Wind Turbine Condition Research Articles

Methods for Advanced Wind Turbine Condition Monitoring and Early Diagnosis: A Literature Review

Condition monitoring and early fault diagnosis for wind turbines have become essential industry practice as they help improve wind farm reliability, overall performance and productivity. If not detected and rectified at early stages, some faults can be catastrophic with significant loss or revenue along with interruption to the business relying mainly on wind energy. The failure of Wind turbine results in system downtime and repairing or replacement expenses that significantly reduce the annual income. Such failures call for more systematized operation and maintenance schemes to ensure the reliability of wind energy conversion systems. Condition monitoring and fault diagnosis systems of wind turbine play an important role in reducing maintenance and operational costs and increase system reliability. This paper is aimed at providing the reader with the overall feature for wind turbine condition monitoring and fault diagnosis which includes various potential fault types and locations along with the signals to be analyzed with different signal processing methods.

A portable wind turbine condition monitoring system and its field applications

A portable wind turbine condition monitoring system and its field applications

An unsupervised spatiotemporal graphical modeling approach for wind turbine condition monitoring

Abstract The vast installment of wind turbines and the development of condition monitoring system provides large amounts of operational data for condition monitoring and health management, while the lack of labeled data becomes one of the major challenges for the data analytics. To address this issue, this work presents an unsupervised anomaly detection approach for wind turbine condition monitoring, where a spatiotemporal graphical modeling method, spatiotemporal pattern network (STPN), is applied to extract the spatial and temporal features between the variables in the system, and an energy-based model, stacked Restricted Boltzmann Machine (RBM) is used to capture the system-wide patterns and then applied for condition monitoring. Case studies on three data sets are carried out including: (1) anomaly detection on a benchmark model for fault detection and isolation, (2) anomaly detection on an experimental data set with the normal condition and 11 fault conditions and (3) online condition monitoring using real data from a wind farm in northwest China. The results show that the proposed approach is capable of detecting the anomalies without the need for labeling data.

Numerical modeling of the seismic performance of monopile supported wind turbines in sandy soils susceptible to liquefaction.

Since 1980, as wind farms have moved from coastal to offshore areas, the wind energy industry has been completely transformed which in turn has led to the increase in the construction of wind turbines. On the other hand, harsher offshore environmental conditions have led to larger lateral loads and anchorages applied to the wind turbines and specifically to their piles than other coastal and offshore structures. Thus, more solid piles are required to ensure proper rigidity and bearing capacity. Liquefaction is one of the most important seismic hazards through which various damages caused to different parts of wind turbines. In order to develop coastal and offshore structures in Iran, a study of liquefaction is of great importance due in part to the high risk of seismicity. In this study, the effect of liquefaction on seismic response of offshore wind turbines is examined taking advantage of a finite element model. To this end, all analyzes have been carried out in both occurrence and non-occurrence of the liquefaction, so that by comparing these two modes, the mechanisms affecting the seismic behavior of wind turbines are understood. As depth increases, the possibility of liquefaction is reduced due to higher pressure. Liquefaction is considered to a depth of 20 m and structural behavior is evaluated based on the level of seismic hazard, the thickness of the susceptible layers, soil compaction, the non-fluidizing top layer, the gradient of the earth, the thickness of the monopole, the dimensions of the wind turbine and different soil layering conditions. According to the mentioned factors, a comprehensive and parametric study of the behavior of wind turbines in seismic zones, and in different loading conditions, pile diameters and soil layering is carried out in soils prone to liquefaction. Since analyzes are performed in both occurrence and non-occurrence of the liquefaction, the number of analyzes and computational cost in this research becomes enormous. Therefore, there is a need for a highly effective software and a practical modeling method that will allow for this comprehensive study. Open Sees software and beam on nonlinear Winkler foundation approach are used to model the soil-pile-structure interaction. The minor differences observed in the laboratory values compared to the numerically calculated ones may refer to the fact that the chamber is not modeled. In the bottom layer, as the depth decreases, the elastic response spectra record larger values which are due to the resonance in the structure.

Analysis of extremely modulated faulty wind turbine data using spectral kurtosis and signal intensity estimator

Abstract The use of signal processing for condition monitoring of wind turbines data has been on-going since several decades. Failure in the analysis of high modulated data may make the machine break. An example of this is the reported real case of bearing failure on a Repower wind turbine, which could not be detected by currently applied methods. The machine had to be out of service immediately after a faulty bearing outer race was visually ascertained. Vibration dataset from this faulty machine was provided to facilitate research into wind turbines analysis and with the hope that the authors of this work can improve upon the existing techniques. In the response to this challenge, the authors of this paper proposed Spectral Kurtosis (SK) and Signal Intensity Estimator (SIE) as proven time-frequency fault indicators to tackle the question of data with different modulation rates. Extensive signal processing using time domain and time-frequency domain analysis was undertaken. It was concluded that SIE is well established mature approach and it provides a more reliable estimate of wind turbine conditions than conventional techniques such as SK, leading to better discrimination between “good” and “bad” machines.

A Novel Degradation Identification Method for Wind Turbine Pitch System

It’s difficult for traditional threshold value method to identify degradation of operating equipment accurately. An novel degradation evaluation method suitable for wind turbine condition maintenance strategy implementation was proposed in this paper. Based on the analysis of typical variable-speed pitch-to-feather control principle and monitoring parameters for pitch system, a multi input multi output (MIMO) regression model was applied to pitch system, where wind speed, power generation regarding as input parameters, wheel rotation speed, pitch angle and motor driving currency for three blades as output parameters. Then, the difference between the on-line measurement and the calculated value from the MIMO regression model applying least square support vector machines (LSSVM) method was defined as the Observed Vector of the system. The Gaussian mixture model (GMM) was applied to fitting the distribution of the multi dimension Observed Vectors. Applying the model established, the Degradation Index was calculated using the SCADA data of a wind turbine damaged its pitch bearing retainer and rolling body, which illustrated the feasibility of the provided method.

Wind Turbine Condition Monitoring Strategy through Multiway PCA and Multivariate Inference

This article states a condition monitoring strategy for wind turbines using a statistical data-driven modeling approach by means of supervisory control and data acquisition (SCADA) data. Initially, a baseline data-based model is obtained from the healthy wind turbine by means of multiway principal component analysis (MPCA). Then, when the wind turbine is monitorized, new data is acquired and projected into the baseline MPCA model space. The acquired SCADA data are treated as a random process given the random nature of the turbulent wind. The objective is to decide if the multivariate distribution that is obtained from the wind turbine to be analyzed (healthy or not) is related to the baseline one. To achieve this goal, a test for the equality of population means is performed. Finally, the results of the test can determine that the hypothesis is rejected (and the wind turbine is faulty) or that there is no evidence to suggest that the two means are different, so the wind turbine can be considered as healthy. The methodology is evaluated on a wind turbine fault detection benchmark that uses a 5 MW high-fidelity wind turbine model and a set of eight realistic fault scenarios. It is noteworthy that the results, for the presented methodology, show that for a wide range of significance, α ∈ [ 1 % , 13 % ] , the percentage of correct decisions is kept at 100%; thus it is a promising tool for real-time wind turbine condition monitoring.

Rotor-Current-Based Fault Diagnosis for DFIG Wind Turbine Drivetrain Gearboxes Using Frequency Analysis and a Deep Classifier

Fault diagnosis of drivetrain gearboxes is a prominent challenge in wind turbine condition monitoring. Many machine learning algorithms have been applied to gearbox fault diagnosis. However, many of the current machine learning algorithms did not provide satisfactory fault diagnosis results due to their shallow architectures. Recently, a class of machine learning models with deep architectures called deep learning has received more attention, because it can learn high-level features of inputs. This paper proposes a new fault diagnosis method for the drivetrain gearboxes of the wind turbines equipped with doubly-fed induction generators (DFIGs) using DFIG rotor current signal analysis. In the proposed method, the instantaneous fundamental frequency of the rotor current signal is first estimated to obtain the instantaneous shaft rotating frequency. Then, the Hilbert transform is used to demodulate the rotor current signal to obtain its envelope, and the resultant envelope signal contains fault characteristic frequencies that are in proportion to the varying DFIG shaft rotating frequency. Next, an angular resampling algorithm is designed to resample the nonstationary envelope signal to be stationary based on the estimated instantaneous shaft rotating frequency. After that, the power spectral density analysis is performed on the resampled envelope signal for the gearbox fault detection. Finally, a classifier with a deep architecture that consists of a stacked autoencoder and a support vector machine is proposed for gearbox fault classification using extracted fault features. Experimental results obtained from a DFIG wind turbine drivetrain test rig are provided to verify the effectiveness of the proposed method.

Sensor Fault Detection and Isolation for a Wireless Sensor Network-Based Remote Wind Turbine Condition Monitoring System

To improve the reliability of wind turbines, various condition monitoring systems (CMSs) have been developed and most of them transmit data using wired communication channels. Recently, wireless sensor networks (WSNs) have been used to transmit data in wind turbine CMS due to the low cost and easy deployment feature of WSNs. However, since wind turbines are installed in harsh environments, the sensors and sensor nodes used in the WSN-based wind turbine CMSs are easily subject to faults, leading to corruption of the signals used for condition monitoring, which decreases the reliability of the CMS. This paper proposes a three-stage method for the detection and isolation of three most common sensor faults, i.e., SHORT fault, CONSTANT fault, and NOISE fault, in WSN-based wind turbine CMS. The proposed sensor fault detection and isolation greatly increases the accuracy and reliability of wind turbine CMSs. Data collected from wind turbines in the field are used to validate the effectiveness of the proposed method.

Wind turbine condition monitoring system as a source of diagnostic information

The issues relating to the necessity of monitoring the wind turbines’ condition and operation are presented in the study. The wind turbines of high powers, are very expensive and complicated machines requiring appropriate control and high quality technical service. The idea of appropriate technical turbine’s maintaining, combines its high availability and productivity, as well as minimization of the costs related to failures and unexpected damages. Remote diagnostic systems allow obtaining the basic information, making it possible to maintain and appropriate control, use and high quality technical service.

Calculating Steady-State Operating Conditions for DFIG-Based Wind Turbines

This paper proposes a noniterative approach to calculate steady-state operating conditions of doubly fed induction generator (DFIG)-based wind turbines. The proposed approach can be applied to calculate steady-state operating conditions for the full-order dynamic model of DFIG-based wind turbines under d–q reference frame. Lossy DFIG back-to-back converters can be well supported, and nonzero reactive power delivery through DFIG grid-side converters can be handled. This noniterative approach encounters no convergence issues, which guarantees fast computation speed. It provides analytical solutions to the steady-state operating conditions of DFIG-based wind turbines. The proposed method could potentially be applied to obtain steady-state operating conditions before performing eigenvalue analysis and electromagnetic simulations for wind farms. Simulations performed on a single-machine-infinite-bus test system and a multibus test system verify the effectiveness of the proposed method.

Condition Monitoring of Wind Turbine Based on Copula Function and Autoregressive Neural Network

The traditional wind turbine fault monitoring is often based on a single monitoring signal without considering the overall correlation between signals. A global condition monitoring method based on Copula function and autoregressive neural network is proposed for this problem. Firstly, the Copula function was used to construct the binary joint probability density function of the power and wind speed in the fault-free state of the wind turbine. The function was used as the data fusion model to output the fusion data, and a fault-free condition monitoring model based on the auto-regressive neural network in the faultless state was established. The monitoring model makes a single-step prediction of wind speed and power, and statistical analysis of the residual values of the prediction determines whether the value is abnormal, and then establishes a fault warning mechanism. The experimental results show that this method can provide early warning and effectively realize the monitoring of wind turbine condition.

Physics-based data analysis for wind turbine condition monitoring

Abstract This article presents methodologies for improving wind turbine condition monitoring using physics-based data analysis techniques. The unique operating conditions of the wind turbine drivetrain are described, and the complex kinematics of the gearbox is analyzed in detail. The pros and cons of the current wind turbine condition monitoring system (CMS) are evaluated. To improve the wind turbine CMS capability, it is suggested to use linear models with unsteady excitations, instead of using nonlinear and nonstationary process models, when dealing the wind turbine dynamics response model. An analysis is undertaken of the damage excitation mechanisms cause for various components in a gearbox, especially for those associated with lower-speed shafts. Physics (mechanics)-based data analysis methods are presented for different component damage excitation mechanisms. Validation results, using the wind farm and manufacturing floor data, are reported.

Mooring design and selection for floating offshore wind turbines on intermediate and deep water depths

The paper focuses on design and selection of mooring systems for Floating Offshore Wind Turbines (FOWTs) for intermediate and deep water depths, based on ultimate, accidental and fatigue limit state design conditions, with reference to power production, parked wind turbine and mooring line fault conditions. Platform static and dynamic offsets are determined by a purposely developed 6DOF programme in the frequency domain, accounting for first and second order motions. The Dutch tri-floater platform, together with the well-known 5 MW NREL wind turbine, is assumed as reference design. Besides, two sites, the one in the Southern North Sea, near Dogger Bank on water depths ranging from 50 up to 80 m, the other one in the Northern North Sea, near the Troll field area on water depths ranging from 200 up to 350 m, are assumed as candidate deployment areas, to investigate the effect of: (i) line number, (ii) platform admissible offset and line scope, (iii) offshore/onshore planned maintenance inspections and (iv) mooring technology on relevant installation and maintenance costs. A preliminary cost analysis is also performed and some recommendations are given to correctly choice the most suitable mooring configuration, as a function of water depth.

Harmonic Assessment for Wind Parks Based on Sensitivity Analysis

The wind power generation is currently a promising renewable energy technology. A critical feature of wind park is the existence of large distributed capacitance because of the cable laying and offshore cable transmission. This phenomenon is likely to cause harmonic resonance issues, and may threat the safety operation of wind parks. Meanwhile, the operation condition of wind turbines and the topology of wind parks affect the system structure and parameters, then affect the system harmonic resonance. Therefore, a dynamic harmonic model of the most widely used doubly fed induction generator was established to study the behaviors of harmonic resonance. At the same time, other primary electrical components in wind parks are also considered, such as power cables and transformers. In this paper, the harmonic resonance of the wind park is investigated based on the measured data and the modal sensitivity analysis. With considering the impact of various parameters on the frequency and amplitude of resonance, the sensitivity indices for these parameters are also calculated. Finally, the frequency shift method is introduced to solve resonance issue.

Nacelle-based Lidar Measurements for the Calibration of a Wake Model at Different Offshore Operating Conditions

Commonly, wake models are calibrated in wind tunnels or using flow simulations with a wide degree of physical details. In general, it is assumed that these methods cannot fully reproduce the real operating conditions of wind turbines. This research aims at investigating the calibration of an analytical single wake model in relation to full-scale measurements. Within this scope, we fitted the wake model to wake measurements realised with a lidar installed on the nacelle of an offshore wind turbine. We studied the parameters returned by the fit separating cases at different levels of atmospheric turbulence and thrust on the wind turbine rotor. Comparing the results with a published calibration based on few LES wind fields representative for partial load conditions, we achieved good agreement when the considered wind turbines operated in similar conditions. For other situations, i.e. at full load, we found different calibrations of the model parameters. Our results show that and how nacelle-based lidar measurements can be complementary in the development of wake models.

Optimisation of Data Acquisition in Wind Turbines with Data-Driven Conversion Functions for Sensor Measurements

Operation and Maintenance (O&M) is an important cost driver of modern wind turbines. Condition monitoring (CM) allows the implementation of predictive O&M strategies helping to reduce costs. In this work a novel approach for wind turbine condition monitoring is proposed focusing on synergistic effects of coexisting sensing technologies. The main objective is to understand the predictability of signals using information from other measurements recorded at different locations of the turbine. The approach is based on a multi-step procedure to pre-process data, train a set of conversion functions and evaluate their performance. A subsequent sensitivity analysis measuring the impact of the input variables on the predicted response reveals hidden relationships between signals. The concept feasibility is tested in a case study using Supervisory Control And Data Acquisition (SCADA) data from an offshore turbine.

Use of Seismic Analyses for the Wind Energy Industry

This paper proposes new seismic-based methods for use in the wind energy industry with a focus on wind turbine condition monitoring. Fourteen Streckeisen STS-2 Broadband seismometers and two three-dimensional (3D) sonic anemometers are deployed in/near an operating wind farm to collect the data used in these proof-of-principle analyses. The interquartile mean (IQM) value of power spectral density (PSD) of the seismic components in 10 min time series is used to characterize the spectral signatures (i.e., frequencies with enhanced variance) in ground vibrations deriving from vibrations of wind turbine subassemblies. A power spectral envelope approach is taken in which the probability density function (PDF) of seismic PSD is developed using seismic data collected under normal turbine operation. These power spectral envelopes clearly show the energy distribution of wind-turbine-induced ground vibrations over a wide frequency range. Singular PSD lying outside the power spectral envelopes can be easily identified and is used herein along with supervisory control and data acquisition (SCADA) data to diagnose the associated suboptimal turbine operating conditions. Illustrative examples are given herein for periods with yaw misalignment and excess tower acceleration. It is additionally shown that there is a strong association between drivetrain acceleration and seismic spectral power in a frequency band of 2.5–12.5 Hz. The long-term goal of the research is development of seismic-based condition monitoring (SBCM) for wind turbines. The primary advantages of SBCM are that the approach is low-cost, noninvasive, and versatile (i.e., one seismic sensor monitoring multiple turbine components).

Special issue: Wind turbine dynamic modeling, condition monitoring and diagnosis

Special issue: Wind turbine dynamic modeling, condition monitoring and diagnosis

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

Abstract Condition monitoring of Wind Turbine based induction generators still faces challenges due to their operation in harsh and aggressive environments. The detection of anomalies at an initial stage is necessary for reducing the downtime of Wind Turbines. Depending on wind speed, the wind turbine generator operates continuously in non-stationary conditions. Under non-stationary conditions, the rotor fault frequency components present at generator output spreads out on bandwidth proportional to the speed which complicates its detectability. Therefore, this paper proposes a potential approach based on the use of wavelet analysis for diagnosing rotor eccentricity fault using WTG electrical signals. Its effectiveness under speed and load varying conditions is validated by experiments on a wind turbine condition monitoring test rig. The effectiveness of the setup was benchmarked with already proven WT test rigs.