- author = "Yikai Ma",
- title = "Maintenance policy optimisation for wind farms",
- school = "University of Warwick",
- year = "2022",
- address = "UK",
- month = sep,
- keywords = "genetic algorithms, genetic programming, Wind power plants, Maintenance and repair, Wind power plants, Design and construction, Maintenance, Decision making, Heuristic algorithms",
-
URL = "
https://wrap.warwick.ac.uk/id/eprint/181189/",
-
URL = "http://webcat.warwick.ac.uk/record=b3948260",
-
URL = "
https://wrap.warwick.ac.uk/id/eprint/181189/1/WRAP_Theses_Ma_2022.pdf",
- size = "136 pages",
-
abstract = "The wind farm industry has undergone significant
growth in recent years. The cost of operating and
maintaining a wind farm can constitute a large
proportion of its overall lifetime cost. Therefore,
designing an effective maintenance policy is
particularly important and has attracted significant
recent attention. Developing such policies for wind
farms is a complicated and challenging problem, with
several aspects needing to be considered, such as the
dependencies among wind turbine components, the
interdependencies between actions, the priority of
activities, various uncertainties and different
resource constraints. An ideal maintenance policy
should simultaneously decide when to trigger and how to
schedule maintenance activities. This research uses
simulation-based optimisation to construct and obtain
approximately optimal maintenance policies for wind
farms, which can simultaneously determine when to
trigger maintenance and how to schedule maintenance
activities. A discrete-event simulation model is used
to simulate the performance of a maintenance policy
under a given scenario, considering all
interdependencies and uncertainties. Different
optimisation algorithms are used to evolve and obtain
effective maintenance policies.
This research proposes three mechanisms to automatically generate maintenance policies, considering single and multi-objective optimisation problems. These policies close several gaps in existing studies. The main contributions of this research are as follows.
1. Three aspects of a maintenance policy are simultaneously considered: reliability based triggering, opportunistic maintenance, and scheduling.
2. Genetic programming is used to automatically construct open-ended maintenance policies for wind farms.
3. Considerations of wind conditions and spare parts inventory are demonstrated.",
- notes = "For more information, please contact the WRAP Team at: wrap@warwick.ac.uk",
