Machine learning in wind turbine O&M

Conor Mckinnon, James Carroll, Alasdair McDonald, Sofia Koukoura

Research output: Contribution to conferencePoster

33 Downloads (Pure)

Abstract

This research will investigate the use of Machine Learning techniques in various applications within the field of Wind Energy The general approach to Machine Learning follows the steps shown on the right Model selection is done through literature review, which depends on the data used This data is then processed and cleaned, through clustering and removal of outliers Features are extracted from the data, either from univariate statistics to find the feature with the least variance from the target, or PCA to reduce the number of features to two abstract features with no physical meaning This data is then used to train and test the model(s) The results produced are then analysed, either using existing alarm data, or through k folds cross validation These results can also inform on model selection
Original languageEnglish
Publication statusPublished - 7 Mar 2019
EventFuture Wind and Marine - University of Strathclyde, Glasgow, United Kingdom
Duration: 7 Mar 2019 → …

Conference

ConferenceFuture Wind and Marine
CountryUnited Kingdom
CityGlasgow
Period7/03/19 → …

Keywords

  • machine learning
  • anomaly detection techniques
  • operations & maintenance
  • neural networks
  • One Class Support Vector Machines (OCSVM)

Fingerprint Dive into the research topics of 'Machine learning in wind turbine O&M'. Together they form a unique fingerprint.

Cite this