Applying network window schema and a simulated annealing technique to optimal VAR planning in large-scale power systems

This paper combines network window schema with a simulated annealing (SA) technique and applies them to reactive volt-ampere (VAR) planning in large-scale power systems. The network window schema alleviates the computational burden in large-scale systems, while the simulated annealing technique finds global optimal solutions for combinatorial complex problems. The solution algorithm based on this technique consists of two stages. The first stage introduces a moving window schema that partitions a large system into several small subsystems; an equivalent method then builds a reduced system for every subsystem. The results from the first stage are mapped into the original system and serve as the initial conditions for the second stage. The second stage compensates for the effect of system partitioning on the quality and feasibility of solutions. According to simulation results, the proposed approach significantly reduces the run time of the SA, and also improves the quality of the solution. For a 358-bus test system, loss reduction was improved by 17.6%, along with a saving in CPU time of approximately 50%.