Comparison of advanced non-parametric models for wind turbine power curves

Research output: Contribution to journalArticle

Abstract

To continuously assess the performance of a wind turbine (WT), accurate power curve modelling is essential. Various statistical methods have been used to fit power curves to performance measurements; these are broadly classified into parametric and non-parametric methods. In this study, three advanced non-parametric approaches, namely: Gaussian Process (GP); Random Forest (RF); and Support Vector Machine (SVM) are assessed for WT power curve modelling. The modelled power curves are constructed using historical WT supervisory control and data acquisition, data obtained from operational three bladed pitch regulated WTs. The modelled power curve fitting performance is then compared using suitable performance, error metrics to identify the most accurate approach. It is found that a power curve based on a GP has the highest fitting accuracy, whereas the SVM approach gives poorer but acceptable results, over a restricted wind speed range. Power curves based on a GP or SVM provide smooth and continuous curves, whereas power curves based on the RF technique are neither smooth nor continuous. This study highlights the strengths and weaknesses of the proposed non-parametric techniques to construct a robust fault detection algorithm for WTs based on power curves.

LanguageEnglish
Pages1503-1510
Number of pages8
JournalIET Renewable Power Generation
Volume13
Issue number9
Early online date20 Mar 2019
DOIs
Publication statusPublished - 8 Jul 2019

Fingerprint

Wind turbines
Support vector machines
Curve fitting
Fault detection
Data acquisition
Statistical methods

Keywords

  • curve fitting
  • decision tree
  • fault detection
  • parameter estimation
  • support vector machines
  • wind turbines

Cite this

@article{71f6ecf35ddb41cb97b78f3cd4e97226,
title = "Comparison of advanced non-parametric models for wind turbine power curves",
abstract = "To continuously assess the performance of a wind turbine (WT), accurate power curve modelling is essential. Various statistical methods have been used to fit power curves to performance measurements; these are broadly classified into parametric and non-parametric methods. In this study, three advanced non-parametric approaches, namely: Gaussian Process (GP); Random Forest (RF); and Support Vector Machine (SVM) are assessed for WT power curve modelling. The modelled power curves are constructed using historical WT supervisory control and data acquisition, data obtained from operational three bladed pitch regulated WTs. The modelled power curve fitting performance is then compared using suitable performance, error metrics to identify the most accurate approach. It is found that a power curve based on a GP has the highest fitting accuracy, whereas the SVM approach gives poorer but acceptable results, over a restricted wind speed range. Power curves based on a GP or SVM provide smooth and continuous curves, whereas power curves based on the RF technique are neither smooth nor continuous. This study highlights the strengths and weaknesses of the proposed non-parametric techniques to construct a robust fault detection algorithm for WTs based on power curves.",
keywords = "curve fitting, decision tree, fault detection, parameter estimation, support vector machines, wind turbines",
author = "Pandit, {Ravi Kumar} and David Infield and Athanasios Kolios",
year = "2019",
month = "7",
day = "8",
doi = "10.1049/iet-rpg.2018.5728",
language = "English",
volume = "13",
pages = "1503--1510",
journal = "IET Renewable Power Generation",
issn = "1752-1416",
publisher = "Institution of Engineering and Technology",
number = "9",

}

TY - JOUR

T1 - Comparison of advanced non-parametric models for wind turbine power curves

AU - Pandit, Ravi Kumar

AU - Infield, David

AU - Kolios, Athanasios

PY - 2019/7/8

Y1 - 2019/7/8

N2 - To continuously assess the performance of a wind turbine (WT), accurate power curve modelling is essential. Various statistical methods have been used to fit power curves to performance measurements; these are broadly classified into parametric and non-parametric methods. In this study, three advanced non-parametric approaches, namely: Gaussian Process (GP); Random Forest (RF); and Support Vector Machine (SVM) are assessed for WT power curve modelling. The modelled power curves are constructed using historical WT supervisory control and data acquisition, data obtained from operational three bladed pitch regulated WTs. The modelled power curve fitting performance is then compared using suitable performance, error metrics to identify the most accurate approach. It is found that a power curve based on a GP has the highest fitting accuracy, whereas the SVM approach gives poorer but acceptable results, over a restricted wind speed range. Power curves based on a GP or SVM provide smooth and continuous curves, whereas power curves based on the RF technique are neither smooth nor continuous. This study highlights the strengths and weaknesses of the proposed non-parametric techniques to construct a robust fault detection algorithm for WTs based on power curves.

AB - To continuously assess the performance of a wind turbine (WT), accurate power curve modelling is essential. Various statistical methods have been used to fit power curves to performance measurements; these are broadly classified into parametric and non-parametric methods. In this study, three advanced non-parametric approaches, namely: Gaussian Process (GP); Random Forest (RF); and Support Vector Machine (SVM) are assessed for WT power curve modelling. The modelled power curves are constructed using historical WT supervisory control and data acquisition, data obtained from operational three bladed pitch regulated WTs. The modelled power curve fitting performance is then compared using suitable performance, error metrics to identify the most accurate approach. It is found that a power curve based on a GP has the highest fitting accuracy, whereas the SVM approach gives poorer but acceptable results, over a restricted wind speed range. Power curves based on a GP or SVM provide smooth and continuous curves, whereas power curves based on the RF technique are neither smooth nor continuous. This study highlights the strengths and weaknesses of the proposed non-parametric techniques to construct a robust fault detection algorithm for WTs based on power curves.

KW - curve fitting

KW - decision tree

KW - fault detection

KW - parameter estimation

KW - support vector machines

KW - wind turbines

UR - http://www.scopus.com/inward/record.url?scp=85067691123&partnerID=8YFLogxK

U2 - 10.1049/iet-rpg.2018.5728

DO - 10.1049/iet-rpg.2018.5728

M3 - Article

VL - 13

SP - 1503

EP - 1510

JO - IET Renewable Power Generation

T2 - IET Renewable Power Generation

JF - IET Renewable Power Generation

SN - 1752-1416

IS - 9

ER -