Short term wind speed forecasting for wind turbine applications using linear prediction method

In this paper a new method, based on linear prediction, is proposed for wind speed forecasting. The method utilizes the ‘linear prediction’ method in conjunction with ‘filtering’ of the wind speed waveform. The filtering eliminates the undesired parts of the frequency spectrum (i.e. smoothing) of the measured wind speed which is less effective in an application, for example, in a wind energy conversion system. The linear prediction method is intuitively explained with some easy to follow case studies to clarify the complex underlying mathematics. For verification purposes, the proposed method is compared with real wind speed data based on experimental results. The results show the effectiveness of the linear prediction method.     

Multiple architecture system for wind speed prediction

A new approach based on multiple architecture system (MAS) for the prediction of wind speed is proposed. The motivation behind the proposed approach is to combine the complementary predictive powers of multiple models in order to improve the performance of the prediction process. The proposed MAS can be implemented by associating the predictions obtained from the different regression algorithms (MLR, MLP, RBF and SVM) making up the ensemble by three fusion strategies (simple, weighted and non-linear). The efficiency of the proposed approach has been assessed on a real data set recorded from seven locations in Algeria during a period of 10 years. The experimental results point out that the proposed MAS approach is capable of improving the precision of the wind speed prediction compared to the traditional prediction methods.     

Support vector machines for wind speed prediction

This paper introduces support vector machines (SVM), the latest neural network algorithm, to wind speed prediction and compares their performance with the multilayer perceptron (MLP) neural networks. Mean daily wind speed data from Madina city, Saudi Arabia, is used for building and testing both models. Results indicate that SVM compare favorably with the MLP model based on the root mean square errors between the actual and the predicted data. These results are confirmed for a system with order 1 to system with order 11.     

Local models-based regression trees for very short-term wind speed prediction

This paper evaluates the performance of different types of Regression Trees (RTs) in a real problem of very short-term wind speed prediction from measuring data in wind farms. RT is a solidly established methodology that, contrary to other soft-computing approaches, has been under-explored in problems of wind speed prediction in wind farms. In this paper we comparatively evaluate eight different types of RTs algorithms, and we show that they are able obtain excellent results in real problems of very short-term wind speed prediction, improving existing classical and soft-computing approaches such as multi-linear regression approaches, different types of neural networks and support vector regression algorithms in this problem. We also show that RTs have a very small computation time, that allows the retraining of the algorithms whenever new wind speed data are collected from the measuring towers.     

A novel method based on Weibull distribution for short-term wind speed prediction

Wind speed prediction (WSP) is essential in order to predict and analyze efficiency and performance of wind-based electricity generation systems. More accurate WSP may provide better opportunities to design and build more efficient and robust wind energy systems. Precious short-term prediction is difficult to achieve; therefore several methods have been developed so far. We notice that the statistics of the alterations, which occur between sequential values of the predicted wind speed data, may differ significantly from observed wind statistics. In this study, we investigate these alterations and compare them and, accordingly, propose a novel method based on Weibull and Gaussian probability distribution functions (PDF) for short-term WSP. The proposed method stands on an algorithm, which examines comparison of the statistical features of the observed and generated wind speed in order to achieve more accurate estimation. We have examined this method on the wind speed data set observed and recorded in Ankara in 2013 and in 2014. The obtained results show that the new algorithm provides better wind speed prediction with an enhanced wind speed model.     

Extraction of synoptic pressure patterns for long-term wind speed estimation in wind farms using evolutionary computing

In this paper we present an evolutionary approach for the problem of discovering pressure patterns under a quality measure related to wind speed and direction. This clustering problem is specially interesting for companies involving in the management of wind farms, since it can be useful for analysis of results of the wind farm in a given period and also for long-term wind speed prediction. The proposed evolutionary algorithm is based on a specific encoding of the problem, which uses a dimensional reduction of the problem. With this special encoding, the required centroids are evolved together with some other parameters of the algorithm. We define a specific crossover operator and two different mutations in order to improve the evolutionary search of the proposed approach. In the experimental part of the paper, we test the performance of our approach in a real problem of pressure pattern extraction in the Iberian Peninsula, using a wind speed and direction series in a wind farm in the center of Spain. We compare the performance of the proposed evolutionary algorithm with that of an existing weather types (WT) purely meteorological approach, and we show that the proposed evolutionary approach is able to obtain better results than the WT approach.     

A joint probability density function of wind speed and direction for wind energy analysis

A very flexible joint probability density function of wind speed and direction is presented in this paper for use in wind energy analysis. A method that enables angular–linear distributions to be obtained with specified marginal distributions has been used for this purpose. For the marginal distribution of wind speed we use a singly truncated from below Normal–Weibull mixture distribution. The marginal distribution of wind direction comprises a finite mixture of von Mises distributions. The proposed model is applied in this paper to wind direction and wind speed hourly data recorded at several weather stations located in the Canary Islands (Spain). The suitability of the distributions is judged from the coefficient of determination R².

The conclusions reached are that the joint distribution proposed in this paper: (a) can represent unimodal, bimodal and bitangential wind speed frequency distributions, (b) takes into account the frequency of null winds, (c) represents the wind direction regimes in zones with several modes or prevailing wind directions, (d) takes into account the correlation between wind speeds and its directions. It can therefore be used in several tasks involved in the evaluation process of the wind resources available at a potential site. We also conclude that, in the case of the Canary Islands, the proposed model provides better fits in all the cases analysed than those obtained with the models used in the specialised literature on wind energy.     

Transfer learning for short-term wind speed prediction with deep neural networks

As a type of clean and renewable energy source, wind power is widely used. However, owing to the uncertainty of wind speed, it is essential to build an accurate forecasting model for large-scale wind power penetration. Numerical weather prediction (NWP) and data-driven modeling are two typical paradigms. NWP is usually unavailable or spatially insufficient. Data-driven modeling is an effective candidate. As to some newly-built wind farms, sufficient historical data is not available for training an accurate model, while some older wind farms may have long-term wind speed records. A question arises regarding whether the prediction model trained by data coming from older farms is also effective for a newly-built farm. In this paper, we propose an interesting trial of transferring the information obtained from data-rich farms to a newly-built farm. It is well known that deep learning can extract a high-level representation of raw data. We introduce deep neural networks, trained by data from data-rich farms, to extract wind speed patterns, and then finely tune the mapping with data coming from newly-built farms. In this way, the trained network transfers information from one farm to another. The experimental results show that prediction errors are significantly reduced using the proposed technique.     

Fuzzy copula model for wind speed correlation and its application in wind curtailment evaluation

Wind parks always produce diverse percentages of their nominal power at the same time, leading to a concern about correlation between wind speeds. The assessments of wind speed correlation have been particularly focused on probabilistic modeling of aleatory uncertainty. However, poor historical data, imprecise parameter estimation and incomplete knowledge of wind speeds lead to another type of uncertainty, possibilistic uncertainty, which requires an explicit analysis. Therefore, a fuzzy copula model is firstly proposed to express the possibilistic uncertainty of wind speed correlation. The advantage of the proposed model is that the copula parameters can be interval numbers, triangular or trapezoidal fuzzy numbers based on the wind speed data and subjective judgment of decision makers. For estimating copula parameters, a complete decision rule and interval estimation method is developed based on cumulative probability and probability distributions of correlated wind speeds. The effectiveness of the proposed model is validated by the application in wind curtailment evaluation while a method is developed to evaluate and quantify wind curtailment in a hybrid power system involving different types of generation. The results demonstrate that the proposed model and method are capable of describing the possibilistic uncertainty and evaluating its effect on wind curtailment. Compared with previous research, the proposed model develops a new universal parameter estimation method and selection rule to provide more interval results, by calculating the membership function of copula parameters and wind curtailment. System planners and operators can apply the fuzzy results to various topics like reserve capacity evaluation or real-time dispatch depending on their level of risk tolerance.     

Characterization of wind speed data according to wind direction

Knowledge of the wind speed distribution and the most frequent wind directions is important when choosing wind turbines and when locating them. For this reason wind evaluation and characterization are important when forecasting output power. The data used here were collected from eleven meteorological stations distributed in Navarre, Spain. We obtained data for the period extending from 1992 to 1995, with each datum encompassing 10 minutes of time. Wind speed data of each station were gathered in eight directional sectors, each one extended over 45 degrees according to the direction from which the wind blows. The stations were grouped in two blocks: those under the influence of the Ebro valley and those in mountainous areas. For each group the Weibull parameters were estimated, (according to the Weibull probability paper because the Weibull distribution gives the best fit in this region). Kurtosis and skewness coefficients were estimated as well. The Weibull parameters, especially the scale parameter c, depend strongly on the direction considered, and both Weibull parameters show an increasing trend as the direction considered moves to the more dominant direction, while both kurtosis and skewness show a corresponding decreasing trend.     

Hybrid boosting algorithms and artificial neural network for wind speed prediction

Energy sources are an important foundation for national economic growth. The future of energy sources depend on the energy controls. The reserves of fossil energy have declined significantly, and environmental pollution has increased dramatically due to excessive fossil fuel consumption. At this point, wind energy can be used as one of the key source of renewable energy. It has a remarkable importance among the low-carbon energy technologies. The primary aim of wind energy production is to reduce dependence on fossil fuels that affect environment adversely. Therefore, wind energy is analyzed to develop new energy resources. The main issue related to evaluation of the wind energy potential is wind speed prediction. Due to the high volatile and irregular nature of wind speed, wind speed prediction is difficult. To cope with complex data structure, this study presents the development of extreme gradient boosting (XGBoost), adaptive boosting (AdaBoost), and artificial neural network (ANN) within particle swarm optimization (PSO) parameter optimization for hourly wind speed prediction. To compare the proposed hybrid methods, various performance measures, the Pearson's test, and the Taylor diagram are used. The results showed that proposed hybrid methods provide reasonable prediction results for wind speed prediction.     

Dynamic models of wind farms with fixed speed wind turbines

The increasing wind power penetration on power systems requires the development of adequate wind farms models for representing the dynamic behaviour of wind farms on power systems. The behaviour of a wind farm can be represented by a detailed model including the modelling of all wind turbines and the wind farm electrical network. But this detailed model presents a high order model if a wind farm with high number of wind turbines is modelled and therefore the simulation time is long. The development of equivalent wind farm models enables the model order and the computation time to be reduced when the impact of wind farms on power systems is studied. In this paper, equivalent models of wind farms with fixed speed wind turbines are proposed by aggregating wind turbines into an equivalent wind turbine that operates on an equivalent wind farm electrical network. Two equivalent wind turbines have been developed: one for aggregated wind turbines with similar winds, and another for aggregated wind turbines under any incoming wind, even with different incoming winds.The proposed equivalent models provide high accuracy for representing the dynamic response of wind farm on power system simulations with an important reduction of model order and simulation time compare to that of the complete wind farm modelled by the detailed model.     

Application of artificial neural networks for the wind speed prediction of target station using reference stations data

In this study, artificial neural networks (ANNs) were applied to predict the mean monthly wind speed of any target station using the mean monthly wind speeds of neighboring stations which are indicated as reference stations. Hourly wind speed data, collected by the Turkish State Meteorological Service (TSMS) at 8 measuring stations located in the eastern Mediterranean region of Turkey were used. The long-term wind data, containing hourly wind speeds, directions and related information, cover the period between 1992 and 2001. These data were divided into two sections. According to the correlation coefficients, reference and target stations were defined. The mean monthly wind speeds of reference stations were used and also corresponding months were specified in the input layer of the network. On the other hand, the mean monthly wind speed of the target station was utilized in the output layer of the network. Resilient propagation (RP) learning algorithm was applied in the present simulation. The hidden layers and output layer of the network consist of logistic sigmoid transfer function (logsig) and linear transfer function (purelin) as an activation function. Finally, the values determined by ANN model were compared with the actual data. The maximum mean absolute percentage error was found to be 14.13% for Antakya meteorological station and the best result was found to be 4.49% for Mersin meteorological station.     

Assessing the suitability of wind speed probabilty distribution functions based on wind power density

Three functions have so far predominantly been used for fitting the measured wind speed probability distribution in a given location over a certain period of time, typically monthly or yearly. In the literature, it is common to fit these functions to compare which one fits the measured distribution best in a particular location. During this comparison process, parameters on which the suitability of the fit is judged are required. The parameters that are mostly used are the mean wind speed or the total wind energy output (primary parameters). It is, however, shown in the present study that one cannot judge the suitability of the functions based on the primary parameters alone. Additional parameters (secondary parameters) that complete the primary parameters are required to have a complete assessment of the fit, such as the discrepancy between the measured and fitted distributions, both for the wind speed and wind energy (that is the standard deviation of wind speed and wind energy distributions). Therefore, the secondary statistical parameters have to be known as well as the primary ones to make a judgement about the suitability of the distribution functions analysed. The primary and secondary parameters are calculated from the 12-month of measured hourly wind speed data and detailed analyses of wind speed distributions are undertaken in the present article.     

A full-scale prediction method for wind turbine rotor noise by using wind tunnel test data

The development of a low-noise wind turbine rotor and propeller is often cost-effective and is in fact a race against time to those who wish to build and test a small-scale rotor instead of an expensive full-scale rotor. The issue of this approach has to do with the interpretation of wind tunnel model test data in terms of both the frequency band and sound pressure level information for the noise scaling effect.This paper discusses a prediction method for the estimation of the noise generated from a full-scale wind turbine rotor using wind tunnel test data measured with both a small-scale rotor and a 2D section of the blade. The 2D airfoil self-noise and the scaled rotor noise were investigated with a series of wind tunnel experiments. Wind tunnel data post-processing considered four aspects: removal of the test condition effect, scaling to full scale, consideration of the wind turbine rotor operating conditions, and the most important terms of full-scale rotor noise as adjustments to address the differences between the wind tunnel test conditions and the full-scale operating conditions.A full-scale rotor noise prediction results comparison was performed by initially dividing the test conditions into the condition of a 2D section noise test and the condition of a small-scale rotor noise test. Based on an airfoil section, the rotor was selected from a blade section at r/R = 0.75. The small-scale rotor was scaled down by a factor of 5.71 for the wind tunnel test.Finally, the full-scale rotor noise data was compared with the wind tunnel test data using a scaling estimation method.     

On wind speed pattern and energy potential in Nigeria

The aim of this paper is to review wind speed distribution and wind energy availability in Nigeria and discuss the potential of using this resource for generation of wind power in the country. The power output from a wind turbine is strongly dependent on the wind speed and accurate information about the wind data in a targeted location is essential. The annual mean wind speeds in Nigeria range from about 2 to 9.5 m/s and the annual power density range between 3.40 and 520 kW/m² based on recent reported data. The trend shows that wind speeds are low in the south and gradually increases to relatively high speeds in the north. The areas that are suitable for exploitation of wind energy for electricity generation as well as for water pumping were identified. Also some of the challenges facing the development of wind energy and suggested solutions were presented.     

Heterogeneous mixture distributions for modeling wind speed,application to the UAE

Heterogeneous mixture distributions (HTM) have not been employed for wind speed modeling of the Arabian Peninsula. In order to improve our understanding of wind energy potential in the Arabian Peninsula, HTM should be tested for the frequency analysis of wind speed. The aim of the current study is to assess the suitability of HTMs and identify the most appropriate probability distribution to model wind speed data in the UAE. Hourly mean wind speed data were used in the current study. Ten homogeneous and heterogeneous mixture distributions were used and constructed by mixing the four following probability distributions: Gamma, Weibull, Extreme value type-one, and Normal distributions. The Weibull and Kappa distributions were also employed as representatives of the conventional non-mixture distributions. Maximum Likelihood, Expectation Maximization algorithm, and Least Squares methods were employed to fit the mixture distributions. Results indicate that mixture distributions give the best fit to wind speed data for all stations. Wind speed data of five stations show strong mixture distributional characteristics. Applications of HTMs show a significant improvement in explaining the whole wind speed regime. The Weibull-Extreme value type-one mixture distribution is considered the most appropriate distribution for wind speed data in the UAE.     

Design requirements for medium-sized wind turbines for remote and hybrid power systems

This paper provides an overview of the design requirements for medium-sized wind turbines intended for use in a remote hybrid power system. The recommendations are based on first-hand experience acquired at the University of Massachusetts through the installation, operation, and upgrade of a 250-kW turbine on a mountain top with difficult access in Western Massachusetts. Experience with the operation of this turbine and the design of its control system, together with a long history in the design and analysis of hybrid power systems, has made it possible to extend the work in Western Massachusetts to remote or hybrid power systems in general. The University test site has many attributes of more remote sites and the overall wind turbine installation is typical of one that could power a hybrid wind system. For example, access to the site is limited due to steep terrain, snow, and environmental restrictions. Also, the power lines feeding the turbine exhibit voltage sags and phase imbalance, especially during start-up. This paper is based on the experience gained from the operation of this wind turbine and assesses the requirements for the design and operation of medium to large wind turbines in remote locations. The work summarizes lessons learned relative to: (1) sensors, communication, and control capabilities; (2) grid connection issues; and (3) weather-related problems. The final section of the paper focuses on design requirements to ensure successful installation and the completion of maintenance and repairs at remote sites.     

Low wind speed turbines and wind power potential in Minnesota, USA

Wind power development in Minnesota largely has been focused in the “windy” southwestern part of the state. This research evaluates the additional power that potentially could be generated via low wind speed turbines, particularly for areas of the state where there has been comparatively little wind energy investment. Data consist of 3 years (2002–2004) of wind speed measurements at 70–75 m above ground level, at four sites representing the range of wind speed regimes (Classes 2–5) found in Minnesota. Power estimates use three configurations of the General Electric 1.5-MW series turbine that vary in rotor diameter and in cut-in, cut-out, and rated speeds. Results show that lower cut-in, cut-out, and rated speeds, and especially the larger rotor diameters, yield increases of 15–30% in wind power potential at these sites. Gains are largest at low wind speed (Class 2) sites and during the summer months at all four sites. Total annual wind power at each site shows some year-to-year variability, with peaks at some sites partially compensating for lulls at others. Such compensation does not occur equally in all years: when large-scale atmospheric circulation patterns are strong (e.g., 2002), the four sites show similar patterns of above- and below-average wind power, somewhat reducing the ability of geographic dispersion to mitigate the effects of wind speed variability.