Capacitor Allocation in Distribution Networks Based on a Two-Stage Analytical-Heuristic Approach

In distribution networks, reduction of search space can be an efficient way to conquer the complexity of capacitor placement. In this paper, based on search space reduction, a two-stage procedure is proposed to efficiently solve capacitor placement problem. For this aim, as the first stage, analytical indices are introduced to determine the importance of network buses for compensation. The indices prioritize buses based on the network losses sensitivity to change of voltage and/or reactive power. As a result of the first stage, a number of most important buses are considered as the candidates for compensation. Then, the second stage is conducted using an optimization method to solve the capacitor placement problem with respect to the candidate buses. To do the second stage, an enhanced metaheuristic method, named enhanced crow search algorithm (ECSA), is developed. Compared to popular methods of search space reduction, loss sensitivity factor (LSF) and loss sensitivity indices (LSI1 and LSI2), simulation results confirm the superiority of the proposed two-stage procedure. On 33-bus network with respect to peak condition, when 3 buses are regarded as the candidate buses (around 10% of the total network buses), the optimal solution is found. In this case, compared to LSF, LSI1 and LSI2, objective function value decreases 2.9, 11.2 and 6.6%, respectively. On 69-bus network, when 6 buses are selected as the candidate buses (around 10% of the total network buses), the optimal solution is obtained. In this case, compared to LSF, LSI1 and LSI2, objective function value decreases around 0.7%.