Reducing sensor complexity for monitoring wind turbine performance using principal component analysis

Wind turbine condition monitoring by the approach of SCADA data analysis

Deep Learning for fault detection in wind turbines

WTCM (Wind Turbine Condition Monitoring) system is important for wind farm operators to realize condition-based O &M (operation & maintenance), in the purpose of reducing O &M cost and improving wind turbine reliability. A WTCM method using only SCADA data based on data mining algorithm is proposed in this paper. Firstly, ARD (Automatic Relevance Determination) algorithm is adopted to determine the effective variables that are relevant to wind turbine condition. Feature vector is then extracted using the effective variables to represent the operation condition of wind turbine. Finally, the condition of a wind turbine is determined using outlier detection algorithm based on the extracted feature vector. Real-world dataset is used to validate the efficiency of the proposed method. Experiment results show that the proposed method can provide advanced failure alarm for wind turbines many days before failure happens. O &M cost can be reduced by condition-based O &M strategy using the result of our proposed WTCM method.

Condition monitoring (CM) has been recognized as one of the most promising and widely applicable approaches to increase the availability and reduce the operation and maintenance (O&M) costs of wind turbines (WTs). However, up to date the potential of CM in the wind industry has not been fully exploited due to manifold reasons including (1) the lack of cost-effective and universal strategies that can deal with the CM issues from various concepts of WTs and (2) the lack of efficient and robust algorithms that can accurately process and interpret CM signals collected from WTs. In addition, WTs are subject to constantly varying loads due to inconsistent wind. In order to mitigate the complex loads and maximize the output power, modern megawatt-scale WTs are designed as variable-speed and pitch-control machines. Consequently, the CM signals collected from large scale WTs are usually non-stationary over time and difficult to process accurately using conventional signal processing methods. Thus, there is an urgent need of advanced CM strategies and dedicated signal processing techniques for WTs. Here a novel WT drivetrain CM strategy and the associated signal processing method, namely, wavelet-transform-based energy tracking technique (WETT), are elaborated in this chapter. The WETT utilizes readily available generator power signal to evaluate the health condition of the whole WT drivetrain system through extracting and assessing the energy of WT power signals at fault characteristic frequencies. The WETT is verified through applying it to detecting the electrical and mechanical faults emulated on a WT drivetrain test rig. Experiment has shown that the WETT can correctly identify the simulated faults and is therefore potentially a successful tool for WT drivetrain CM.

Wind turbine condition monitoring is considered as a key task in wind power industry. A plethora of methodologies based on machine learning have been proposed but the absence of faulty data, at the amount and the variety needed, still set limitations. Therefore Anomaly Detection methodologies are proposed as alternatives for fault detection. Deep learning tools have been introduced in the research field of wind turbines’ monitoring for the purpose of higher detection accuracy. In this work, a deep learning-based anomaly detection method, the Deep Support Vector Data Description (Deep SVDD), is proposed for the monitoring of wind turbines. Compared to the classic SVDD anomaly detection approach, this method combines a deep network, more specifically a Convolutional Neural Network (CNN), with the SVDD detector in order to automatically extract effective features. To test and validate the effectiveness of the proposed method, the Deep SVDD method is applied on SCADA data from a real wind turbine use case, targeting to the ice detection on wind turbine blades. The experimental results show that the method can effectively detect the generation of ice on wind turbines’ blades with a successful detection rate of 91.45%.

Condition monitoring of wind turbine for rotor fault detection under non stationary conditions

Condition monitoring of gear-based mechanical systems undergoing non-stationary operation conditions is in general very challenging. In particular, this issue is remarkable as regards wind energy technology because most of the modern wind turbines are geared and gearbox damages account for at least the 20% of their unavailability time. In this work, a new method for the diagnosis of gearbox damages is proposed: the main idea is that vibrations are measured at the tower, instead that at the gearbox. This implies that measurements can be performed without impacting on the wind turbine operation, as desirable by the point of view of wind turbine practitioners. A test case study is discussed: it deals with a wind farm owned by Renvico, featuring 6 wind turbines with 2 MW of rated power each. The vibration measurements at a wind turbine suspected to be damaged and at reference wind turbines are processed through a multivariate Novelty Detection algorithm in the feature space. The application of this algorithm is justified by univariate statistical tests on the time-domain features selected and by a visual inspection of the data set via Principal Component Analysis. Finally, a novelty index based on the Mahalanobis distance is used to detect the anomalous conditions at the damaged wind turbine.

Improvement of condition monitoring (CM) systems for wind turbines (WTs) and reduction of the cost of wind energy are possible if knowledge about the condition of different WT components is available. CM based on the WT drive train shaft torque signal can give a better understanding of the gearbox failure mechanisms as well as provide a method for detecting mass imbalance and aerodynamic asymmetry. The major obstacle preventing the industrial application of CM based on the shaft torque signal is the costly measurement equipment which is impractical for long‐term use on operating WTs. This paper suggests a novel approach for low‐cost, indirect monitoring of the shaft torque from standard WT measurements. The shaft torque is estimated recursively from measurements of generator torque, high speed shaft and low speed shaft angular speeds using the well‐known Kalman filter theory. The performance of the augmented Kalman filter with fading memory (AKFF) is compared with the augmented Kalman filter (AKF) using simulated data of the WT for different load conditions, measurement noise levels and WT fault scenarios. A multiple‐model algorithm, based on a set of different Kalman filters, is designed for practical implementation of the shaft torque estimator. Its performance is validated for a scenario where there are frequent changes of operating points. The proposed cost‐effective shaft torque estimator overcomes all major problems, which prevent the industrial application of CM systems based on shaft torque measurements. Future work will be focused on validating the method using experimental data and developing suitable signal processing algorithms for fault detection. Copyright © 2013 John Wiley & Sons, Ltd.

As wind energy adoption is growing, ensuring the efficient operation and maintenance of wind turbines becomes essential for maximizing energy production and minimizing costs and downtime. Many AI applications in wind energy, such as in condition monitoring and power forecasting, may benefit from using operational data not only from individual wind turbines but from multiple turbines and multiple wind farms. Collaborative distributed AI which preserves data privacy holds a strong potential for these applications. Federated learning has emerged as a privacy-preserving distributed machine learning approach in this context. We explore federated learning in wind turbine condition monitoring, specifically for fault detection using normal behaviour models. We investigate various federated learning strategies, including collaboration across different wind farms and turbine models, as well as collaboration restricted to the same wind farm and turbine model. Our case study results indicate that federated learning across multiple wind turbines consistently outperforms models trained on a single turbine, especially when training data is scarce. Moreover, the amount of historical data necessary to train an effective model can be significantly reduced by employing a collaborative federated learning strategy. Finally, our findings show that extending the collaboration to multiple wind farms may result in inferior performance compared to restricting learning within a farm, specifically when faced with statistical heterogeneity and imbalanced datasets.

Some large wind turbines use a synchronous generator, directly-coupled to the turbine, and a fully rated converter to transform power from the turbine to the mains. This paper considers condition monitoring and diagnosis of mechanical and electrical faults in such a variable speed machine. A new condition monitoring technique is proposed in this paper, which removes the negative influence of variable wind in machine condition monitoring. This technique has a versatile function, able to detect both the mechanical and electrical faults in the wind turbine. Its effectiveness is validated by the experiments on a wind turbine condition monitoring test rig. Furthermore, a potential approach for diagnosing wind turbine drive-train mechanical faults using wind turbine generator electrical signals is introduced. The diagnosis of rotor imbalance in the wind turbine will be used as an illustrative example, heralding the detection of wind turbine electromechanical faults by power analysis. The paper offers a simpler and cheaper condition monitoring and fault diagnosis system for wind turbines.

A robust fleet-based anomaly detection framework applied to wind turbine vibration data

This paper proposes a two-stage data-driven methodology in condition monitoring (CM) of wind turbines (WTs). In the first stage, a fast and powerful network, namely a parallel generative adversarial network (PGAN) is proposed to resolve the problem of limited available information by generating artificial data. In the second stage, a robust deep network is designed based on a one-module Gabor filter oriented convolutional neural network and reformulation of a new loss function, namely robust deep Gabor network (RDGN). The experimental dataset of 3MW wind turbines in Ireland is used to verify the effectiveness of the proposed method and demonstrate the superiority of the proposed two-stage method in comparison with several state-of-the-art methods in terms of accuracy and reliability.

Chapter 2 - Wind turbine inspection and condition monitoring

This study details the development of a fully automated pipeline for the condition monitoring of wind turbine drive trains. Vibration data is collected using hardware designed and manufactured in-house and used directly to monitor the condition of the drive trains. The complex nature of wind turbine vibration signals, due to the large number of components and highly variable operating conditions, makes drive train condition monitoring a challenging task. This paper details the full data measurement and analysis flow from sensor to insights and proposes a hybrid automated pipeline with signal processing and data-driven techniques to address the complexity of dealing with wind turbine vibration data. The vibration signals are directly employed to estimate the wind turbine’s instantaneous angular speed to compensate for any rotation speed fluctuations. Pre-processing is performed on the speed-independent signals to evaluate condition indicators in both the time and spectral domain for the vibration signals and their envelopes. Machine learning is then employed to distinguish the healthy state of the machine from a faulty one using the computed condition indicators. Besides the scalar indicators, also two-dimensional vibration decompositions such as the cyclic spectral correlation maps are used as inputs to the machine learning pipeline. This comprehensive and automated approach ensures both an early and reliable fault detection. Experimental results demonstrate that the fully automated hybrid pipeline can effectively be used for fleet-based health tracking of offshore wind turbine drivetrains.

Condition monitoring for wind turbines is essential for the further development of wind farms. Currently, many of the works are focused on the installation of new sensors to predict turbine failures, which raises the cost of wind projects. Wind turbines operate in a wide variety of environmental conditions, such as different temperatures and wind speeds that vary throughout the year season. Typically, most or all of the data available in a turbine is healthy data (operation without failure), so data-driven supervised classification methods have data imbalance problems (more data from one class). Also, when historical pre-failure data do not exist, those methods cannot be used. Taking into account the aforementioned difficulties, the stated strategy in this work is based on a principal component analysis anomaly detector for main bearing failure prognosis and its contributions are: i) this methodology is based only on healthy SCADA data, ii) it works under different seasons of the year providing its usefulness, iii) it is based only on external variables and one temperature related to the element under diagnosis, thus avoiding data containing information from other fault types, iv) it accomplishes the main bearing failure prognosis (several months beforehand), and v) the performance of the proposed strategy is validated on a real in production wind turbine.