Optimizing offshore outage planning via renewable forecasts

Summary

In this case study, you will discover details regarding:

1. Reasons why Elia is facing increasing challenges in planning offshore grid outages due to renewable energy variability and reliance on manual processes
2. The innovative forecast-optimization workflow developed by N-SIDE and UMONS, which utilizes tailored wind power forecasting and Mixed-Integer Programming (MIP) to enhance the efficiency of maintenance scheduling
3. The advantages of implementing such tools include minimizing “wasted resources” (false alarms) and avoiding “missed opportunities” for maintenance, along with supporting decision-making through dedicated user interfaces

Introduction

PROOF stands for Predictions of Renewables Optimized for Offshore using Forecasting. This project is a team effort by N-SIDE, UMONS, Elia, the Belgian transmission system operator, and with the support of the Energy Transition Fund. Its goal is to improve outage planning based on offshore wind power generation forecasting. This is crucial for achieving maintenance targets while avoiding costly cancellations triggered by uncertainties in renewable energy generation.

Partner Challenge

Planning offshore-related maintenance works and outages is highly challenging. This is due to the mix of renewable energy production, aging assets, and onshore grid constraints. Outages need to be scheduled without compromising grid security and without offshore renewable energy curtailment. 

Current planning suffers from issues because the process is completely manual and relies heavily on the planners’ experience. This can cause last-minute cancellations or delays. These issues lead to wasted resources and increased costs.

Solution/Approach

PROOF provides an integrated Forecast-Optimization workflow. The core is a refined Time Series Forecaster. It provides fine-grained granularity in both time and space, providing 15-minute interval forecasts per Belgian offshore wind farm, up to seven days ahead. In addition, the forecasting engine also provides a flow forecast for the HVDC Nemo cable between Belgium and the United Kingdom, as well as a wave height forecast.

These are fed into a Mixed-Integer Programming (MIP) optimization engine. The optimizer maximizes the executed requests and aligns with user-preferred days (which can correspond to tentatively planned days), while minimizing curtailment costs and change costs. Crucially, the model enforces various constraints, including power thresholds (which must be respected at all working time periods of the day), wave height safety checks, forbidden days, and service center resource availability. The optimizer enables the planning of complex, conditional outages that depend not only on wind production but also on outage combinations.

Benefits

PROOF improves the short-term outage planning process by transitioning to a proactive, forecast-based system. Through testing* against external benchmarks and an ideal “perfect information” scenario, the solution demonstrated significant improvements in operational decision quality and resource optimization.

• Superior Forecasting Accuracy: The final operational model’s aggregated week-ahead Mean Absolute Error (MAE%) was 14.73%, demonstrating outperformance compared to existing benchmarks (21.76% and 17.98%).
• Reduced Resource Waste: Achieved a low 4.00% rate of cancelled outages, significantly outperforming high-waste benchmarks (~7%), to drastically reduce the financial impact of mobilizing crews for non-feasible windows.
• Increased Scheduling Precision: Minimized missed opportunities to 8.22% (versus ~14-22% in benchmarks), ensuring that favorable weather conditions are fully utilized to maintain optimal power generation.

Moreover, PROOF allows for the identification of new maintenance opportunities thanks to the consideration of generation outages and the combination of transmission outages, while the manual process only looks at every single request individually.

*Tests were carried out for 3 of the 10 Belgian offshore wind farms, for the period December 2022 to November 2023.

Conclusion, Outcome, and Results

The PROOF R&D project concluded in December 2025. The integrated system successfully validated its capability to handle complex conditional outages and align maintenance schedules with minimized generation loss. In the coming year, the Elia outage planning team will test and compare the developed tool with their current practices.