Search for: All records

In cold climates, ice formation on wind turbines causes power reduction produced by a wind farm. This paper introduces a framework to predict icing at the farm level based on our recently developed Temporal Convolutional Network prediction model for a single turbine using SCADA data.First, a cross-validation study is carried out to evaluate the extent predictors trained on a single turbine of a wind farm can be used to predict icing on the other turbines of a wind farm. This fusion approach combines multiple turbines, thereby providing predictions at the wind farm level. This study shows that such a fusion approach improves prediction accuracy and decreases fluctuations across different prediction horizons when compared with single-turbine prediction. Two approaches are considered to conduct farm-level icing prediction: decision fusion and feature fusion. In decision fusion, icing prediction decisions from individual turbines are combined in a majority voting manner. In feature fusion, features of individual turbines are averaged first before conducting prediction. The results obtained indicate that both the decision fusion and feature fusion approaches generate farm-level icing prediction accuracies that are 7% higher with lower standard deviations or fluctuations across different prediction horizons when compared with predictions for a single turbine. 

Wind tunnel experiments were performed to quantify the coupling mechanisms between incoming wind flows, power output fluctuations, and unsteady tower aerodynamic loads of a model wind turbine under periodically oscillating wind environments across various yaw misalignment angles. A high-resolution load cell and a data logger at high temporal resolution were applied to quantify the aerodynamic loads and power output, and time-resolved particle image velocimetry system was used to characterize incoming and wake flow statistics. Results showed that due to the inertia of the turbine rotor, the time series of power output exhibits a distinctive phase lag compared to the incoming periodically oscillating wind flow, whereas the phase lag between unsteady aerodynamic loads and incoming winds was negligible. Reduced-order models based on the coupling between turbine properties and incoming periodic flow characteristics were derived to predict the fluctuation intensity of turbine power output and the associated phase lag, which exhibited reasonable agreement with experiments. Flow statistics demonstrated that under periodically oscillating wind environments, the growth of yaw misalignment could effectively mitigate the overall flow fluctuation in the wake region and significantly enhance the stream-wise wake velocity cross correlation intensities downstream of the turbine hub location. 

Summary Wake steering is very effective in optimizing the power production of an array of turbines aligned with the wind direction. However, the wind farm behaves as a porous obstacle for the incoming flow, inducing a secondary flow in the lateral direction and a reduction of the upstream wind speed. This is normally referred to as blockage effect. Little is known on how the blockage and the secondary flow influence the loads on the turbines when an intentional yaw misalignment is applied to steer the wake. In this work, we assess the variation of the loads on a virtual 4 by 4 array of turbines with intentional yaw misalignment under different levels of turbulence intensity. We estimate the upstream distance at which the incoming wind is influenced by the wind farm, and we determine the wind farm blockage effect on the loads. In presence of low turbulence intensity in the incoming flow, the application of yaw misalignment was found to induce a significant increase of damage equivalent load (DEL) mainly in the most downstream row of turbines. We also found that the sign (positive or negative) of the yaw misalignment affects differently the dynamic loads and the DEL on the turbines. Thus, it is important to consider both the power production and the blade fatigue loads to evaluate the benefits of intentional yaw misalignment control especially in conditions with low turbulence intensity upstream of the wind farm. 

Abstract This paper presents results from wind tunnel experiments to evaluate power gains from wake steering via yaw control. An experimental scaled wind farm with 12 turbines in an aligned rectangular array is used. Wake steering is performed by yawing turbines using a closed-loop algorithm termed the Log-of-Power Proportional Integral Extremum Seeking Control (LP-PIESC). Two configurations are considered. In the first configuration, the turbines in the first two upstream rows are controlled. In the second case, yaw control is applied to the turbines in the first upstream row and the third row. For both cases, uncontrolled turbines have no yaw misalignment. The results show that by independent parallel maximization of the power sum of a reduced number of turbines, it is possible to obtain a close approximation of the true maximum power. The data shows that the LP-PIESC algorithm can converge relatively fast compared to traditional ESC algorithms. 

Icing on the blades of wind turbines during winter seasons causes a reduction in power and revenue losses. The prediction of icing before it occurs has the potential to enable mitigating actions to reduce ice accumulation. This paper presents a framework for the prediction of icing on wind turbines based on Supervisory Control and Data Acquisition (SCADA) data without requiring the installation of any additional icing sensors on the turbines. A Temporal Convolutional Network is considered as the model to predict icing from the SCADA data time series. All aspects of the icing prediction framework are described, including the necessary data preprocessing, the labeling of SCADA data for icing conditions, the selection of informative icing features or variables in SCADA data, and the design of a Temporal Convolutional Network as the prediction model. Two performance metrics to evaluate the prediction outcome are presented. Using SCADA data from an actual wind turbine, the model achieves an average prediction accuracy of 77.6% for future times of up to 48 h. 

Each year a growing number of wind farms are being added to power grids to generate sustainable energy. The power curve of a wind turbine, which exhibits the relationship between generated power and wind speed, plays a major role in assessing the performance of a wind farm. Neural networks have been used for power curve estimation. However, they do not produce a confidence measure for their output, unless computationally prohibitive Bayesian methods are used. In this paper, a probabilistic neural network with Monte Carlo dropout is considered to quantify the model or epistemic uncertainty of the power curve estimation. This approach offers a minimal increase in computational complexity and thus evaluation time. Furthermore, by adding a probabilistic loss function, the noise or aleatoric uncertainty in the data is estimated. The developed network captures both model and noise uncertainty which are found to be useful tools in assessing performance. Also, the developed network is compared with the existing ones across a public domain dataset showing superior performance in terms of prediction accuracy. The results obtained indicate that the developed network provides the quantification of uncertainty while maintaining accurate power estimation. 

An efficient strategy for maximizing the power production of a power plant is to control in a coordinated way only turbines that are aerodynamically coupled through wake effects. The implementation of such control strategy requires the knowledge of which clusters of turbines are coupled through wake interaction. In a previous study, we identified turbine clusters in real-time by evaluating the correlation among the power production signals of the turbines in the farm. In this study we reproduce the more challenging scenario with large scale variation of the wind direction. Different time windows of data needed to compute the correlation coefficients are tested and characterized in term of accuracy and promptness of the identification. 

This paper presents regression and classification methods to estimate wind direction in a wind farm from operational data. Two neural network models are trained using supervised learning. The data are generated using high-fidelity large eddy simulations (LES) of a virtual wind farm with 16 turbines, which are representative of the data available in actual SCADA systems. The simulations include the high-fidelity flow physics and turbine dynamics. The LES data used for training and testing the neural network models are the rotor angular speeds of each turbine. Our neural network models use sixteen angular speeds as inputs to produce an estimate of the wind direction at each point in time. Training and testing of the neural network models are done for seven discrete wind directions, which span the most interesting cases due to symmetry of the wind farm layout. The results of this paper are indicative of the potential that existing neural network models have to obtain estimates of wind direction in real time.