A novel adequate bi-level reactive power planning strategy

• A bi-level strategy is investigated to solve the reactive power planning problem.
• The weakest buses are identified by the refined heuristic process in the first level.
• Two versions of Differential Evolution Algorithm (DEA) are proposed in the second level.
• The proposed DEA versions are competitive compared to other optimization techniques.
• The control parameters of the proposed DEA are optimally adopted.

Planning of reactive power sources is a serious issue for secure and economic operation of power systems. In this paper, a bi-level strategy is proposed to optimize the Reactive Power Planning (RPP) problem. In the first level, the weakest buses are selected to be the optimal placements to install the additional VAR sources and its corresponding suitable sizes are determined using a proposed Refined Heuristic Process (RHP). In the second level, two modified versions of Differential Evolution Algorithm (DEA) are proposed for optimizing the RPP control variables which able to minimize both the allocation costs of additional VAR sources throughout the system, and the system operational costs of real power losses. To validate the effectiveness of proposed strategy, several applications are carried out on three power systems networks namely IEEE 14-bus, IEEE 30-bus test systems and the West Delta region system as a part of the Egyptian Unified network. The proposed strategy is evaluated compared with other optimization methods as Genetic Algorithms (GA), Particle Swarm Optimization (PSO), and the commonly used Differential Evolution (DE) version (DE/rand/1). The robustness of the proposed versions of DEA is proven compared to other optimization techniques. Added to that, the control parameters of the proposed DEA are optimally identified. Numerical results show that the proposed version of DEA achieves highest reduction in the operation and investment costs compared to other optimizing algorithms in the literature which denotes that the proposed version of DEA can be efficiently applied to the RPP problem.