A wake-induced two-phase planning framework for offshore wind farm maintenance with stochastic mixed-integer program

Abstract

Offshore wind farms are essential in fulfilling global renewable energy targets; yet, their maintenance poses substantial challenges due to remote marine locations and harsh environmental conditions. Effective maintenance strategies and implementation are crucial to enhance operational efficiency and reduce costs. Current models frequently overlook an integrated perspective on periodic factors and uncertainties inherent in metocean conditions and failure rates, leading to suboptimal planning and increased costs. This study bridges this gap by introducing a comprehensive maintenance planning framework that incorporates these uncertainties. We formulate an annual planning model as a stochastic mixed-integer linear programming problem. The annual planning process aims to minimize operations and maintenance costs, including losses from downtime, by employing wake model constraints and accounting for stochastic scenarios. By tackling the scheme challenge, we garner strategic allocations of maintenance resources, which specifies the requisite number of operational vehicles and teams to be deployed over the year. Tentatively, we establish a long-term strategy and devise a short-term program that encompasses failure parameters and weather-related conditions. To further refine planning, we address weekly short-term scheduling problems to elaborate detailed maintenance schedules. Each week, maintenance tasks are adjusted based on actual stochastic conditions, yielding precise, real-world schedules. Our weekly scheduling considers not only preventive and corrective maintenance but also opportunistic maintenance. In particular, we harness a flexible scheduling approach to accommodate the efficiency of maintenance vessels. Computational tests demonstrate that our framework remarkably reduces downtime losses by 19.0% and recovery delays by 38.2%, leveraging scheduling flexibility.