Gearbox Failure Research Articles

Diagnosis of Wind Turbine Yaw System Based on Self-Attention–Long Short-Term Memory (LSTM)

Addressing the challenges and significant risks associated with diagnosing faults in wind turbine yaw systems, along with the typically low diagnostic accuracy, this study introduces a Long Short-Term Memory (LSTM) neural network augmented by a self-attention mechanism (SAM) as a novel fault diagnosis technique for wind turbine yaw systems. The method integrates the automatic weighting capability of the self-attention mechanism on input features with the advantage of LSTM in processing time series data, thereby effectively capturing key information and long-term dependencies in the operating data of the yawing system. This combination enhances the accuracy of fault feature extraction to more accurately identify various types of fault modes within the yawing system. Six types of feature parameters are extracted from the raw data collected by the SCADA (Supervisory Control And Data Acquisition) system of the wind turbine and are utilized as inputs for the diagnostic model. These parameters are then fed into the self-attention–LSTM neural network model to diagnose the health status of the yaw system, including yaw bearing damage, yaw gearbox failure, yaw motor failure, and sensor failure. The experimental results demonstrate that the accuracy of LSTM fault diagnosis, when enhanced with the self-attention mechanism, can reach 98.67% with an appropriate amount of training samples, verifying its significant advantages in terms of accuracy and stability of fault diagnosis. The proposed fault diagnosis method exhibits a better model fitting effect, strong generalization ability, and high accuracy compared to other methods, providing robust support for the reliable operation and maintenance of wind turbines.

Integrated Multi-Scale Analysis and Advanced Prototype Model for Early Detection of Gearbox Failures Using Infrared Thermal Image Data Under Dynamic Conditions

Integrated Multi-Scale Analysis and Advanced Prototype Model for Early Detection of Gearbox Failures Using Infrared Thermal Image Data Under Dynamic Conditions

Dynamic Characteristics of Gear System Under Compound Fault of Tooth Root Crack and Bearing Outer Ring

Research on single faults like tooth root cracks and bearing outer ring failures is well-established. However, existing studies often replace the actual crack propagation path with a simplified linear crack model, neglecting the variation in stress intensity factors. They do not account for the coupling effect between gear teeth when calculating gear mesh stiffness. In practice, gearbox failures are predominantly compound failures involving both gears and bearings. This paper simulates the actual crack propagation path based on Workbench, introduces a dual-tooth meshing zone method to address the tooth-to-tooth coupling effect, and establishes a dynamic model of a system with compound faults of tooth root cracks and bearing outer ring defects. The dynamic response of the system under different levels of deterioration is then investigated. The findings indicate that the crack propagation path exhibits a concave shape, with the curvature of the path increasing as the crack deteriorates. The time-varying mesh stiffness calculated by considering tooth-to-tooth coupling effects aligns more closely with the finite element results. By analyzing the evolution patterns of the impact components and sidebands during the deterioration process, the dynamic evolution trends of the coupling phenomenon between tooth root cracks and bearing outer ring faults are clarified.

Study of the Gearbox Meshing Coupling Vibration Law Based on Mechanical Signal and Frequency Signal

In this study, the failure of a certain type of gearbox of a high-speed locomotive group with cracks is examined, and a gearbox failure assessment method that considers the coupled vibration is established and combined with mechanical signal and frequency signal to determine the basis for judging the failure of a faulty gearbox. First, according to the mechanical model of the finite element calculation, we determined the stress weak links and then the layout response stress measurement points and acceleration measurement points. We then calculated the gearbox ratio, meshing frequency, vibration frequency, mechanical response, modal response, and other frequency characteristics using the Hilbert–Huang transform (HHT) method and the fast Fourier transform (FFT) algorithm to analyze the vibration signals generated by different speeds and wheel out-of-roundness conditions. These were used to calculate the frequency of the different vibration sources of the mechanical response on the weak areas. The frequency correlations of the different vibration sources on the mechanical response in the weak areas were then analyzed, and the vibration transmission law of the gearbox case was obtained. The fault determination criterion was then determined, and the final cause of the fault was obtained.

Sensor concepts for gear damage detection in wind power drives

Abstract Gearbox failure in offshore wind turbines can be caused by fatigue damages like tooth fank fracture (TFF) or pitting damage. For damage detection in current applications, simple threshold values for vibration amplitudes are used to initiate a shutdown of the wind turbine. At this time significant damage to the gear is already present and an external maintenance is inevitable. The objective of this experimental study is to determine a better sensor concept for detection of gear damage in wind power drives during operation. The damage needs to be detected at a very early stage before the health condition of the gears becomes critical. Evaluation criteria for comparison of sensor concepts include the earliest possible detection of damage, robust and reliable detection under varying operating conditions, as well as practicality and cost-effectiveness. The test results show, vibration measurement in radial direction and in high frequency ranges is most suitable and most robust for pitting detection during operation. With the best sensor concept in place, it is possible to detect gear damage like pitting at a very early stage of 0.3 % pitting size.

Study of Variation in Physiochemical Properties of a Worm Gearbox Lubricant by Blending Castor Oil in the Base Lubricant

In both industrial & heavy-duty applications, oil analysis can be a very effective strategy for preventing gearbox failure, cutting downtime, and managing maintenance costs. In any application, gearboxes are essential to production, the unplanned downtime and the resulting productivity losses can prove to be extremely expensive. Used oil analysis is a preventative maintenance procedure that can point out and pinpoint mechanical breakdown inside the system before it gets serious or becomes too expensive to fix. A thorough oil examination can identify wear, corrosion, and other changes that could be harmful to the health and durability of the machinery. To reduce machine wear, prevent contamination, or optimize working conditions in challenging environments, lubricants are carefully formulated by mixing additives in the base oil. A lubricant with a high acid content can cause varnish and sludge to accumulate, which can clog oil filters and cause machine parts to corrode. When a lubricant degrades, acidic byproducts are produced because of the base stock and additives' process of decomposition in the presence of heat and air. To assess the presence of acidic constituents in the oil or its acid concentration, Total Acid Number (TAN) oil testing is employed. The present work studies the effects on Total Acid Number (TAN) and Total Base Number (TBN) properties of commercial lubricant (EP 90) when castor oil is mixed in percentage concentration in it. The findings of this research have unveiled that the acidic characteristics of the base lubricant (EP90 commercial gear oil) increases as the percentage of castor oil additive is increased in the base lubricant oil.

Wind turbine gearbox condition monitoring using AI-enabled virtual indicators*

Abstract At present, the condition monitoring of wind turbine gearbox mostly involves selecting monitoring indicators in advance and establishing normal behavior models based on deep learning. However, single physical monitoring indicator cannot fully characterize the operating status of the wind turbine gearbox. To address the above issues, a wind turbine gearbox condition monitoring method using convolutional neural network (CNN)-long and short term memory network (LSTM)-autoencoder (AE) enabled virtual indicators is proposed in this paper. This method first establishes a cleaning principle to preprocess the data of supervisory control and data acquisition (SCADA) system and screen out effective SCADA data further. A CNN-LSTM-AE monitoring model was trained using SCADA data during normal operating process, various characteristic parameters of the gearbox are integrated to construct the AI-enabled virtual indicators, and meanwhile warning thresholds was determined based on the probability density distribution of virtual indicators to analyze the operating status of the gearbox. Finally, the proposed method was validated using real SCADA data from two wind turbines in a cooperated wind farm with my group. Compared with a single physical indicator, the virtual indicator enables to detect gearbox early faults 5 d in advance, indicating that the proposed method can effectively alert wind turbine gearbox failures.

Experimental Characterization of Common-Mode Voltages and Currents in Gearbox Bearings of a Doubly-Fed Induction Generator Wind Turbine on a Test Bench

Wind turbine (WT) drivetrain failures can lead to expensive repairs and downtime, significantly impacting the Levelized Costs of Electricity (LCOE). Gearbox failures are a major contributor to maintenance costs, with High-Speed-Shaft (HSS) bearings having the highest failure rate. Frequently occurring premature HSS bearing damage indicates design inadequacies. Currently, bearing design standards (DIN/ISO 281) are primarily based on fatigue. Other damage mechanisms like adhesion or damages due to electrical currents passing through the bearings are not considered. WT gearbox bearings are exposed to electrical currents, mainly caused by common-mode (CM) voltages. Bearing currents lead to surface damages and also may promote the formation of white etching areas (WEA) or cracks (WEC) and thus need to be investigated further. The CM voltage is an inherent feature of the power converter, which generates CM currents in the electrical machine. The resulting high voltages at the HSS bearings can lead to electrical discharge currents in the rolling contact, which damage the raceways of the bearing. This paper presents an experimental method for characterization of the currents and voltages that pass from the generator to the HSS of the gearbox in a multi-megawatt WT drivetrain on a test bench. CM currents and voltages measured at the converter are correlated with the shaft voltages at the HSS bearing. The results indicate, that the CM currents directly pass the HSS and its bearings into the gearbox housing. The CM currents show peaks indicating discharge events with over 100 A. Furthermore, at the discharge events, high-frequency oscillations with over 20 MHz are observed in CM current and bearing voltage.

Boosting field data using synthetic SCADA datasets for wind turbine condition monitoring

State-of-the-art Deep Learning (DL) methods based on Supervisory Control and Data Acquisition (SCADA) system data for the detection and prognosis of wind turbine faults require large amounts of failure data for successful training and generalisation, which are generally not available. This limitation prevents benefiting from the superior performance of these methods, especially in SCADA-based failure prognosis. Data augmentation approaches have been proposed in the literature for generating failure data instances within a SCADA sequence to reduce the imbalance between healthy and faulty state data points, which is relevant to fault detection tasks. However, the successful implementation of DL-based failure prognosis methods requires the availability of multiple run-to-failure SCADA sequences. This paper proposes a data-driven method for generating synthetic run-to-failure SCADA sequences with custom operational and environmental conditions and progression of degradation. An Artificial Neural Network (ANN) is trained with signals that represent these factors to reconstruct the SCADA signals. Then, it is used to generate synthetic SCADA datasets based on data available from a wind turbine that experienced a gearbox failure. Synthetic data sets generated are evaluated on the basis of the similarity of their signal distributions, the temporal dynamics within each signal, and the temporal dynamics among different SCADA signals with those in similar field datasets. The results show that the generated synthetic datasets are consistent with their field counterparts, with a comparatively lower diversity in their dynamic behaviour in time.

Preprocessing and Modeling Approach for Gearbox Pitting Severity Prediction under Unseen Operating Conditions and Fault Severities

Gear pitting is a common gearbox failure mode that can lead to unplanned machine downtime, inefficient power transmission and a higher risk of sudden catastrophic failure. Consequently, there is strong incentive to create machine learning models that are capable of detecting and quantifying the severity of gearbox pitting faults. The performance of machine learning models is however highly dependent on the availability of training data and since training data for a wide variety of different operating conditions and fault severities is rarely available in practice, machine learning models must be designed to be robust to unseen operating conditions and fault severities. Furthermore, models should be capable of identifying data outside of the training data distribution and adjusting the confidence in a prediction accordingly. This work presents a strategy for pitting severity estimation in gearboxes under unseen operating conditions and fault severities in response to the PHM North America 2023 Conference Data Challenge. The strategy includes the design of dedicated validation sets for quantifying model performance on unseen data, an investigation into the most appropriate preprocessing methods, and a specialized convolutional neural network with an integrated out of distribution detection model for identifying samples from foreign operating conditions and fault severities. The results show that the best models are capable of some generalization to unseen operating conditions, but the generalization to unseen pitting severities is more challenging.

Feature-oriented unified dictionary learning-based sparse classification for multi-domain fault diagnosis

Aiming at the instability of sparse representation of multi-domain features caused by domain discrepancy and individual differences in feature sparse discriminative ability, the Feature-oriented Unified Dictionary Learning based Sparse Classification (FUDL-SC) is proposed for multi-domain fault diagnosis. Blending sources alignment is designed to unify the transformations of multiple domains at the sample global level to reduce the discrepancy in the data distribution across domains. The FUDL framework is established by learning dictionary atoms specific to each feature of different classes in the unified domain, thus generating a Feature-oriented Unified Dictionary (FUD). Reconstruction Scoring Matrix (RSM) is constructed to measure the sparse discrimination performance of individual FUD. The iterative update of the RSM is incorporated into the FUDL model, thus enhancing FUD's sparsity-preserving and discriminating ability in multiple domains. The efficacy of FUDL-SC is experimentally demonstrated on a multi-condition bearing failure dataset, a multi-device bearing failure dataset, and a small-sample gearbox failure dataset. The comparative studies verify that FUDL-SC can achieve higher recognition accuracy, better stability, and greater robustness than numerous state-of-the-art methods in multi-domain fault diagnosis.

A Study on a Model for Determining the Cause of Gantry Crane Gearbox Failures Using Machine Learning and Design of Experiments

A Study on a Model for Determining the Cause of Gantry Crane Gearbox Failures Using Machine Learning and Design of Experiments

Gear fault diagnosis based on complex network theory and error-correcting output codes: Multi class support vector machine

Gearbox failures have a detrimental effect on the machine performance that affects the production capacity and economic benefits of a manufacturing industry in an adverse manner. Early detection of faults in gears helps to prevent sudden machine failures. Previously, many signal processing and artificial intelligence techniques have been successfully applied to diagnose gear faults by analyzing the machine vibrations. These techniques have repeatedly emphasized on the need to extract quality features from gear vibration signals and using artificial intelligence techniques that can identify multiple types of faults simultaneously. As such, in this article, new features called as visibility graph features based on complex network theory have been proposed to extract fault information from vibration signals. In addition, a new artificial intelligence framework for multifault classification known as error-correcting output codes-multiclass support vector machine is proposed to detect gear faults. To the best of authors’ knowledge, these features along with the error-correcting output codes-artificial intelligence framework have been rarely utilized for fault diagnosis purpose. The proposed approach has been applied on experimental gear vibration data to validate its effectiveness. First, the vibration signals are transformed to graphs, and various graph properties are computed to serve as fault features. Then, the fault features are supplied to train the error-correcting output codes-multiclass support vector machine model and learn the fault patterns. Finally, the gear fault classification accuracies are determined for various gear test conditions and compared with those of existing methods. It is observed that the proposed approach provides an average improvement of 4.22%, 9.93% and 11.8% over the time-domain features, time–frequency domain features and voting-based multiclass support vector machine classifier, respectively, for different operating conditions. Furthermore, the suggested technique augments the classification accuracies by 10%, 7.5% and 6.7%, respectively, as compared with the deep-learning models, namely standard stacked sparse autoencoder, deep neural network and convolution neural network.

A deep learning approach for health monitoring in rotating machineries using vibrations and thermal features

Gearbox failures can lead to substantial damage, significant financial losses due to maintenance downtimes, and, in some instances, fatalities. This study introduces an intelligent gear fault diagnosis system employing a convolutional neural network (CNN), utilizing vibration and thermal features extracted from healthy, chipped, and broken tooth gear health categories. The performance of the CNN is compared with conventional machine learning models, including Naïve Bayes (NB), random forest (RF), and support vector machine (SVM) classifiers. Experimental investigations highlight CNN’s remarkable performance. With vibration features, the CNN achieved 96.78% accuracy, surpassing SVM (84.89%), NB (81.56%), and RF (85.11%). The CNN attained 100% accuracy when utilizing thermal features, while SVM, NB, and RF achieved 91.11%, 88.89%, and 96.51% accuracies, respectively.

Automatizacija praćenja stanja reduktora vetroturbina - uporedna studija tehnike mašinskog učenja zasnovane na analizi vibracija

Wind turbines play a role in the adoption of renewable energy production, but they are susceptible to shutdowns that require thorough monitoring. Gearbox failures are an issue leading to maintenance and operational downtime. This study investigates the application of machine learning methods to enhance the diagnosis of gearbox problems using vibration analysis. Through the application of fault scenarios that impact bearings and gears, the researchers successfully extracted time domain features from vibration data of a 750 kW turbine testbed in order to detect indications of damage. Support Vector Machine (SVM), Naive Bayes, and K Nearest Neighbour (KNN) machine learning models were used to classify gearbox faults. Among these models, Naive Bayes achieved an accuracy rate of 95.7%, which exceeded the established benchmarks. The probabilistic approach was able to successfully associate symptom characteristics with fault patterns. Intelligent monitoring systems could improve maintenance efficiency. This data-driven approach highlights the potential of machine learning in supporting wind power development by eliminating gearbox inefficiencies and improving turbine reliability, and further research is being conducted to ensure that this approach works in concert with diversity and in the real world. This shows how machine learning is contributing to advances in renewable energy by helping to analyze predictive problems and prevent costly gearbox failures.

Failures and methods of ensuring the service life of friction discs of gearboxes

The resource of a complex machine is determined by the resource of its units and components. As a result of the normal operation of energyintensive mobile vehicles of the K-7 series, it was established that the service life of the hydromechanical gearbox does not exceed 3000 engine hours, and the experimental probability of its failure is more than 0.55 of all machine units. The cause of failure of the hydromechanical gearbox is failure of the clutches and drive shaft. Failures of friction clutches are associated with significant overheating of the friction discs, their wear, welding and changes in spatial geometry. In the practice of normal operation of mobile machines, the drive shaft is repaired or replaced with a new one. Failure of friction discs occurs due to changes in spatial geometry (experimental probability of failure - 0.76), wear (experimental probability of failure - 0.63), destruction (experimental probability of failure - 0.38) and welding (experimental probability of failure - 0.34) friction discs. The work examines the use of the method of restoring friction discs by flat dimensional grinding, and conducts comparative life tests of restored friction discs.

Incipient fault detection of planetary gearbox under steady and varying condition

As an important component in rotating machines, gearbox failure will lead to costly economic losses. Generally, incipient fault features of gearbox are weak and concealed in a set of time-varying vibration signals that are challenging to identify effectively. Based on that, a new method is proposed for incipient fault detection(FD) under steady and variable conditions of gearboxes based on improved octave convolution in this paper. First, the vibration signals are transferred into images via the symmetrized dot pattern(SDP) method. Then, the proposed method enhances image detail learning by adding convolution kernels and introducing residual connections in the high-frequency component of the octave convolution. Meanwhile, self-attention units are introduced in the information interaction branches. After that, the improved octave convolution is applied to the ResNet50 backbone network(IOC-ResNet50) to mine deep fault features. Compared with other published methods, the results indicate that the proposed performs superiorly in both time-varying conditions and steady state.

Block feature selection based on NSGA-II applied to fault diagnosis of gearboxes

A hybrid model with multiple sub-classifier is a powerful classifier for intelligent fault diagnosis of gearboxes. Each sub-classifier is used to deal with different gear or bearing faults, and thus each should have its own optimal feature group. However, these sub-classifiers in the hybrid model used the same feature group in most reported feature selection methods. To overcome this problem, this study brings some insights into optimization on sub-classifiers. This proposed method, known as block feature selection (BFS), optimizes the feature group for only one fault type in each sub-classifier. The selected groups are then amalgamated to form the optimal features for all types of gearbox failures, with corresponding labels in the hybrid model. Each sub-classifier is considered to be one block. In each block, the concept of NSGA-II is used to select features for its corresponding sub-classifier. Then, a new sorting algorithm is proposed to find the best feature group of sub-classifier. In general, BFS can optimize features of each sub-classifier to find the best feature group for hybrid model. The effectiveness of BFS is validated by experimental data from bearing and gear faults in a planetary gearbox rig. The robustness of BFS is verified by incorporating white noise at different levels.

The effect of center distance error on the service life of polymer gears

Gear assembly errors are an important contributor to a premature gearbox failure. This study investigates the effects of center distance deviation on polymer gearbox life. The main focus was on the influence on the stress state of the gear. Reference spur gears (m = 1 mm, z = 20, b = 6 mm) with a standard involute gear profile and a gear ratio of 1 (same pinion and gear) were studied. A combination of numerical and experimental methods was used to obtain the best insight into the gear meshing process. Simulations of the gear meshing process were performed using a verified finite element analysis model. Torques of 0.8 Nm and 1.0 Nm were applied and gear meshing was simulated at center distances of 20.00 mm, 20.05 mm, and 20.10 mm. It was found that, as the center distance increased, the contact ratio decreased, which consequently leads to an increase in tooth root stress. Experimental gears tests were additionally conducted on a dedicated gear testing machine. The gears were tested under the same conditions previously simulated using the FEM model. Steel/PA6 gear pairs were tested, with at least three test repetitions for each combination of load and center distance. A significant reduction in gear service life of more than 30% was identified for larger center distances, while no pronounced effects on operating temperature were observed.

Best Practice Data Sharing Guidelines for Wind Turbine Fault Detection Model Evaluation

In this paper, a set of best practice data sharing guidelines for wind turbine fault detection model evaluation is developed, which can help practitioners overcome the main challenges of digitalisation. Digitalisation is one of the key drivers for reducing costs and risks over the whole wind energy project life cycle. One of the largest challenges in successfully implementing digitalisation is the lack of data sharing and collaboration between organisations in the sector. In order to overcome this challenge, a new collaboration framework called WeDoWind was developed in recent work. The main innovation of this framework is the way it creates tangible incentives to motivate and empower different types of people from all over the world to share data and knowledge in practice. In this present paper, the challenges related to comparing and evaluating different SCADA-data-based wind turbine fault detection models are investigated by carrying out a new case study, the “WinJi Gearbox Fault Detection Challenge”, based on the WeDoWind framework. A total of six new solutions were submitted to the challenge, and a comparison and evaluation of the results show that, in general, some of the approaches (Particle Swarm Optimisation algorithm for constructing health indicators, performance monitoring using Deep Neural Networks, Combined Ward Hierarchical Clustering and Novelty Detection with Local Outlier Factor and Time-to-failure prediction using Random Forest Regression) appear to exhibit high potential to reach the goals of the Challenge. However, there are a number of concrete things that would have to have been done by the Challenge providers and the Challenge moderators in order to ensure success. This includes enabling access to more details of the different failure types, access to multiple data sets from more wind turbines experiencing gearbox failure, provision of a model or rule relating fault detection times or a remaining useful lifetime to the estimated costs for repairs, replacements and inspections, provision of a clear strategy for training and test periods in advance, as well as provision of a pre-defined template or requirements for the results. These learning outcomes are used directly to define a set of best practice data sharing guidelines for wind turbine fault detection model evaluation. The guidelines can be used by researchers in the sector in order to improve model evaluation and data sharing in the future.