Application of Ordinal Optimization to Reactive Volt-Ampere Sources Planning Problems

Reactive volt-ampere sources planning is an effort to determine the most effective investment plan for new reactive sources at given load buses while ensuring appropriate voltage profile and satisfying operational constraints. Optimization of reactive volt-ampere sources planning is not only a difficult problem in power systems, but also a large-dimension constrained optimization problem. In this paper, an ordinal optimization-based approach containing upper and lower level is developed to solve this problem efficiently. In the upper level, an ordinal search (OS) algorithm is utilized to select excellent designs from a candidate-design set according to the system’s structural information exploited from the simulations executed in the lower level. There are five stages in the ordinal search algorithm, which gradually narrow the design space to search for a good capacitor placement pattern. The IEEE 118-bus and IEEE 244-bus systems with four load cases are employed as the test examples. The proposed approach is compared with two competing methods; the genetic algorithm and Tabu search, and a commercial numerical libraries (NL) mixed integer programming tool; IMSL Numerical Libraries. Experimental results illustrate that the proposed approach yields an outstanding design with a higher quality and efficiency for solving reactive volt-ampere sources planning problem.