Minimization of Wind Farm Operational Cost Based on Data-Driven Models

Scheduling a wind farm in the presence of uncertain wind speed conditions is presented. Two scheduling models, the base model and the stochastic optimization model, are developed by integrating mathematical programming and data mining. A migrated particle swarm optimization algorithm is developed for solving the two scheduling models. The solution computed by this algorithm determines the operational status and control settings of a wind turbine. The cost of operating a wind farm according to the solutions of both scheduling models closely matches the cost computed based on a schedule under a perfect information scenario. The computational results provide insights into the management and operation of wind farms.