scholarly journals Model-free estimation of available power using deep learning

2021 ◽  
Vol 6 (1) ◽  
pp. 111-129
Author(s):  
Tuhfe Göçmen ◽  
Albert Meseguer Urbán ◽  
Jaime Liew ◽  
Alan Wai Hou Lio
Keyword(s):  
Wind Speed ◽  
Wind Turbine ◽  
Transfer Learning ◽  
Short Term Memory ◽  
High Accuracy ◽  
Normal Operation ◽  
Zero Bias ◽  
Model Free ◽  
Single Input ◽  
Model Free Approach

Abstract. In order to assess the level of power reserves during down-regulation, the available power of a wind turbine needs to be estimated. The current practice in available power estimation is heavily dependent on the pre-defined performance parameters of the turbine and the curtailment strategy followed. This paper proposes a single-input model-free approach dynamic estimation of the available power using recurrent neural networks. Accordingly, it combines wind turbine control considerations and modern forecasting methodologies for a model-free, single-input estimation of available power. It enables a robust real-time implementation of dynamic delta control, as well as higher-accuracy provision of the reserves to the system operators. The model-free approach requires only 1 Hz wind speed measurements as input and estimates 1 Hz available power as output. The neural network is trained, tested and validated using the DTU 10 MW reference wind turbine HAWC2 model under realistic atmospheric conditions. The unsteady patterns in the turbulent flow are represented via long short-term memory (LSTM) neurons which are trained during a period of normal operation. The adaptability of the network to changing inflow conditions is ensured via transfer learning, where the last LSTM layer is updated using new measurements. It is seen that the sensitivity of the networks to changing wind speed is much higher than that of turbulence, and the updates are to be implemented solely based on the altering inflow velocity. The validation of the trained LSTM networks on time series with 7, 9 and 11 m s−1 mean wind speeds demonstrates high accuracy (less than 1 % bias) and capability of transfer-learning online. Including highly turbulent inflow cases, the networks have shown to comply with the most recent grid codes, which require the quality of the available power estimations to be evaluated with high accuracy (less than 3.3 % standard deviation of the error around zero bias) at 1 min intervals.

scholarly journals Model-free Estimation of Available Power using Deep Learning

2019 ◽  
Author(s):  
Tuhfe Göçmen ◽  
Jaime Liew ◽  
Albert Meseguer Urban ◽  
Alan Wai Hou Lio
Keyword(s):  
Wind Speed ◽  
Wind Turbine ◽  
Transfer Learning ◽  
Short Term Memory ◽  
Power Estimation ◽  
High Accuracy ◽  
Normal Operation ◽  
Atmospheric Conditions ◽  
Model Free ◽  
Model Free Approach

Abstract. In order to assess the level of power reserves during down-regulation, the available power of a wind turbine needs to be estimated. The current practice in available power estimation is heavily dependent on the pre-defined performance parameters of the turbine and the curtailment strategy followed. This paper proposes a model-free approach for a single input, dynamic estimation of the available power using recurrent neural networks. The unsteady patterns in the turbulent flow are represented via Long Short-Term Memory (LSTM) neurons which are trained during a period of normal operation. The model-free approach requires only 1-Hz wind speed measurements as the input and generates 1-Hz available power estimation as the output. The neural network is trained, tested and validated using the DTU 10 MW reference wind turbine HAWC2 model under realistic atmospheric conditions. The adaptability of the network to changing inflow conditions is ensured via transfer learning, where the last LSTM layer is updated using new measurements. It is seen that the sensitivity of the networks to changing wind speed is much higher than that of turbulence, and the updates are to be implemented solely based on the altering inflow velocity. The validation of the trained LSTM networks on time series with 7, 9 and 11 m/s mean wind speeds demonstrates high accuracy (less than 1 % bias) and capability of transfer-learning. Including highly turbulent inflow cases, the networks have shown to easily comply with the most recent grid codes, which require the quality of the available power estimations to be evaluated with high accuracy (less than 3.3 % standard deviation of the error) at 1-min intervals.

scholarly journals Proactive Critical Energy Infrastructure Protection via Deep Feature Learning

Energies ◽  
2020 ◽  
Vol 13 (10) ◽  
pp. 2622
Author(s):  
Konstantina Fotiadou ◽  
Terpsichori Helen Velivassaki ◽  
Artemis Voulkidis ◽  
Dimitrios Skias ◽  
Corrado De Santis ◽  
...  
Keyword(s):  
Fault Detection ◽  
Wind Turbine ◽  
Short Term Memory ◽  
Feature Learning ◽  
Slight Modification ◽  
Critical Energy ◽  
Normal Operation ◽  
High Detection Rate ◽  
Energy Infrastructure ◽  
Critical Energy Infrastructure

Autonomous fault detection plays a major role in the Critical Energy Infrastructure (CEI) domain, since sensor faults cause irreparable damage and lead to incorrect results on the condition monitoring of Cyber-Physical (CP) systems. This paper focuses on the challenging application of wind turbine (WT) monitoring. Specifically, we propose the two challenging architectures based on learning deep features, namely—Long Short Term Memory-Stacked Autoencoders (LSTM-SAE), and Convolutional Neural Network (CNN-SAE), for semi-supervised fault detection in wind CPs. The internal learnt features will facilitate the classification task by assigning each upcoming measurement into its corresponding faulty/normal operation status. To illustrate the quality of our schemes, their performance is evaluated against real-world’s wind turbine data. From the experimental section we are able to validate that both LSTM-SAE and CNN-SAE schemes provide high classification scores, indicating the high detection rate of the fault level of the wind turbines. Additionally, slight modification on our architectures are able to be applied on different fault/anomaly detection categories on variant Cyber-Physical systems.

scholarly journals Numerical Simulation of the Aeroelastic Response of Wind Turbines in Typhoons Based on the Mesoscale WRF Model

2019 ◽  
Vol 12 (1) ◽  
pp. 34
Author(s):  
Long Wang ◽  
Cheng Chen ◽  
Tongguang Wang ◽  
Weibin Wang
Keyword(s):  
Wind Speed ◽  
Wind Turbine ◽  
Wind Direction ◽  
Flow Direction ◽  
Wrf Model ◽  
Normal Operation ◽  
Average Wind Speed ◽  
Aeroelastic Response ◽  
Response Characteristics ◽  
Wind Speeds

A new simulation method for the aeroelastic response of wind turbines under typhoons is proposed. The mesoscale Weather Research and Forecasting (WRF) model was used to simulate a typhoon’s average wind speed field. The measured power spectrum and inverse Fourier transform method were coupled to simulate the pulsating wind speed field. Based on the modal method and beam theory, the wind turbine model was constructed, and the GH-BLADED commercial software package was used to calculate the aerodynamic load and aeroelastic response. The proposed method was applied to assess aeroelastic response characteristics of a commercial 6 MW offshore wind turbine under different wind speeds and direction variation patterns for the case study of typhoon Hagupit (2008), with a maximal wind speed of 230 km/h. The simulation results show that the typhoon’s average wind speed field and turbulence characteristics simulated by the proposed method are in good agreement with the measured values: Their difference in the main flow direction is only 1.7%. The scope of the wind turbine blade in the typhoon is significantly larger than under normal wind, while that under normal operation is higher than that under shutdown, even at low wind speeds. In addition, an abrupt change in wind direction has a significant impact on wind turbine response characteristics. Under normal operation, a sharp variation of the wind direction by 90 degrees in 6 s increases the wind turbine (WT) vibration scope by 27.9% in comparison with the case of permanent wind direction. In particular, the maximum deflection of the wind tower tip in the incoming flow direction reaches 28.4 m, which significantly exceeds the design standard safety threshold.

Finite Element Analysis of Concrete Filled Double Skin Steel Tubes for Wind Turbine Tower

2014 ◽  
Vol 680 ◽  
pp. 551-556
Author(s):  
Wei Kong ◽  
Hong Liang Wang ◽  
Ying Cai
Keyword(s):  
Finite Element ◽  
Wind Speed ◽  
Wind Turbine ◽  
Normal Operation ◽  
Mode Shapes ◽  
Strength Analysis ◽  
Element Analysis ◽  
Finite Element Software ◽  
Wind Turbine Tower ◽  
Double Skin

In order to save the steel consumption,ensure the better economy of wind turbine tower,this paper designeda new concrete filled double skin steel tube for wind turbine tower,based on the parameters of 1. 5 MW wind turbine tower.A three-dimensional finite element model of wind turbine tower was built by using the finite element software ANSYS,then the static strength analysis and modal analysis were carried out,in which the stress and displacement at the top of the tower were calculated under three kinds of working conditions: normal operation with rated wind speed,normal operation with cutout wind speed and shutdown under extreme wind conditions,the natural frequency and mode shapes of the tower were obtained as well. The results show that the tower does not resonate with blades,and its structure can meet the strength and stiffness requirements of engineering.

scholarly journals Enhancement of power system performance with wind farm disturbances

2019 ◽  
Vol 8 (2) ◽  
pp. 159
Author(s):  
Mahmoud M. El-Sharkawy ◽  
Mahmoud A. Attia ◽  
Almoataz Y. Abdelaziz
Keyword(s):  
Wind Speed ◽  
Power System ◽  
Wind Turbine ◽  
Pitch Angle ◽  
Wind Farm ◽  
Search Algorithm ◽  
Harmony Search ◽  
Normal Operation ◽  
Ground Fault ◽  
Variable Wind

This paper discussed how wind farm disturbance, especially wind speed variability, can affect the performance of the power system. Also, it discussed how blades angels of wind turbine can be controlled to increase the energy efficacy of the power system. It showed that the optimized pitch angel controller using harmony search algorithm could enhance blades angels’ adjustments performance. First part, the paper explained the advantages of doubly fed induction generator in wind turbine system. Paper also enumerated the most probable ways to tame the wind speed variability challenge focusing on the pitch angle controller technique. After that, paper compared the system parameter`s result before optimization with values after optimized pitch angle controller gain (Kp). This comparison would be held in three cases, case of variable wind speed with normal operation condition, constant wind speed with line to ground fault condition and variable wind speed with line to ground fault condition. Finally, It demonstrated the MATLAB/SIMULINK model used illustrating results appeared and conclusion.

Long-Term Extreme Response Analysis for a Fixed Offshore Wind Turbine Considering Blade-Pitch-Actuator Fault and Normal Transient Events

2014 ◽  
Author(s):  
Qinyuan Li ◽  
Zhen Gao ◽  
Torgeir Moan
Keyword(s):  
Wind Speed ◽  
Wind Turbine ◽  
Wind Direction ◽  
Normal Operation ◽  
Offshore Wind Turbine ◽  
Bending Moments ◽  
Start Up ◽  
Extreme Response ◽  
Transient Events

In this paper, the 50-year long-term 1-hour extreme responses of a fixed jacket-type offshore wind turbine with consideration of one-blade-pitch-actuator-stuck fault and the effect of normal transient events such as normal shut-down and start-up process is studied. The long-term extreme results are found based on each short-term extreme response distributions at different environmental conditions. Structure responses such as tower and jacket bottom shear and bending moments as well as blade root bending moments will be focused in this paper. To study the long-term effect of the fault and transient events, the service life of a wind turbine is divided into normal part, faulted part, and transient part. Normal part includes both normal operation and parking of the wind turbine at different wind speed range without any faults. Faulted part includes the parked and emergency shut-down condition of the wind turbine under the fault assuming that the faults are detected soon after they occur but require a longer time before fully repaired. Transient part includes the start-up and shut-down process during the normal operation when wind speed is beyond operation range. The contribution of each part to the long-term extreme response distribution is calculated by weighting factors based on the probability of occurrence of each part. From the results, it is found that in general, the blade-pitch-actuator-stuck fault and the normal transient events generally increase the extreme responses of the wind turbine. The jacket wind turbine is more affected compared to its land based counterpart. In this study since the wind direction is aligned with wind turbine, it is found that the fault primarily increases the tower bottom shear force perpendicular to the wind direction and the bending moments with the axis parallel to the wind as well as the torsional moment, while normal transient events, especially the start-up process at cut-out speed, causes a much greater increase compared to the fault. It contribute mostly to the shear forces parallel and bending moment with axis perpendicular to the wind direction. The azimuth of the blades is found to be very important for blade responses during start-up process especially at higher wind speed.

Design and Performance Analysis of Distributed Equal Angle Spiral Vertical Axis Wind Turbine

2020 ◽  
Vol 14 (1) ◽  
pp. 120-132
Author(s):  
Li Zheng ◽  
Zhang Wenda ◽  
Han Ruihua ◽  
Tian Yongsheng
Keyword(s):  
Wind Speed ◽  
Wind Energy ◽  
Wind Turbine ◽  
Wind Turbines ◽  
Vertical Axis ◽  
Energy Utilization ◽  
Normal Operation ◽  
Utilization Rate ◽  
Speed Ratio ◽  
Vertical Axis Wind Turbine

Background: The wind turbine is divided into a horizontal axis and a vertical axis depending on the relative positions of the rotating shaft and the ground. The advantage of the choke wind turbine is that the starting torque is large and the starting performance is good. The disadvantage is that the rotation resistance is large, the rotation speed is low, the asymmetric flow occurs when the wind wheel rotates, the lateral thrust is generated, and the wind energy utilization rate is lowered. How to improve the wind energy utilization rate of the resistance wind turbine is an important issue to be solved by the wind power technology. Objective: The nautilus isometric spiral wind turbines studied in this paper have been introduced and analyzed in detail, preparing for the further flow analysis and layout of wind turbines, improving the wind energy utilization rate of wind turbines, introducing patents of other structures and output characteristics of its generator set. Methods: Combined with the flow field analysis of ANSYS CFX software, the numerical simulation of the new wind turbine was carried out, and the aerodynamic performance of the new vertical axis wind turbine was analyzed. The mathematical model and control model of the generator were established by the maximum power control method, and the accuracy of the simulation results was verified by the measured data. Results: The basic parameters of the new wind turbine tip speed ratio, torque coefficient and wind energy utilization coefficient are analyzed. Changes in wind speed, pressure and eddy viscosity were investigated. Three-dimensional distribution results of wake parameters such as wind speed and pressure are obtained. By simulating the natural wind speed, the speed and output current of the generator during normal operation are obtained. Conclusion: By analyzing the wind performance and power generation characteristics of the new wind turbine, the feasibility of the new wind turbine is determined, which provides reference and reference for the optimal design and development of the wind turbine structure.

Deep neural network-based wind speed forecasting and fatigue analysis of a large composite wind turbine blade

2018 ◽  
Vol 233 (8) ◽  
pp. 2794-2812 ◽  
Author(s):  
Pravin A Kulkarni ◽  
Ashwinkumar S Dhoble ◽  
Pramod M Padole
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Wind Speed ◽  
Wind Turbine ◽  
Short Term Memory ◽  
Fatigue Analysis ◽  
Short Term ◽  
Term Memory ◽  
Long Short Term Memory ◽  
Nonlinear Autoregressive

The purpose of this paper is to analyze the modern deep neural networks such as nonlinear autoregressive network with external inputs and a recurrent neural network called long short-term memory for wind speed forecast for long-term and use the prediction for fatigue analysis of a large 5 MW wind turbine blade made of composite materials. The use of machine learning algorithms of advanced neural network applied for engineering problems is increasing recently. The present paper therefore brings as important connection between these latest machine learning methods and engineering analysis of complex wind turbine blades which are also the focus of researchers in renewable system design and analysis. First, a nonlinear autoregressive network with external inputs neural network model using Levenberg–Marquardt back propagation feed forward algorithm is developed with 5 years of environment parameters as input. Similarly, a long short-term memory based model is developed and compared. The chosen long short-term memory model is used for developing two-year wind speed forecast. This wind pattern is used to create time varying loads on blade sections and cross-verified with National Renewable Energy Laboratory tools. A high-fidelity CAD model of the NREL 5 MW blade is developed and the fatigue analysis of the blade is carried out using the stress life approach with load ratio based on cohesive zone modeling. The blade is found to have available life of about 23.6 years. Thus, an integrated methodology is developed involving high-fidelity modeling of the composite material blade, wind speed forecasting using multiple environmental parameters using latest deep learning methods for machine learning, dynamic wind load calculation, and fatigue analysis for National Renewable Energy Laboratory blade.

Dynamic Optimization of Drivetrain Gear Ratio to Maximize Wind Turbine Power Generation—Part 1: System Model and Control Framework

2012 ◽  
Vol 135 (1) ◽  
Author(s):  
John F. Hall ◽  
Dongmei Chen
Keyword(s):  
Wind Speed ◽  
Wind Turbine ◽  
Dynamic Optimization ◽  
Low Cost ◽  
Gear Ratio ◽  
Normal Operation ◽  
Power Coefficient ◽  
Speed Ratio ◽  
Electric Generator ◽  
Disk Brake

Wind is considered to be one of the most promising resources in the renewable energy portfolio. Still, to make wind energy conversion more economically viable, it is necessary to extract greater power from the wind while minimizing the cost associated with the technology. This is particularly important for small wind turbines, which have the highest cost per kilowatt of energy produced. One solution would be a variable ratio gearbox (VRG) that can be integrated into the simple and low-cost fixed-speed induction generator. Through discrete variable rotor speed operation, the VRG-enabled system affects the wind speed ratio, the power coefficient, and ultimately the power produced. To maximize electrical production, mechanical braking is applied during the normal operation of the wind turbine. A strategy is used to select gear ratios (GRs) that produce torque slightly above the maximum amount the generator can accept while simultaneously applying the mechanical brake, so that full load production may be realized over greater ranges of the wind speed. To characterize the performance of the system, a 100 kW, fixed speed, stall-regulated wind turbine, has been developed for this study. The VRG-enabled wind turbine control system is presented in two papers. Part 1 focuses on the turbine simulation model, which includes the rotor, VRG-enabled drivetrain, disk brake, and electric generator. A technique for estimating the performance of a disk brake, in the wind turbine context, is also presented. Part 2 of the research will present a dynamic optimization algorithm that is used to establish the control protocol for competing performance objectives.

Sign in / Sign up

Export Citation Format

Share Document