A Decision-dependent Stochastic Approach for Wind Farm Maintenance Scheduling Considering Wake Effect

Abstract

With higher proportion of wind generation, the operation and maintenance plan of wind farms plays an increasingly critical role in power systems. Failure probability based risk management of wind turbines (WT) contributes to cost efficient maintenance schedules and reliable operation of a wind farm. In this paper, focusing on a large-scale wind farm, a decision-dependent stochastic wind farm maintenance strategy is studied considering the influence of wake effect. Based on the failure probability expression of WTs and wake effect model, the expected wind speed at every WT is formulated as a decision-dependent uncertain parameter, which depends on the failure probability and maintenance status of its upstream WTs. Then, considering uncertainties in electricity price and wind speed and failure probability of WTs, the maintenance decision-making model of the wind farm is established as a decision-dependent stochastic programming model to maximize total expected wind power revenue. The established non-convex stochastic optimization problem is solved using particle swarm algorithm, and methods to improve computation efficiency are presented. Case study results verified the validity of the proposed method.