Figure - available from: Microsystem Technologies
This content is subject to copyright. Terms and conditions apply.
Source publication
In this work, is presented the fabrication and characterization of optical fiber refractometer based on lossy mode resonances for turbine lubricant degradation. Three different optical fiber refractometers have been fabricated by means of the deposition of indium oxide thin-films. The sensitivity of the devices has been characterized as a function...
Citations
... Numerous condition monitoring technologies for wind turbines have been developed to date, with the most common being vibration analysis [2], acoustic analysis [3], lubrication oil analysis [4], and others. However, these methods are limited in practical wind farm applications due to the requirement for additional measurement sensors and data acquisition equipment. ...
Condition monitoring of wind turbines is critical for increasing the reliability of the turbines and reducing their operation and maintenance costs. Supervisory control and data acquisition (SCADA) systems have been widely regarded as a promising technique to monitor the health status of turbines due to their abundance and cost-effective operation data. However, SCADA data are fundamentally multivariate time series with inherent spatio-temporal correlations. Therefore, it is still difficult to extract such correlations and then accurately identify the health status. This paper proposes a novel multi-view spatio-temporal feature fusion approach (MVSTCNN) based on convolutional neural networks (CNN) for condition monitoring of wind turbines. Specifically, multiple CNN modules with convolutional kernels of varying sizes are designed to extract correlations among several sensor variables and the temporal dependency concealed in each variable in parallel. A main advantage of the proposed method is its capacity to capture multiscale local information and global information simultaneously in both temporal and spatial dimensions, which improves the performance of condition monitoring. Real SCADA data from a wind farm is utilized to evaluate the effectiveness and superiority of the proposed approach. The result demonstrates that the proposed approach is effective for early fault detection in wind turbines.
... The picture also shows that maintenance works of wind turbines gearbox need 18.38 days to finish it which is in the second for per failure due to their size and robust link to other components making it harder to access, repair, or even replace [6]. Such long maintenance time and according another research [7][8][9] can fully illustrates above opinion that it is hard to repair the faults of wind turbine gearbox in the wind field. Moreover, the expected operation life of gearbox is usually publicized as 20 years from merchant. ...
As an renewable and clean energy of the world, wind energy has gained more and more attention and its fault diagnosis becomes more and more important. The gearbox, as the kernel component of the wind turbine system, it’s robust conditions have a great influence on the whole wind turbines system. Wind turbine gearbox has complex structure, which is usually composed of solar planetary gearbox and cylindrical gearbox. In the process of operation, various kinds of faults easily occur, resulting in serious losses. Once the wind turbine gearbox is not functioning as smoothly as it could be, it may result in large economic losses for the company and owner. At the same time, the failure rate of wind turbine gearbox has always been high because of complicated mechanic structure and special motion. Therefore, the tasks of reducing the downtime and increasing the productivity of wind turbine gearbox are urgent. This paper reviewed some research results of faults diagnosis on wind turbines gearbox, such as time-frequency analysis method, vibration based methods, nondestructive testing methods, etc. Meanwhile, this paper finds out some key problems and the channel of the resolution of the issue in order to supply some information for the further research of wind turbines gearbox.
... Dans le domaine de l'éolien la maintenance conditionnelle peut être effectuée selon les approches suivantes : -L'analyse de la lubrification [14,15,16,17,18,19,20], -L'analyse des émissions acoustiques [21,22,23,24,25,26], -L'analyse des vibrations [27,28,29,30,31,32,33,34,35,36], -L'analyse des signaux électriques [37,38,39,40]. ...
Cette thèse est consacrée à la détection de défauts mécaniques dans les machines synchrones à partir des mesures électriques à vitesse variable. L'application visée est l'éolien. L'approche proposée est basée sur les méthodes d'order tracking dans lesquelles les signaux d'analyse sont échantillonnés en fonction de l'angle mécanique. Dans ce cas, les composantes spectrales deviennent indépendantes de la vitesse de rotation et l'analyse fréquentielle peut être exploitée. L'order tracking est généralement mis en œuvre à partir d'une mesure de position. Dans ce travail nous étudions des méthodes permettant d'estimer cette position à partir des mesures électriques (courants et tensions). Plusieurs méthodes sont proposées et classées en fonction du nombre de mesures disponibles. Elles sont ensuite comparées par rapport à leur aptitude à isoler la composante de défaut recherchée par order tracking. Ces méthodes sont testées en simulation et sur un banc d'essai éolien du LIAS. Dans un second temps, une méthode statistique est mise en œuvre pour finaliser le diagnostic. A vitesse variable, ce type de méthode est difficile à mettre en œuvre et nous montrons que l'order tracking permet de simplifier l'analyse.
... There are some other studies on condition monitoring and fault detection of the WT using signal-based [29][30][31][32], knowledge-based [33][34][35][36][37], and data-driven based methods [38][39][40][41][42][43], which relies on the analysis of historical data, and the health condition can be predicted accurately, these methods are innovative and provides theoretical reference for further research. Moreover, due to the critical role the lubrication system plays, methods for lubrication analysis in wind turbine gearbox had been proposed, such as Lossy Mode Resonance (LMR) based optical fiber refractometer [44], Online lubrication oil condition monitoring and remaining useful life prediction using viscosity and dielectric constant sensors along with a particle filtering technique [45,46], oil debris count and size detection [47]. ...
To reduce the operation and maintenance (O&M) cost with optimized O&M strategy, this paper proposes a fuzzy synthetic method for real-time condition assessment of wind turbine gearbox (WTG). Firstly, a real-time condition assessment framework is proposed with statistical analysis and fuzzy mathematics. Then, the dynamic threshold value and variable weight are obtained to evaluate the operating status of the WTG, after filtering and statistical analysis of the data samples which came from two-month of running data of twenty-five wind turbines in 2017. Case studies are performed using 12 groups of actual monitoring data from 2 megawatts (MW) wind turbines with the proposed fuzzy synthetic method. Following this, this paper compares the proposed model to the traditional fuzzy assessment method. The results show that a strong correlation exists between variables and the condition assessment through the proposed method, which can be used to predict the practical operating status of the WTG more accurately, especially the potential failures.
... Results showed that a faulty gear tooth could be identified. Based on Lossy Mode Resonance (LMR), Sanchez et al. [7] designed an optical fibre refractometer to monitor gearbox lubricant degradation at temperature operating conditions of WT's. Study showed that the use of thinner coating leads to higher sensitivity in the results. ...
Today, condition monitoring (CM) is unarguably the most important field in any industrial applications. CM of wind turbines (WT’s) has in the past few years grown substantially. Although numerous initiatives to develop CM techniques and make operations more efficient were launched, most developed tools failed to respond on time to unpredictable events. One area that shows great potential in the battle against machine damages and their exploits is the diagnosis and prognosis of WT gears. In the world of big varying modulated data, analysis of health conditions of WT gears by traditional CM methods is no longer sufficient. Example for this is the high dimensionality and very extremely modulated vibration dataset, provided by Suzlon company. Suzlon unworkably attempted to online discriminate its machines using a set of well-known CM analysis methods. However, only visual inspection could identify the faulty WT gear. Hence, Suzlon flagged up a top priority need to identify more efficient online tools for improving CM processes. In the response to this essential need, the author employs Signal Intensity Estimator (SIE) method and some machine learning (ML) algorithms to analyse Suzlon dataset. A conclusion was reached that these techniques could successfully provide a reliable estimate of WT's conditions.
... Reference [6] proposes a method based Autocorrelation Time Synchronization Average (ATSA) to deal with the physical interaction between the ring and the gear in the gearbox. In reference [7] , a fiber refractometer is fabricated and the degradation of the gearbox synthetic lubricant is characterized. Physical models have an advantage in predicting long-term trends in components. ...
Wind turbines condition monitoring and fault warning have important practical value for wind farms to reduce maintenance costs and improve operation levels. Due to the increase in the number of wind farms and turbines, the amount data of wind turbines have increased dramatically. This problem has caused a need for efficiency and accuracy in monitoring the operation condition of the turbine. In this paper, the idea of deep learning is introduced into the wind turbine condition monitoring. After selecting the variables by the method of adaptive elastic network, the Convolutional Neural Network (CNN) and the Long and Short Term Memory network (LSTM) are combined to establish the logical relationship between observed variables. Based on training data and hardware facilities, the method is used to process the temperature data of gearbox bearing. The purpose of artificial intelligence monitoring and over-temperature fault warning of high-speed side of bearing is realized efficiently and conveniently. Example analysis experiments verify the high practicability and generalization of the proposed method.
... Vibration analysis and oil monitoring have become two commonly used techniques for wind turbine condition monitoring [3][4][5][6]. However, both techniques are sophisticated and expensive in their practical application, since additional investments, including installing extra sensors and data acquisition devices, are required. ...
Health monitoring and early fault detection of wind turbines have attracted considerable attention due to the benefits of improving reliability and reducing the operation and maintenance costs of the turbine. However, dynamic and constantly changing operating conditions of wind turbines still pose great challenges to effective and reliable fault detection. Most existing health monitoring approaches mainly focus on one single operating condition, so these methods cannot assess the health status of turbines accurately, leading to unsatisfactory detection performance. To this end, this paper proposes a novel general health monitoring framework for wind turbines based on supervisory control and data acquisition (SCADA) data. A key feature of the proposed framework is that it first partitions the turbine operation into multiple sub-operation conditions by the clustering approach and then builds a normal turbine behavior model for each sub-operation condition. For normal behavior modeling, an optimized deep belief network is proposed. This optimized modeling method can capture the sophisticated nonlinear correlations among different monitoring variables, which is helpful to enhance the prediction performance. A case study of main bearing fault detection using real SCADA data is used to validate the proposed approach, which demonstrates its effectiveness and advantages.
... Soon after this, In 2 O 3 coated optical fibers refractometers showing a repeated response with maximum sensitivity of 4926 and 4255 nm/RIU for TE and TM resonances respectively was also fabricated by Zamarreno et al. [21] , with the advantage of permitting realization of dual-peak reference measurements in different regions of the spectrum Fig. 7 . In 2015, Sanchez put forward In 2 O 3 coated optical fibers by a DC-Sputter deposition technology [59] . The 1st LMR experiments These values correspond to sensitivities of 5680 nm/RIU. ...
... During the next 2015 years, there are three papers about temperature sensor based on LMR applied in measuring degradation of industrial lubricant [59,92,93] . In industrial lubricating oil can reduce the gear wear and tear and prolonging the service life. ...
... From movement of resonance wavelength, degree of degradation of lubricating oil can be obtained. The sensitivity of the LMR sensors is 0.675 nm/°C for thin film In 2 O 3 [59] in the temperature range between 60 °C-100 °C, and 1.26 and 2.2 nm/°C for ITO [92] and SnO 2 [93] , respectively. At the same time, because of different refractive index of different lubricating oil, this kind of sensors can be used to identify lubricant. ...
This review paper presents the achievements and present developments in lossy mode resonances-based optical fiber sensors in different sensing field, such as physical, chemical and biological, and briefly look forward to its future development trend in the eyes of the author. Lossy mode resonances (LMR) is a relatively new physical optics phenomenon put forward in recent years. Fiber sensors utilizing LMR offered a new way to improve the sensing capability. LMR fiber sensors have diverse structures such as D-shaped, cladding-off, fiber tip, U-shaped and tapered fiber structures. Major applications of LMR sensors include refraction sensors and biosensors. LMR-based fiber sensors have attracted considerable research and development interest, because of their distinct advantages such as high sensitivity and label-free measurement. This kind of sensor is also of academic interest and many novel and great ideas are continuously developed.
... The layer morphology is compact but also rod-like (see Figure 25). A sensor prepared this manner was successfully evaluated for monitoring the aging of gear box oil as a function of its refractive index change [116]. Titania has been also employed to prepare optical fibre sensors. ...
... The resulting coating is homogeneous and it is formed by compact nanorods. Reprinted[116] from with permission from Springer. ...
The measurement of chemical and biomedical parameters can take advantage of the features exclusively offered by optical fibre: passive nature, electromagnetic immunity and chemical stability are some of the most relevant ones. The small dimensions of the fibre generally require that the sensing material be loaded into a supporting matrix whose morphology is adjusted at a nanometric scale. Thanks to the advances in nanotechnology new deposition methods have been developed: they allow reagents from different chemical nature to be embedded into films with a thickness always below a few microns that also show a relevant aspect ratio to ensure a high transduction interface. This review reveals some of the main techniques that are currently been employed to develop this kind of sensors, describing in detail both the resulting supporting matrices as well as the sensing materials used. The main objective is to offer a general view of the state of the art to expose the main challenges and chances that this technology is facing currently.
An increasing number of deep autoencoder-based algorithms for intelligent condition monitoring and anomaly detection have been reported in recent years to improve wind turbine reliability. However, most existing studies have only focused on the precise modeling of normal data in an unsupervised manner; few studies have utilized the information of fault instances in the learning process, which results in suboptimal detection performance and low robustness. To this end, we first developed a deep autoencoder enhanced by fault instances, that is, a triplet-convolutional deep autoencoder (triplet-Conv DAE), jointly integrating a convolutional autoencoder and deep metric learning. Aided by fault instances, triplet-Conv DAE can not only capture normal operation data patterns but also acquire discriminative deep embedding features. Moreover, to overcome the difficulty of scarce fault instances, we adopted an improved generative adversarial network-based data augmentation method to generate high-quality synthetic fault instances. Finally, we validated the performance of the proposed anomaly detection method using a multitude of performance measures. The experimental results show that our method is superior to three other state-of-the-art methods. In addition, the proposed augmentation method can efficiently improve the performance of the triplet-Conv DAE when fault instances are insufficient.
