Wind turbine performance assessment & power curve outlier rejection using copula modelling

Research output: Contribution to conferencePoster

Abstract

The conventional means of assessing performance of a wind turbine is through consideration of its power curve. However, this representation fails to capture plausibility of measurement and cannot provide anomaly detection capabilities, which may assist in the detection of plant degradation. Although the probabilistic form of the power curve is complex, Copula models are presented here as a means of expressing the operational power curve as a joint distribution of wind speed and power output. This probabilistic model is demonstrated as an efficient way to remove outliers from operational SCADA data, simplifying and accelerating the process of identifying plant maloperation.

Conference

Conference2018 Global Offshore Wind
Abbreviated titlegow18
CountryUnited Kingdom
CityManchester
Period19/06/1820/06/18
Internet address

Fingerprint

Wind turbines
Degradation
Statistical Models

Keywords

  • wind turbine
  • performance
  • power curve
  • copula model

Cite this

Zorzi, G., Stephen, B., & McMillan, D. (2018). Wind turbine performance assessment & power curve outlier rejection using copula modelling. Poster session presented at 2018 Global Offshore Wind, Manchester, United Kingdom.
@conference{c5fc5ec9ff384a679dc2450dfafa3a1f,
title = "Wind turbine performance assessment & power curve outlier rejection using copula modelling",
abstract = "The conventional means of assessing performance of a wind turbine is through consideration of its power curve. However, this representation fails to capture plausibility of measurement and cannot provide anomaly detection capabilities, which may assist in the detection of plant degradation. Although the probabilistic form of the power curve is complex, Copula models are presented here as a means of expressing the operational power curve as a joint distribution of wind speed and power output. This probabilistic model is demonstrated as an efficient way to remove outliers from operational SCADA data, simplifying and accelerating the process of identifying plant maloperation.",
keywords = "wind turbine, performance, power curve, copula model",
author = "Giorgio Zorzi and Bruce Stephen and David McMillan",
year = "2018",
month = "6",
day = "19",
language = "English",
note = "2018 Global Offshore Wind, gow18 ; Conference date: 19-06-2018 Through 20-06-2018",
url = "https://events.renewableuk.com/gow18",

}

Zorzi, G, Stephen, B & McMillan, D 2018, 'Wind turbine performance assessment & power curve outlier rejection using copula modelling' 2018 Global Offshore Wind, Manchester, United Kingdom, 19/06/18 - 20/06/18, .

Wind turbine performance assessment & power curve outlier rejection using copula modelling. / Zorzi, Giorgio; Stephen, Bruce; McMillan, David.

2018. Poster session presented at 2018 Global Offshore Wind, Manchester, United Kingdom.

Research output: Contribution to conferencePoster

TY - CONF

T1 - Wind turbine performance assessment & power curve outlier rejection using copula modelling

AU - Zorzi, Giorgio

AU - Stephen, Bruce

AU - McMillan, David

PY - 2018/6/19

Y1 - 2018/6/19

N2 - The conventional means of assessing performance of a wind turbine is through consideration of its power curve. However, this representation fails to capture plausibility of measurement and cannot provide anomaly detection capabilities, which may assist in the detection of plant degradation. Although the probabilistic form of the power curve is complex, Copula models are presented here as a means of expressing the operational power curve as a joint distribution of wind speed and power output. This probabilistic model is demonstrated as an efficient way to remove outliers from operational SCADA data, simplifying and accelerating the process of identifying plant maloperation.

AB - The conventional means of assessing performance of a wind turbine is through consideration of its power curve. However, this representation fails to capture plausibility of measurement and cannot provide anomaly detection capabilities, which may assist in the detection of plant degradation. Although the probabilistic form of the power curve is complex, Copula models are presented here as a means of expressing the operational power curve as a joint distribution of wind speed and power output. This probabilistic model is demonstrated as an efficient way to remove outliers from operational SCADA data, simplifying and accelerating the process of identifying plant maloperation.

KW - wind turbine

KW - performance

KW - power curve

KW - copula model

M3 - Poster

ER -

Zorzi G, Stephen B, McMillan D. Wind turbine performance assessment & power curve outlier rejection using copula modelling. 2018. Poster session presented at 2018 Global Offshore Wind, Manchester, United Kingdom.