A Comparison of Multi-Objective Optimisation of Two Wind Turbine Controller Designs

This paper presents a comparison between two pitch controllers for a 10MW wind turbine, a baseline using PI and a state-space model, where each controller's parameters have been chosen using Bayesian optimization. Four objectives are simultaneously examined: energy output; pitch action; blade l...

Full description

Saved in:
Bibliographic Details
Published inJournal of physics. Conference series Vol. 1618; no. 2; pp. 22046 - 22054
Main Authors Yu, W., Engels, W. P., Stock-Williams, C.F.W.
Format Journal Article
LanguageEnglish
Published Bristol IOP Publishing 01.09.2020
Subjects
Bayesian optimization
Control systems design
Controllers
Energy consumption
multi-objective
Multiple objective analysis
Parameters
Pareto front
Pareto optimization
Physics
Pitch (inclination)
State space models
state-space model controller
wind turbine
Wind turbines
Online AccessGet full text

Cover

More Information
Summary:This paper presents a comparison between two pitch controllers for a 10MW wind turbine, a baseline using PI and a state-space model, where each controller's parameters have been chosen using Bayesian optimization. Four objectives are simultaneously examined: energy output; pitch action; blade load; and tower load. First, it is shown that, for both controllers, using parameters selected from the obtained Pareto fronts achieves higher energy production with less energy consumption of pitch activities and with less load in blade and tower, compared with settings chosen by an expert. Second, in this robust comparison of the capabilities of both controllers, the increased degrees of freedom and reduced simplification in the more advanced controller are seen to have greater potential to improve each of the objectives.
ISSN:1742-6588
1742-6596
DOI:10.1088/1742-6596/1618/2/022046