excellent convergence during the initial iterations. First, the single-phase power flow problem is formulated in rectangular coordinate and the MDEPSO and MHPSO algorithms are applied to solve this problem. To compare the performances of these algorithms, test cases of 6 bus, IEEE 30 bus and IEEE 118 bus systems are considered. The results show that MDEPSO algorithm provides more accurate results. The DE, MDEPSO, MHPSO algorithms are applied to determine the maximum loadability limit of power systems. The results are compared with the Continuation Power Flow (CPF). The results show that the evolutionary algorithms are superior to CPF. MDEPSO is better than DE and MHPSO. The results obtained using the evolutionary algorithms are same as that are obtained using NR. The optimization technique eliminates the formation and inversion of Jacobian matrix and hence reduces the complexity of the problem. newline