EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

4219. Investigating condition-based and predictive strategies for heavy maintenance at offshore wind farms

Invited abstract in session WD-34: New problems in logistics under uncertainty, stream Stochastic, Robust and Distributionally Robust Optimization.

Wednesday, 14:30-16:00
Room: 43 (building: 303A)

Authors (first author is the speaker)

1. Vibeke Hvidegaard Petersen
NTNU
2. Christian Rognes
NTNU
3. Herman Grytbakk Tronsaune
NTNU
4. Magnus Stålhane
NTNU

Abstract

Until recently maintenance at offshore wind farms have been conducted using a preventive- or condition-based maintenance strategy. However, as more information becomes available, and better prediction models are developed, the goal is to further reduce the cost of energy from offshore wind farms by moving to a predictive maintenance strategy. In such a strategy, probabilistic models of the future conditions of each component of a wind turbine are used as a basis for planning the maintenance.
Heavy maintenance where large components such as blades, generators, etc., are replaced, is one of the costliest types of maintenance as it, besides new components, requires chartering of a specialized jack-up vessel.
In this work we investigate how a predictive maintenance strategy affect the cost of heavy maintenance compared to other maintenance strategies. To do this, different strategies are evaluated through a framework simulating the 25-year long lifetime of a wind farm. Each year in the simulation framework, an optimization problem determining how many components to order and the chartering length of a jack-up vessel is solved. As these decisions must be taken well in advance of the maintenance period where component failures and weather conditions governing when the vessel can operate are unknown, this problem is formulated as a two-stage stochastic program. The vessel and component deployment is then simulated as the component failures and weather conditions are realized.

Keywords

Status: accepted

Back to the list of papers