Prognosis of a Wind Turbine Gearbox Bearing Using Supervised Machine Learning

Abstract

Deployment of large-scale wind turbines requires sophisticated operation and maintenance strategies to ensure the devices are safe, profitable and cost-effective. Prognostics aims to predict the remaining useful life (RUL) of physical systems based on condition measurements. Analyzing condition monitoring data, implementing diagnostic techniques and using machinery prognostic algorithms will bring about accurate estimation of the remaining life and possible failures that may occur. This paper proposes to combine two supervised machine learning techniques, namely, regression model and multilayer artificial neural network model, to predict the RUL of an operational wind turbine gearbox using vibration measurements. Root Mean Square (RMS), Kurtosis (KU) and Energy Index (EI) were analysed to define the bearing failure stages. The proposed methodology was evaluated through a case study involving vibration measurements of a high-speed shaft bearing used in a wind turbine gearbox.

An Improved LightGBM Algorithm for Online Fault Detection of Wind Turbine Gearboxes

“…More and more attention has been paid to the fault diagnosis methods based on machine learning [9,10]. In machine learning, the boost algorithm combines weakly predictive models into a strongly predictive model, which is adjusted by increasing the weight of the error samples to improve the accuracy of the algorithm [11][12][13][14]. However, the boost algorithm needs to use the lower limit of the accuracy of the weak classifier in advance and has limited application in industrial fault diagnosis.…”

Estimating Residual Life Distributions of Complex Operational Systems Using a Remaining Maintenance Free Operating Period (RMFOP)-Based Methodology

“…For instance, wind turbine systems suffer from stochastic loadings due to various wind speeds day by day, which makes it challenging to determine prognosis. Considering today’s large-scale wind farms and the long distances from operation centres, the costs of manual maintenance will also be massive [ 2 ]. A similar challenge is also experienced by railway asset managers.…”

Predicting Frequency, Time-To-Repair and Costs of Wind Turbine Failures

“…The current study uses the risk factors determined in [13,14] to improve the prediction of the frequency of wind turbine failures and associated repair times Bayesian updating. In addition, we employ machine learning techniques on the risk factors, to assess wind turbine failure costs [13,14].Bayesian updating has not been used before in probabilistic assessments of time-to-failure (TTF) and time-to-repair (TTR) and machine learning techniques were not used previously in determining costs of failures for wind turbines although they were used for component condition monitoring [15][16][17] and for other applications [18][19][20][21]. The major contributions of this paper are two-fold: (1) Bayesian updating is used at the first time in probabilistic assessments of time-to-failure and time-to-repair of wind turbines.…”

“…Bayesian updating has not been used before in probabilistic assessments of time-to-failure (TTF) and time-to-repair (TTR) and machine learning techniques were not used previously in determining costs of failures for wind turbines although they were used for component condition monitoring [15][16][17] and for other applications [18][19][20][21]. The major contributions of this paper are two-fold: (1) Bayesian updating is used at the first time in probabilistic assessments of time-to-failure and time-to-repair of wind turbines.…”