Optimal allocation of phasor measurement units using practical constraints in power systems

Abstract

The purpose of the research is to find a strategic placement of phasor measurement units (PMUs) that can maintain the observability of the power system using a binary variant of particle swarm optimisation method (BPSO). In recent years, integer linear programming (ILP) is the common optimisation method used to solve this problem. However, recent studies have shown that deterministic methods such as ILP are still unable to find the most optimal placement of PMUs. Therefore, there is still room for further investigation using heuristic algorithm due to its search-based nature. It is important to determine the most strategic placement of PMUs to ensure that the PMUs are fully utilised due to their expensive price tag. The challenge in using heuristic algorithm lies in its weakness when involving large-sized problem since it is prone to stuck in local optima as shown in earlier studies where most of them applied their proposed approach to small bus systems. To prevent the BPSO method from being stuck in local optima, a mutation strategy is proposed in addition to the V-shaped sigmoid function replacing the S-shaped sigmoid function. The proposed method is designed to intensify the local search of the algorithm around the current best solution to help instigate the particles from being stuck in local optima and consequently finding a better solution. This is shown to be a crucial factor based on the numerical results obtained where it outperforms all methods for all bus systems tested including the IEEE 300-bus system in terms of measurement redundancy and the number of PMUs. This study also considers other problem constraints such as zero-injection bus, single PMU loss and also PMU’s channels limit. Most of the existing studies considered PMU to have an unlimited number of channels whereas in practical, PMU does have a finite number of channels that it can use.