Search

Wind prediction has consistently been in the spotlight as a crucial element in achieving efficient wind power generation and reducing operational costs. In recent years, with the rapid advancement of artificial intelligence (AI) technology, its application in the field of wind prediction has made significant strides. Focusing on the process of AI-based wind prediction modeling, this paper provides a comprehensive summary and discussion of key techniques and models in data preprocessing, feature extraction, relationship learning, and parameter optimization. Building upon this, three major challenges are identified in AI-based wind prediction: the uncertainty of wind data, the incompleteness of feature extraction, and the complexity of relationship learning. In response to these challenges, targeted suggestions are proposed for future research directions, aiming to promote the effective application of AI technology in the field of wind prediction and address the crucial issues therein. [ABSTRACT FROM AUTHOR]

Wind energy forecasting is a critical aspect for wind energy producers, given that the chaotic nature and the intermittence of meteorological wind cause difficulties for both the integration and the commercialization of wind-produced electricity. For most European producers, the quality of the forecast also affects their financial outcomes since it is necessary to include the impact of imbalance penalties due to the regularization in balancing markets. To help wind farm owners in the elaboration of offers for electricity markets, the EOLO predictor model can be used. This tool combines different sources of data, such as meteorological forecasts, electric market information, and historic production of the wind farm, to generate an estimation of the energy to be produced, which maximizes its financial performance by minimizing the imbalance penalties. This research study aimed to evaluate the performance of the EOLO predictor model when it is applied to the different Spanish electricity markets, focusing on the statistical analysis of its results. Results show how the wind energy forecast generated by EOLO anticipates real electricity generation with high accuracy and stability, providing a reduced forecast error when it is used to participate in successive sessions of the Spanish electricity market. The obtained error, in terms of RMAE, ranges from 8%, when it is applied to the Day-ahead market, to 6%, when it is applied to the last intraday market. In financial terms, the prediction achieves a financial performance near 99% once imbalance penalties have been discounted. [ABSTRACT FROM AUTHOR]

The design and placement of a new building in an urban environment is optimised so that pedestrians do not feel uncomfortable due to high wind speeds. Architects and city planners typically use Computational Fluid Dynamics (CFD) simulations to predict the effects of wind on new buildings. CFD is too time consuming to be used in generative design tools, which are necessary because they allow designers to visualise and iterate on their ideas, collaborate effectively and work more efficiently. These tools enhance the design process, facilitate communication and feedback, and save time by automating tasks. Finding a generative design model based on limited data is challenging due to issues such as insufficient representation, increased risk of overfitting, lack of statistical significance, potential bias and difficulties in validation and evaluation. In the design framework proposed in this paper, the computationally intensive CFD solver could be replaced by a fast machine learning surrogate predicting wind speed. The replacement is a U-Net convolutional neural network trained on high-precision Reynolds' averaged Navier–Stokes (RANS) simulations (CFD simulations). The proposed workflow is demonstrated by optimising the placement of a hypothetical new building on a city square in Gothenburg, Sweden. The surrogate used in the optimisation process calculates an optimal building design within seconds instead of hours saving 11 h of simulation time. The actual area with strong wind effects is reduced by half. [Display omitted] • Interactive design framework for new buildings in existing city structures. • 2D-U-Nets for limited RANS simulations to predict wind velocity and pressure. • Optimising building parameters to reduce high velocities in existing city structures. • Design optimisation completes in seconds instead of 11 hours, saving significant time. [ABSTRACT FROM AUTHOR]

• Prediction of dominating speed and direction of 1D time series wind data. • 1D single CNN (1DS) and multiple CNN (1DM) as classification for prediction. • Validation with real world meteorological data of Stuttgart and Netherlands. • Classification methods work on original wind data without smoothening and filtering. • User defined future time frame of prediction. This paper proposes two one-dimensional (1D) convolutional neural networks (CNNs) for predicting dominant wind speed and direction for the temporal wind dataset. The proposed 1D Single CNN (1DS) takes as input the consecutive temporal values in terms of the wind speed and direction and predicts in future dominating speed and direction, separately, after the last value in the input. The developed 1D Multiple CNN (1DM) combines several 1DS but with different views of the same input, therefore, learning more information compared to the 1DS. The proposed algorithms have been trained and tested using the historical wind datasets of Stuttgart (Germany) and Netherlands for different months. Total accuracies reached up to 95.2%, 95.1% for predicting the dominant wind speed and direction, respectively, using the 1DS and up to 98.8%, 99.7% for predicting the dominant wind speed and direction, respectively, using the 1DM. Unlike other methods which use regression techniques with manually designed features for predicting speed and direction, the proposed methods have used classification techniques with the 1DS and 1DM learning their features automatically on original wind dataset. Further, predicted dominant speed and direction in this work would be helpful in wind turbine installation whose power output depends on above parameters. [ABSTRACT FROM AUTHOR]

This paper presents a system for identification of wind features, such as gusts and wind shear. These are of particular interest in the context of energy-efficient navigation of Small Unmanned Aerial Systems (UAS). The proposed system generates real-time wind vector estimates and a novel algorithm to generate wind field predictions. Estimations are based on the integration of an off-the-shelf navigation system and airspeed readings in a so-called direct approach. Wind predictions use atmospheric models to characterize the wind field with different statistical analyses. During the prediction stage, the system is able to incorporate, in a big-data approach, wind measurements from previous flights in order to enhance the approximations. Wind estimates are classified and fitted into a Weibull probability density function. A Genetic Algorithm (GA) is utilized to determine the shaping and scale parameters of the distribution, which are employed to determine the most probable wind speed at a certain position. The system uses this information to characterize a wind shear or a discrete gust and also utilizes a Gaussian Process regression to characterize continuous gusts. The knowledge of the wind features is crucial for computing energy-efficient trajectories with low cost and payload. Therefore, the system provides a solution that does not require any additional sensors. The system architecture presents a modular decentralized approach, in which the main parts of the system are separated in modules and the exchange of information is managed by a communication handler to enhance upgradeability and maintainability. Validation is done providing preliminary results of both simulations and Software-In-The-Loop testing. Telemetry data collected from real flights, performed in the Seville Metropolitan Area in Andalusia (Spain), was used for testing. Results show that wind estimation and predictions can be calculated at 1Hz and a wind map can be updated at 0.4Hz. Predictions show a convergence time with a 95% confidence interval of approximately 30 s. [ABSTRACT FROM AUTHOR]

In this study, we investigate the accuracy of wind-speed prediction at a designated target site using wind-speed data from reference stations that employ an ANN (artificial neural network). The reference and target sites fall into three geographical categories: plains, coast, and mountains of South Korea. Accurate wind-speed predictions are calculated by means of a correlation coefficient between the actual and simulated wind-speed data obtained by ANN. We investigate the effect of the geological characteristics of each category and the distance between reference and target sites on the accuracy of wind-speed prediction using ANN. [ABSTRACT FROM AUTHOR]

Abstract: Wind energy has an important part to play as renewable energy resource in a sustainable world. For a reliable integration of wind energy high-dimensional wind time-series have to be analyzed. Fault analysis and prediction are an important aspect in this context. The objective of this work is to show how methods from neural computation can serve as forecasting and monitoring techniques, contributing to a successful integration of wind into sustainable and smart energy grids. We will employ support vector regression as prediction method for wind energy time-series. Furthermore, we will use dimension reduction techniques like self-organizing maps for monitoring of high-dimensional wind time-series. The methods are briefly introduced, related work is presented, and experimental case studies are exemplarily described. The experimental parts are based on real wind energy time-series data from the National Renewable Energy Laboratory (NREL) western wind resource data set. [Copyright &y& Elsevier]

A class of nonlinear quaternion-valued adaptive filtering algorithms is proposed based on locally analytic nonlinear activation functions. To circumvent the stringent standard analyticity conditions which are prohibitive to the development of nonlinear adaptive quaternion-valued estimation models, we use the fact that stochastic gradient learning algorithms require only local analyticity at the operating point in the estimation space. It is shown that the quaternion-valued exponential function is locally analytic, and, since local analyticity extends to polynomials, products, and ratios, we show that a class of transcendental nonlinear functions can serve as activation functions in nonlinear and neural adaptive models. This provides a unifying framework for the derivation of gradient-based learning algorithms in the quaternion domain, and the derived algorithms are shown to have the same generic form as their real- and complex-valued counterparts. To make such models second-order optimal for the generality of quaternion signals (both circular and noncircular), we use recent developments in augmented quaternion statistics to introduce widely linear versions of the proposed nonlinear adaptive quaternion valued filters. This allows full exploitation of second-order information in the data, contained both in the covariance and pseudocovariances to cater rigorously for second-order noncircularity (improperness), and the corresponding power mismatch in the signal components. Simulations over a range of circular and noncircular synthetic processes and a real world 3-D noncircular wind signal support the approach. [ABSTRACT FROM AUTHOR]

A novel complex echo state network (ESN), utilizing full second-order statistical information in the complex domain, is introduced. This is achieved through the use of the so-called augmented complex statistics, thus making complex ESNs suitable for processing the generality of complex-valued signals, both second-order circular (proper) and noncircular (improper). Next, in order to deal with nonstationary processes with large nonlinear dynamics, a nonlinear readout layer is introduced and is further equipped with an adaptive amplitude of the nonlinearity. This combination of augmented complex statistics and enhanced adaptivity within ESNs also facilitates the processing of bivariate signals with strong component correlations. Simulations in the prediction setting on both circular and noncircular synthetic benchmark processes and real-world noncircular and nonstationary wind signals support the analysis. [ABSTRACT FROM AUTHOR]

Wind forecasts over a varying period of time are needed for a variety of applications in the coastal and ocean region, like planning of construction and operation-related works as well as prediction of power output from wind turbines located in coastal areas. Such forecasting is currently done by adopting complex atmospheric models or by using statistical time-series analysis. Because occurrence of wind in nature is extremely uncertain no single technique can be entirely satisfactory. This leaves scope for alternative approaches. The present work employs the technique of neural networks in order to forecast daily, weekly as well as monthly wind speeds at two coastal locations in India. Both feed forward as well as recurrent networks are used. They are trained based on past data in an auto-regressive manner using back-propagation and cascade correlation algorithms. A generally satisfactory forecasting as reflected in its higher correlation and lower deviations with actual observations is noted. The neural network forecasting is also found to be more accurate than traditional statistical time-series analysis. [Copyright &y& Elsevier]

Modeling forest fire spread is a very complex problem, and the existing models usually need some input parameters which are hard to get. How to predict the time series of forest fire spread rate based on passed series may be a key problem to break through the current technical bottleneck. In the process of forest fire spreading, spread rate and wind speed would affect each other. In this paper, three kinds of network models based on Long Short-Term Memory (LSTM) are designed to predict fire spread rate, exploring the interaction between fire and wind. In order to train these LSTM-based models and validate their effectiveness of prediction, several outdoor combustion experiments are designed and carried out. Process data sets of forest fire spreading are collected with an infrared camera mounted on a UAV, and wind data sets are recorded using a anemometer simultaneously. According to the close relationship between wind and fire, three progressive LSTM based models are constructed, which are called CSG-LSTM, MDG-LSTM and FNU-LSTM, respectively. A Cross-Entropy Loss equation is employed to measure the model training quality, and then prediction accuracy is computed and analyzed by comparing with the true fire spread rate and wind speed. According to the performance of training and prediction stage, FNU-LSTM is determined as the best model for the general case. The advantage of FNU-LSTM is further demonstrated by doing comparison experiments with the normal LSTM and other LSTM based models which predict both fire spread rate and wind speed separately. The experiment has also demonstrated the ability of the model to the real fire prediction on the basis of two historical wildland fires. [ABSTRACT FROM AUTHOR]

• Models uniting wavelet transform, deep learning and ensemble learning were proposed. • Tree-based models facilitate the prediction of DBN. • Feature extension of DBN improve the prediction accuracy of tree-based models. • After extraction, bagging and boosting algorithm vary in accuracy and stability. The utilization of wind power is influenced by the fluctuation of the wind, to further strengthen the prediction accuracy of wind speed, two novel hybrid models uniting signal processing, deep learning and ensemble learning were proposed. Firstly, the wind speed series was disaggregated by the wavelet transform (WT). Then to enhance the forecasting precision of the subseries, the deep belief network (DBN) was applied to extract the high dimensional features. Besides, to overcome the limitation of the conventional DBN, the forecasts for each subseries processed by DBN were executed by the light gradient boosting machine (LGBM) and the random forest (RF). Some experiments have been accomplished, where the promotion of high dimensional feature extraction through DBN was explored. Meanwhile, the development of forecasting accuracy by applying tree-based models was confirmed, and the differences between these two hybrid models were discussed. It is shown that: (1) In comparison with the persistence method, the Elman neural network (ENN), DBN, LGBM, and RF, the hybrid models show a great boost in prediction accuracy. (2) The high dimensional feature extraction through DBN is in favor of improving the predicting accuracy of tree-based models, and tree-based models would facilitate the prediction. (3) Between the proposed models, the hybrid model integrated with RF slightly outperforms the other with LGBM in prediction accuracy, but the one with LGBM gets more stable predictions. [ABSTRACT FROM AUTHOR]