Modeling the effects of the environment on wind turbine failure modes using neural networks

Graeme Wilson, David McMillan

Research output: Contribution to conferencePaper

Abstract

Neural networks were used to investigate if any relationship existed between maximum daily gust speed, average daily wind speed and temperature and wind turbine failure modes. Using five years of weather station data a typical site characteristic was determined using the neural network. This was then compared to a characteristic produced using only weather data for days when failures occurred. These failure and normal characteristics were then compared to determine if any relationships existed. It was found that in some sub-assemblies the failure trends differed to the normal conditions trend, suggesting that there may be relationships.

Conference

ConferenceIET International Conference on Sustainable Power Generation and Supply
Abbreviated titleSUPERGEN 2012
CountryChina
CityHangzhou
Period8/09/129/09/12
Internet address

Fingerprint

Wind turbines
Failure modes
Neural networks
Temperature

Keywords

  • wind turbines
  • reliability
  • asset management
  • SAP data
  • environment
  • effects
  • wind turbine failure modes
  • neural networks

Cite this

Wilson, G., & McMillan, D. (2012). Modeling the effects of the environment on wind turbine failure modes using neural networks. Paper presented at IET International Conference on Sustainable Power Generation and Supply , Hangzhou, China.
Wilson, Graeme ; McMillan, David. / Modeling the effects of the environment on wind turbine failure modes using neural networks. Paper presented at IET International Conference on Sustainable Power Generation and Supply , Hangzhou, China.
@conference{56c544be98d345799b3258c7fa782f78,
title = "Modeling the effects of the environment on wind turbine failure modes using neural networks",
abstract = "Neural networks were used to investigate if any relationship existed between maximum daily gust speed, average daily wind speed and temperature and wind turbine failure modes. Using five years of weather station data a typical site characteristic was determined using the neural network. This was then compared to a characteristic produced using only weather data for days when failures occurred. These failure and normal characteristics were then compared to determine if any relationships existed. It was found that in some sub-assemblies the failure trends differed to the normal conditions trend, suggesting that there may be relationships.",
keywords = "wind turbines, reliability, asset management, SAP data, environment, effects, wind turbine failure modes , neural networks",
author = "Graeme Wilson and David McMillan",
year = "2012",
month = "9",
language = "English",
note = "IET International Conference on Sustainable Power Generation and Supply , SUPERGEN 2012 ; Conference date: 08-09-2012 Through 09-09-2012",
url = "http://www.ukchinanet.com/UK-China{\%}20Network_conference{\%}20website{\%}20HZ.html",

}

Wilson, G & McMillan, D 2012, 'Modeling the effects of the environment on wind turbine failure modes using neural networks' Paper presented at IET International Conference on Sustainable Power Generation and Supply , Hangzhou, China, 8/09/12 - 9/09/12, .

Modeling the effects of the environment on wind turbine failure modes using neural networks. / Wilson, Graeme; McMillan, David.

2012. Paper presented at IET International Conference on Sustainable Power Generation and Supply , Hangzhou, China.

Research output: Contribution to conferencePaper

TY - CONF

T1 - Modeling the effects of the environment on wind turbine failure modes using neural networks

AU - Wilson, Graeme

AU - McMillan, David

PY - 2012/9

Y1 - 2012/9

N2 - Neural networks were used to investigate if any relationship existed between maximum daily gust speed, average daily wind speed and temperature and wind turbine failure modes. Using five years of weather station data a typical site characteristic was determined using the neural network. This was then compared to a characteristic produced using only weather data for days when failures occurred. These failure and normal characteristics were then compared to determine if any relationships existed. It was found that in some sub-assemblies the failure trends differed to the normal conditions trend, suggesting that there may be relationships.

AB - Neural networks were used to investigate if any relationship existed between maximum daily gust speed, average daily wind speed and temperature and wind turbine failure modes. Using five years of weather station data a typical site characteristic was determined using the neural network. This was then compared to a characteristic produced using only weather data for days when failures occurred. These failure and normal characteristics were then compared to determine if any relationships existed. It was found that in some sub-assemblies the failure trends differed to the normal conditions trend, suggesting that there may be relationships.

KW - wind turbines

KW - reliability

KW - asset management

KW - SAP data

KW - environment

KW - effects

KW - wind turbine failure modes

KW - neural networks

UR - http://www.ukchinanet.com/UK-China%20Network_conference%20website%20HZ.html

UR - http://www.ukchinanet.com/sample/Final%20conference_program.pdf

M3 - Paper

ER -

Wilson G, McMillan D. Modeling the effects of the environment on wind turbine failure modes using neural networks. 2012. Paper presented at IET International Conference on Sustainable Power Generation and Supply , Hangzhou, China.