A current development trend in wind energy is characterized by the installation of wind turbines (WT) with increasing rated power output. Higher towers and larger rotor diameters increase rated power leading to an intensification of the load situation on the drive train and the main gearbox. However, current main gearbox condition monitoring systems (CMS) do not record the 6‑degree of freedom (6-DOF) input loads to the transmission as it is too expensive. Therefore, this investigation aims to present an approach to develop and validate a low-cost virtual sensor for measuring the input loads of a WT main gearbox. A prototype of the virtual sensor system was developed in a virtual environment using a multi-body simulation (MBS) model of a WT drivetrain and artificial neural network (ANN) models. Simulated wind fields according to IEC 61400‑1 covering a variety of wind speeds were generated and applied to a MBS model of a Vestas V52 wind turbine. The turbine contains a high-speed drivetrain with 4‑points bearing suspension, a common drivetrain configuration. The simulation was used to generate time-series data of the target and input parameters for the virtual sensor algorithm, an ANN model. After the ANN was trained using the time-series data collected from the MBS, the developed virtual sensor algorithm was tested by comparing the estimated 6‑DOF transmission input loads from the ANN to the simulated 6‑DOF transmission input loads from the MBS. The results show high potential for virtual sensing 6‑DOF wind turbine transmission input loads using the presented method.
White etching crack (WEC) failure is a failure mode that affects bearings in many applications, including wind turbine gearboxes, where it results in high, unplanned maintenance costs. WEC failure is unpredictable as of now, and its root causes are not yet fully understood. While WECs were produced under controlled conditions in several investigations in the past, converging the findings from the different combinations of factors that led to WECs in different experiments remains a challenge. This challenge is tackled in this paper using machine learning (ML) models that are capable of capturing patterns in high-dimensional data belonging to several experiments in order to identify influential variables to the risk of WECs. Three different ML models were designed and applied to a dataset containing roughly 700 high- and low-risk oil compositions to identify the constituting chemical compounds that make a given oil composition high-risk with respect to WECs. This includes the first application of a purpose-built neural network-based feature selection method. Out of 21 compounds, eight were identified as influential by models based on random forest and artificial neural networks. Association rules were also mined from the data to investigate the relationship between compound combinations and WEC risk, leading to results supporting those of previous analyses. In addition, the identified compound with the highest influence was proved in a separate investigation involving physical tests to be of high WEC risk. The presented methods can be applied to other experimental data where a high number of measured variables potentially influence a certain outcome and where there is a need to identify variables with the highest influence.
The ongoing trend of building larger wind turbines (WT) to reach greater economies of scale is contributing to the reduction in cost of wind energy, as well as the increase in WT drivetrain input loads into uncharted territories. The resulting intensification of the load situation within the WT gearbox motivates the need to monitor WT transmission input loads. However, due to the high costs of direct measurement solutions, more economical solutions, such as virtual sensing of transmission input loads using stationary sensors mounted on the gearbox housing or other drivetrain locations, are of interest. As the number, type, and location of sensors needed for a virtual sensing solutions can vary considerably in cost, in this investigation, we aimed to identify optimal sensor locations for virtually sensing WT 6-degree of freedom (6-DOF) transmission input loads. Random forest (RF) models were designed and applied to a dataset containing simulated operational data of a Vestas V52 WT multibody simulation model undergoing simulated wind fields. The dataset contained the 6-DOF transmission input loads and signals from potential sensor locations covering deformations, misalignments, and rotational speeds at various drivetrain locations. The RF models were used to identify the sensor locations with the highest impact on accuracy of virtual load sensing following a known statistical test in order to prioritize and reduce the number of needed input signals. The performance of the models was assessed before and after reducing the number of input signals required. By allowing for a screening of sensors prior to real-world tests, the results demonstrate the high promise of the proposed method for optimizing the cost of future virtual WT transmission load sensors.
White Etching Cracks (WEC) in gearbox bearings is a major concern in the wind turbine industry, which can lead to a premature failure of the gearbox. Though many hypotheses regarding the generation of WEC have been proposed over the decades, the answer is still disputable. To trace back the failures to earlier stages before they occur, an innovative sensor-set has been utilized on a test rig to monitor the influencing factors that lead to WEC. This paperwork seeks to recognize abnormal patterns from recorded sensor data and derive statements of sensible sensor combinations in WEC early detection. A Long Short Term Memory (LSTM) network-based autoencoder is proposed for the anomaly detection (AD) task. Employing an auto-associative sequence-to-sequence predictor, a model is trained to reconstruct the normal time series data without WEC. The reconstruction error of testing time series data is evaluated for the determination of its anomaly. The results show that the specified LSTM autoencoder framework can qualitatively distinguish anomalies from collected multivariate time series data. Moreover, the anomaly score evaluated via reconstruction-error-based metrics can discriminate normal and abnormal behaviors in the study. This investigation’s results entail a significant step towards early WEC risk detection and more cost-efficient wind turbine technology if this approach can be further applied on stream data with plausible thresholds in monitoring system.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.