Research On Distributed Computing Of The Optimization Problem Of Under-voltage Load Shedding Parameters

In recent years, voltage instability has become a threat to normal operation of the power system. Under-voltage load shedding (UVLS) is an economical and effective countermeasure against voltage instability, and selecting proper under-voltage load shedding parameters is a critical problem.The optimization problem of UVLS parameters is a long-term voltage stability-constrained optimal power flow problem which has very large amount of computation. It is hard to be solved by traditional centralized computation. The research on high-powered distributed computing method for the optimization of UVLS parameters becomes a question for discussion.In this paper, a distributed computing method based on enhanced particle swarm optimization algorithm for the optimization of UVLS parameters is presented. The problem is described as an optimization model with an L∞norm objective function. Two rules for the comparison of particles are introduced to keep infeasible solutions near the boundary and improve search efficiency. The strategy of performing global search and local search alternately is adopted to speed up convergence. Furthermore, an approximate strategy of contingency simulation is proposed which can reduce the computational burden effectively. At the same time, this paper adopts PC cluster with a high performance/price ratio and the communication program written by Delphi to form the platform of distributed computing, based on which the optimization problem of UVLS parameters is solved successfully. The results of a real world China power grid show that the proposed method not only improves the computation speed evidently, but also achieves satisfactory optimization results.