No abstract

Wind turbines are often plagued by premature component failures, with drivetrain bearings being particularly subjected to these failures. To identify failing components, vibration condition monitoring has emerged and grown substantially. The fast Fourier transform (FFT) is the major signal processing method of vibrations. Recently, the wavelet transforms have been used more frequently in bearing vibration research, with one alternative being the discrete wavelet transform (DWT). Here, the low-frequency component of the signal is repeatedly decomposed into approximative and detailed coefficients using a predefined mother wavelet. An extension to this is the wavelet packet transform (WPT), which decomposes the entire frequency domain and stores the wavelet coefficients in packets. How wavelet transforms and FFT compare regarding fault detection in wind turbine drivetrain bearings has been largely overlooked in literature when applied on field data, with non-ideal placement of sensors and uncertain parameters influencing the measurements. This study consists of a comprehensive comparison of the FFT, a three-level DWT, and the WPT when applied on enveloped vibration measurements from two 2.5-MW wind turbine gearbox bearing failures. The frequency content is compared by calculating a robust condition indicator by summation of the harmonics and shaft speed sidebands of the bearing fault frequencies. Results show a higher performance of the WPT when used as a field vibration measurement analysis tool compared with the FFT as it detects one bearing failure earlier and more clearly, leading to a more stable alarm setting and avoidable, costly false alarms. KEYWORDS bearing failure, condition monitoring, discrete wavelet transform, wavelet packet transform, wind turbine gearbox bearings 1 INTRODUCTION Wind power is today the fastest growing renewable energy source in the world, with an installed capacity of 591 GW in 2018 and a predicted growth up to 908 GW in 2023. 1 However, wind turbines designed for a 20-year lifetime still experience premature failures with the root cause not yet fully understood. When compiling failures occurring in all the subsystems within the wind turbine, gearbox failures have been shown to cause the longest downtime and are thereby also associated with the highest cost per failure. 2,3 Out of the two main component types, the bearings experience most failures, around 76% of the time and with the gearbox output and generator shaft bearings being most represented, while the gears fail 17% of the time and other sources 7%. 4The method considered most effective to minimize the costs of these failures in rotating equipment is condition monitoring, where vibrations is the most common method as it can give early warnings on the health of bearings and gears before their degradation threaten the surrounding components. 5 Statistical methods in the time domain as well as frequency domain methods such as the fast Fourier transform (FFT) has throughoutThe peer review history for this article is available at...

Although the discrete wavelet transform has been used for diagnosing bearing faults for two decades, most work in this field has been done with test rig data. Since field data starts to be made more available, there is a need to shift into application studies. The choice of mother wavelet, ie, the predefined shape used to analyse the signal, has previously been investigated with simulated and test rig data without consensus of optimal choice in literature. Common between these investigations is the use of the wavelet coefficients' Shannon entropy to find which mother wavelet can yield the most useful features for condition monitoring. This study attempts to find the optimal mother wavelet selection using the discrete wavelet transform. Datasets from wind turbine gearbox accelerometers, consisting of enveloped vibration measurements monitoring both healthy and faulty bearings, have been analysed. The bearing fault frequencies' excitation level has been analysed with 130 different mother wavelets, yielding a definitive measure on their performance. Also, the applicability of Shannon entropy as a ranking method of mother wavelets has been investigated. The results show the discrete wavelet transforms ability to identify faults regardless of mother wavelet used, with the excitation level varying no more than 4%. By analysing the Shannon entropy, broad predictions to the excitation level could be drawn within the mother wavelet families but no direct correlation to the main results. Also, the high computational effort of high order Symlet wavelets, without increased performance, makes them unsuitable.

Condition monitoring is central to the efficient operation of wind farms due to the challenging operating conditions, rapid technology development, and a large number of aging wind turbines. In particular, predictive maintenance planning requires the early detection of faults with few false positives. Achieving this type of detection is a challenging problem due to the complex and weak signatures of some faults, particularly the faults that occur on the gearbox bearings of a turbine drivetrain. The results of former studies addressing condition-monitoring tasks using dictionary learning indicate that unsupervised feature learning is useful for diagnosis and anomaly detection purposes. However, these studies are based on small sets of labeled data from test rigs operating under controlled conditions that focus on classification tasks, which are useful for quantitative method comparisons but gives little insight into how useful these approaches are in practice or how can be used by existing condition-monitoring systems. Here, we investigate an unsupervised dictionary learning method for condition monitoring using vibration data recorded over 46 months under typical industrial operations. Thus, we contribute real-world industrial vibration data that are made publicly available and novel test results. In this study, dictionaries are learned from gearbox vibrations in six different turbines, and the dictionaries are subsequently propagated over a few years of monitoring data when faults are known to occur. We perform the experiment using two different sparse coding algorithms to investigate if the algorithm selected affects the features of abnormal conditions. We propose a dictionary distance metric derived from the dictionary learning process as a condition indicator and find the time periods of abnormal dictionary adaptation starting six months before a drivetrain bearing replacement and one year before the resulting gearbox replacement. In addition, we investigate the distance between dictionaries learned from geographically close turbines of the same type under healthy conditions. We find that the features learned are similar and that a dictionary learned from one turbine can be useful for monitoring a similar turbine.

No abstract