Using a new modified harmony search algorithm to solve multi-objective reactive power dispatch in deterministic and stochastic models

The optimal reactive power dispatch (ORPD) problem is a very important aspect in power system planning, and it is a highly non-linear, non-convex optimization problem because it consists of both the continuous and discrete control variables. Since a power system has an inherent uncertainty, this paper presents both the deterministic and stochastic models for the ORPD problem in multi-objective and single-objective formulations, respectively. The deterministic model considers three main issues in the ORPD problem including the real power loss, voltage deviation, and voltage stability index. However, in the stochastic model, the uncertainties in the demand and equivalent availability of shunt reactive power compensators have been investigated. To solve them, we proposed a new modified harmony search algorithm (HSA), implemented in single and multi-objective forms. Since, like many other general purpose optimization methods, the original HSA often traps into the local optima, an efficient local search method called chaotic local search (CLS) and a global search operator are proposed in the internal architecture of the original HSA algorithm to improve its ability in finding the best solution because the ORPD problem is very complex, with different types of continuous and discrete constrains, i.e. excitation settings of generators, sizes of fixed capacitors, tap positions of tap changing transformers, and amount of reactive compensation devices. Moreover, the fuzzy decision-making method is employed to select the best solution from the set of Pareto solutions. The proposed model is individually examined and applied on different test systems. The simulation results show that the proposed algorithm is suitable and effective for the reactive power dispatch problem compared to the other available algorithms.