Prediction of wind turbine generator failure using two‐stage cluster‐classification methodology

Abstract

Reducing wind turbine downtime through innovations surrounding asset management has the potential to greatly influence the levelised cost of energy (LCoE) for large wind farm developments. Focusing on generator bearing failure and vibration data, this paper presents a two‐stage methodology to predict failure within 1 to 2 months of occurrence. Results are obtained by building up a database of failures and training machine learning algorithms to classify the bearing as healthy or unhealthy. This is achieved by first using clustering techniques to produce subpopulations of data based on operating conditions, which this paper demonstrates can greatly influence the ability to diagnose a fault. Secondly, this work classifies individual clusters as healthy or unhealthy from vibration‐based condition monitoring systems by applying order analysis techniques to extract features. Using the methodology explained in the report, an accuracy of up to 81.6% correct failure prediction was achieved.

Wind turbine drivetrains: state-of-the-art technologies and future development trends

“…One classification of machine learning techniques for wind turbine condition monitoring is to divide them into supervised (classification and regression) and unsupervised (clustering) learning. An example of classification for fault detection, isolation and failure mode diagnosis on the gearbox is illustrated by Koukoura et al (2019), whereas Turnbull et al (2019) applied a combination of clustering and classification techniques to group similar operating conditions and detect generator faults. An example of regression to detect anomalies in vibration indicators can be found in Verstraeten et al (2019).…”

Cost Benefit of Implementing Advanced Monitoring and Predictive Maintenance Strategies for Offshore Wind Farms

“…Findings showed that most models use SCADA or simulated data, with almost two-thirds of methods using classification and the rest relying on regression. Neural networks, support vector machines and decision trees are most commonly used, with more recent papers now combining techniques and data sources to optimise performance such as [12,13].…”

A Wind Turbine Vibration Monitoring System for Predictive Maintenance Based on Machine Learning Methods Developed under Safely Controlled Laboratory Conditions

“…In the literature, several authors report on predictive maintenance using machine learning models [6][7][8][9][10][11], deep learning based on turbine performance curves, and condition monitoring to detect vibration anomalies from cleaning and processing vibration time windows signals for information extraction [12,13]. Moreover, image processing has been used in the analysis of rotary drives to determine their reference condition [14][15][16][17][18].…”