Enhancing Short-Term Wind Power Forecasting through Multiresolution Analysis and Echo State Networks

On differences and similarities of applied wind comfort criteria

This study analyzed the wind speed data of the met mast in the first commercial-scale offshore wind farm of Taiwan from May 2017 to April 2018. The mean wind speed and standard deviation, wind rose, histogram, wind speed profile, and diurnal variation of wind speed with associated changes in wind direction revealed some noteworthy findings. First, the standard deviation of the corresponding mean wind speed is somewhat high. Second, the Hellmann exponent is as low as 0.05. Third, afternoons in winter and nights and early mornings in summer have the highest and lowest wind speed in a year, respectively. Regarding the histogram, the distribution probability of wind is bimodal, which can be depicted as a mixture of two gamma distributions. In addition, the corresponding change between the hourly mean wind speed and wind direction revealed that the land–sea breeze plays a significant role in wind speed distribution, wind profile, and wind energy production. The low Hellmann exponent is discussed in detail. To further clarify the effect of the land–sea breeze for facilitating future wind energy development in Taiwan, we propose some recommendations.

The fluctuation and uncertainty of wind power generation bring severe challenges to secure and economic operation of power systems. Because wind power forecasting error is unavoidable, probabilistic forecasting becomes critical to accurately quantifying the uncertainty involved in traditional point forecasts of wind power and to providing meaningful information to conduct risk management in power system operation. This paper proposes a novel direct quantile regression approach to efficiently generate nonparametric probabilistic forecasting of wind power generation combining extreme learning machine and quantile regression. Quantiles with different proportions can be directly produced via an innovatively formulated linear programming optimization model, without dependency on point forecasts. Multistep probabilistic forecasting of 10-min wind power is newly carried out based on real wind farm data from Bornholm Island in Denmark. The superiority of the proposed approach is verified through comparisons with other well-established benchmarks. The proposed approach forms a new artificial neural network-based nonparametric forecasting framework for wind power with high efficiency, reliability, and flexibility, which can be beneficial to various decision-making activities in power systems.

This paper aims to produce a low-complexity predictor for the hourly mean wind speed and direction from 1 to 6 h ahead at multiple sites distributed around the UK. The wind speed and direction are modelled via the magnitude and phase of a complex-valued time series. A multichannel adaptive filter is set to predict this signal on the basis of its past values and the spatio-temporal correlation between wind signals measured at numerous geographical locations. The filter coefficients are determined by minimizing the mean square prediction error. To account for the time-varying nature of the wind data and the underlying system, we propose a cyclo-stationary Wiener solution, which is shown to produce an accurate predictor. An iterative solution, which provides lower computational complexity, increased robustness towards ill-conditioning of the data covariance matrices and the ability to track time-variations in the underlying system, is also presented. The approaches are tested on wind speed and direction data measured at various sites across the UK. Results show that the proposed techniques are able to predict wind speed as accurately as state-of-the-art wind speed forecasting benchmarks while simultaneously providing valuable directional information.

Wind speed is the most critical factor that determines wind power potential and generation. In this paper, the wind speed data of multiple years from various observation sites in North Dakota, U.S. was analyzed to assess the wind power potential. The study first applied probability density functions (PDFs) to characterize the wind speed data and fit the distributions at various heights for each observation site. The fitted distributions were then used to estimate the wind power potential based on the theoretical cubic power relationship between energy potential and wind speed. Due to the complexity of functions, the numerical integration approach was employed. The following major findings were obtained from this empirical study: (1) Weibull distribution is not always the best function to fit wind speed data, while gamma and lognormal distributions produce better fitting in many occasions; (2) For different height levels at one observation site, the best performing distributions may be different; (3) The estimation accuracies of wind energy potential based on the fitted wind speed distributions range from -4% to 3.8%; (4) The rank of energy potential estimation accuracies is not always consistent with that of goodness-of-fit for wind speed distributions. In addition, a simplified approach that only relies on the hourly mean wind speed to estimate wind power potential is evaluated. Based on the theoretical cubic relationship for wind power estimation, it was found that the simplified approach may provide significantly lower estimates of wind power potential by 42-54%. As such, this approach will become more practical if this amount of difference is to be compensated. Key words: Wind speed; Distribution; Goodness-of-fit; Wind power potential; North Dakota

Neste trabalho o ciclo diário do vento à superfície é investigado com ênfase na variação sazonal e nas circulações locais. Os dados da estação automática instalada a 25-30 km de Petrolina, e da estação convencional localizada na cidade são utilizados. Os resultados mostram que as velocidades médias horárias aumentam (diminuem) no decorrer do trimestre abril-maio-junho (outubro-novembro-dezembro), devido à intensificação (enfraquecimento) da alta subtropical do Atlântico Sul. No trimestre julho-agosto-setembro os valores são máximos e muito próximos, com direção no intervalo 120-130 graus. O ciclo diário tem um máximo de intensidade às 8-9 HL (hora local) e dois mínimos, um às 5-6 HL e outro às 17-18 HL, em todos os meses. Os hodógrafos do vento local nas duas estações mostram circulações fracas de janeiro a maio, uma característica que pode estar relacionada com o maior teor de umidade do solo e, consequentemente, com valores mais baixos do gradiente do fluxo de calor sensível. No restante do ano, período de estiagem, as circulações locais são comparativamente fortes. Os hodógrafos apresentam forma irregular que indica um regime preferencial na direção sudeste-noroeste. As circulações locais são mais intensas às 9 e 18 HL, horários de máxima e mínima velocidade no ciclo diário, respectivamente.

Reliability analysis of a microwave tower for fluctuating mean wind with directional effect

Accurate wind power forecasting (WPF) is crucial to enhance availability and reap the benefits of integration into power grids. The time lag of wind power generation lags the time of wind speed changes, especially in ultra-short-term forecasting. The prediction model is sensitive to outliers and sudden changes in input historical meteorological data, which may significantly affect the robustness of the WPF model. To address this issue, this paper proposes a novel hybrid machine learning model for highly accurate forecasting of wind power generation in ultra-short-term forecasting. The raw wind power data were filtered and classified with the local outlier factor (LOF) and the voting tree (VT) model to obtain a subset of inputs with the best relevance. The time-varying properties of the fluctuating sub-signals of the wind power sequences were analyzed with the optimized variational mode decomposition (OVMD) algorithm. The Northern Goshawk optimization (NGO) algorithm was improved by incorporating a logical chaotic initialization strategy and chaotic adaptive inertia weights. The improved NGO algorithm was used to optimize the least squares support vector regression (LSSVR) prediction model to improve the computational speed and prediction results. The proposed model was compared with traditional machine learning models, deep learning models, and other hybrid models. The experimental results show that the proposed model has an average R2 of 0.9998. The average MSE, average MAE, and average MAPE are as low as 0.0244, 0.1073, and 0.3587, which displayed the best results in ultra-short-term WPF.

Very short-term forecasting of wind power generation using hybrid deep learning model

This paper investigates short term forecast of wind power generation using Support Vector Machine (SVM) methods. We propose a similar pattern matching technique for data pre-processing. The proposed technique is able to select the most appropriate segments from the available historical data for the forecasting. These segments are used for the training of parameters in the SVM model, which will then be able to provide the desired forecast. We study the reliability of the technique in wind speed forecasting and power generation forecasting. The proposed technique is evaluated by using the public data sets obtained from the National Renewable Energy Laboratory.

An approach of a mean hourly wind speed forecasting in wind farm is proposed in this paper. It applies support vector regression as well as forecasting error estimation. Firstly, support vector regression is applied to the mean hourly wind speed forecasting. Secondly, a support vector classifier is trained to estimate the forecasting error. Finally, the forecasting results can tailor themselves to the estimated forecasting error, and thus improve the forecasting precision. To test the approach, three-year data from a wind farm is given as a support vector regression process, and a support vector classifier is trained in addition to estimate the forecasting error. Experimental results show that the proposed approach can achieve higher quality of mean hourly wind speed forecasting; also it has lower mean square error compared with the traditional support vector regression forecasting.

Accurate forecasting of wind farm power generation is essential for successful operation and management of wind farms and to minimize risks associated with their integration into energy systems. However, due to the inherent wind intermittency, wind power forecasts are highly prone to error and often far from being perfect. The purpose of this paper is to develop statistical methods for quantifying uncertainties associated with wind power generation forecasts. Prediction intervals (PIs) with a prescribed confidence level are constructed using the delta and bootstrap methods for neural network forecasts. The moving block bootstrap method is applied to preserve the correlation structure in wind power observations. The effectiveness and efficiency of these two methods for uncertainty quantification is examined using two month datasets taken from a wind farm in Australia. It is demonstrated that while all constructed PIs are theoretically valid, bootstrap PIs are more informative than delta PIs, and are therefore more useful for decision-making.

Wind energy is an emission free low-carbon energy source that has employed great advances in technology in recent years. The unpredictable and random characteristics of wind, including variables like wind speed, direction, barometric pressure and temperature present distinct challenges for precise wind power prediction. Accurate forecasts of wind power generation are crucial for the effective management of energy grids, ensuring stability, and the seamless integration of wind energy into existing power infrastructures. This study introduces a novel, Transformer-based, Dynamic context-aware power forecasting model tailored to enhance the accuracy of wind power forecasting. This model adapts the Transformer architecture, incorporating innovative modifications to tackle the complications of wind power forecasting. It improves the extraction of contextual features while simplifying the model’s structure. The results demonstrate the Normalized values of Root mean square error (NRMSE) 1.603, 2.219, 2.76 and 2.69 for Dynamic context-aware, Transformer, LSTM and GRU model respectively. Dynamic context-aware model outperforms other models achieving the lowest error rates in accuracy for prediction to provide comprehensive knowledge to decision-makers interested in wind turbines and energy optimization.

A substantial part of South Africa is subject to more than one strong wind source. The effect of that on extreme winds is that higher quantiles are usually estimated with a mixed strong wind climate estimation method, compared to the traditional Gumbel approach based on a single population. The differences in the estimated quantiles between the two methods depend on the values of the Gumbel distribution parameters for the different strong wind mechanisms involved. Cluster analysis of the distribution parameters provides a characterization of the effect of the relative differences in their values, and therefore the dominance of the different strong wind mechanisms. For gusts, cold fronts tend to dominate over the coastal and high-lying areas, while other mechanisms, especially thunderstorms, are dominant over the lower-lying areas in the interior. For the hourly mean wind speeds cold fronts are dominant in the south-west, south and east of the country. On the West Coast the ridging of the Atlantic Ocean high-pressure system dominate in the south, while the presence of a deep trough or coastal low pressure system is the main strong wind mechanism in the north. In the central interior cold fronts tend to share their influence almost equally with other synoptic-scale mechanisms.

Evaluation of hybrid forecasting approaches for wind speed and power generation time series