Reducing sensor complexity for monitoring wind turbine performance using principal component analysis

Availability and reliability are among the priority concerns for deployment of distributed generation (DG) systems, particularly when operating in a harsh environment. Condition monitoring (CM) can meet the requirement but has been challenged by large amounts of data needing to be processed in real...

Full description

Saved in:
Bibliographic Details
Published in:Renewable energy 2016-11, Vol.97, p.444-456
Main Authors: Wang, Yifei, Ma, Xiandong, Joyce, Malcolm J.
Format: Article
Language:English
Subjects:
Algorithms
Condition monitoring
Distributed generation
Feature extraction
Frequency domains
Mathematical models
Monitoring
Principal component analysis
Principal component analysis (PCA)
Sensors
Wind turbine
Wind turbines
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Availability and reliability are among the priority concerns for deployment of distributed generation (DG) systems, particularly when operating in a harsh environment. Condition monitoring (CM) can meet the requirement but has been challenged by large amounts of data needing to be processed in real time due to the large number of sensors being deployed. This paper proposes an optimal sensor selection method based on principal component analysis (PCA) for condition monitoring of a DG system oriented to wind turbines. The research was motivated by the fact that salient patterns in multivariable datasets can be extracted by PCA in order to identify monitoring parameters that contribute the most to the system variation. The proposed method is able to correlate the particular principal component to the corresponding monitoring variable, and hence facilitate the right sensor selection for the first time for the condition monitoring of wind turbines. The algorithms are examined with simulation data from PSCAD/EMTDC and SCADA data from an operational wind farm in the time, frequency, and instantaneous frequency domains. The results have shown that the proposed technique can reduce the number of monitoring variables whilst still maintaining sufficient information to detect the faults and hence assess the system’s conditions. •Proposes an optimal sensor selection methodology based on PCA for condition monitoring of wind turbines.•Examines the method with both simulation data in PSCAD/EMTDC and SCADA data of an operational wind farm.•Examines the method with data in the time, frequency, and instantaneous frequency domains.•Proves sufficient information is still maintained from retained variables for fault diagnosis.•We demonstrate the feasibility of the proposed sensor selection methodology.
ISSN:0960-1481
1879-0682
DOI:10.1016/j.renene.2016.06.006