Normal Behaviour Models for Wind Turbine Vibrations: Comparison of Neural Networks and a Stochastic Approach

Abstract

Abstract:To monitor wind turbine vibrations, normal behaviour models are built to predict tower top accelerations and drive-train vibrations. Signal deviations from model prediction are labelled as anomalies and are further investigated. In this paper we assess a stochastic approach to reconstruct the 1 Hz tower top acceleration signal, which was measured in a wind turbine located at the wind farm Alpha Ventus in the German North Sea. We compare the resulting data reconstruction with that of a model based on a neural network, which has been previously reported as a data-mining algorithm suitable for reconstructing this signal. Our results present evidence that the stochastic approach outperforms the neural network in the high frequency domain (1 Hz). Although neural network retrieves accurate step-forward predictions, with low mean square errors, the stochastic approach predictions better preserve the statistics and the frequency components of the original signal, retaining high accuracy levels. The implementation of our stochastic approach is available as open source code and can easily be adapted for other situations involving stochastic data reconstruction. Based on our findings we argue that such an approach could be implemented in signal reconstruction for monitoring purposes or for abnormal behaviour detection.

Fault Diagnosis Based on an Approach Combining a Spectrogram and a Convolutional Neural Network with Application to a Wind Turbine System

“…It was proposed that the vibration signal-to-noise ratio of the motor can be increased and then de-noised using the empirical wavelet transform, after which the fault feature can be extracted. To monitor wind turbine vibrations, in 2017, reference [7] presented normal behaviour models to predict tower-top accelerations and drivetrain vibrations. In reference [8], the implementation and performance analysis of a vibration-response-based Gaussian Mixture Model Random Coefficient (GMM-RC) model-based structural health monitoring (SHM) framework for structures with time-dependent dynamics under significant uncertainty (operational and environmental) have been presented.…”

A Markovian reliability approach for offshore wind energy system analysis in harsh environments

“…It is a very important tool for analyzing the probability distribution in possible states within the state space . The recent application of the stochastic approach, Markov process, and neural network for offshore energy system analysis have shown the capacity of the Markovian process in dynamic system modeling. However, the application of the Markovian process in dynamic reliability prediction for the offshore wind turbine generators has not been demonstrated.…”

Behavior Anomaly Indicators Based on Reference Patterns—Application to the Gearbox and Electrical Generator of a Wind Turbine

“…These indicators are based on patterns that were previously obtained observing the typical normal behavior of the components monitored [12][13][14].…”