Predictive optimal control of wind farm reactive sources

This paper presents two different optimization models for predictive optimal control of wind farm reactive sources. The models are formulated as a reactive power dispatch problem. The objective of this optimization task is to reduce the operation cost of the on load tap changing transformers by reducing the short term tap changes using a predictive optimal control. Wind generation is predicted by neural network. The stochastic nature of the wind generation is modeled as a binary tree for the stochastic tree method. The proposed methods are tested on an offshore wind farm and solved using particle swarm optimization algorithm.