Reactive power optimization is a kind of complicated nonlinear optimization problem. In early works, mathematical analysis programming methods with mature and integrated theory were used, which can obtain the final result very quickly, but unfortunately the functions must be differentiable, that's not easy to attain. Nowadays artificial intelligent algorithms that are on the basis of random search are booming. This series of algorithms with self-adapting step size can search the solution directly and more likely to acquire global optimization. In this paper, a new voltage-reactive power optimization method, which considers the equipment operational cost and confines the number of operated equipment in each optimization, is introduced. The reactive power optimization result of IEEE 30 node system and the network of Bayannaoer region by Artificial Fish Swarm Algorithm (AFSA) shows that AFSA has a strong robustness and possesses the excellent the value in theory and practice. It also shows that AFSA is a successful and feasible approach for reactive power optimization.
|