Very short-term wind speed forecasting with Bayesian structural break model

This paper examines a new time series method for very short-term wind speed forecasting. The time series forecasting model is based on Bayesian theory and structural break modeling, which could incorporate domain knowledge about wind speed as a prior. Besides this Bayesian structural break model predicts wind speed as a set of possible values, which is different from classical time series model's single-value prediction This set of predicted values could be used for various applications, such as wind turbine predictive control, wind power scheduling. The proposed model is tested with actual wind speed data collected from utility-scale wind turbines.     

An investigation of variable speed operation of diesel generators in hybrid energy systems

This paper provides a preliminary assessment of the performance and economic potential of a hybrid energy system (wind/diesel) power system which includes a variable speed diesel generator. Recent development in power electronics would be utilized to allow asynchronous operation of the diesel generator, while simultaneously delivering constant frequency electric power to the local electrical grid. In addition to the variable speed diesel, the system can include wind and/or solar electric sources. A hybrid energy system model recently developed at the University of Massachusetts is used to simulate this system configuration and other more conventional wind/diesel hybrid energy systems. Experimental data from a series of variable speed diesel generator tests were used to generate a series of fuel consumption curves used in the analytical model. In addition to performance (fuel savings) comparisons for fixed and variable speed systems, economic cost of energy calculations for the various system designs are presented. It is shown that the proposed system could offer both performance and economic advantages.     

A wind map of Bangladesh

Utilization of wind energy in Bangladesh has been slow mainly due to lack of quality wind data. Recent measurements in some places have shown significant wind energy potentials in Bangladesh. In this paper, a wind map is presented which incorporates several microscale features, such as terrain roughness, elevation etc. with a mesoscale model. Several meso-maps were obtained from global databases and a suitable model was chosen and modified for a 30-m elevation. Ground data from various sources were collected and modified for height and land condition adjustments based on local knowledge and GIS information. It was found that, the generated wind map and the modified ground data resemble. Annual average wind speed at 30 m height along the coastal belt is above 5 m/s. Wind speed in northeastern parts is above 4.5 m/s while inland wind speed is around 3.5 m/s for most part of Bangladesh. Small-scale wind turbines could be installed and tested in locations such as St. Martins Island, Cox’s Bazar, Patenga, Bhola, Barguna, Dinajpur, Thakurgaon and Panchagar.     

Short-Term Prediction of Wind Farm Power: A Data Mining Approach

This paper examines time series models for predicting the power of a wind farm at different time scales, i.e., 10-min and hour-long intervals. The time series models are built with data mining algorithms. Five different data mining algorithms have been tested on various wind farm datasets. Two of the five algorithms performed particularly well. The support vector machine regression algorithm provides accurate predictions of wind power and wind speed at 10-min intervals up to 1 h into the future, while the multilayer perceptron algorithm is accurate in predicting power over hour-long intervals up to 4 h ahead. Wind speed can be predicted fairly accurately based on its historical values; however, the power cannot be accurately determined given a power curve model and the predicted wind speed. Test computational results of all time series models and data mining algorithms are discussed. The tests were performed on data generated at a wind farm of 100 turbines. Suggestions for future research are provided.      

Forecasting and simulating wind speed in Corsica by using an autoregressive model

Alternative approaches for generating wind speed time series are discussed. The method utilized involves the use of an autoregressive process model. The model has been applied to three Mediterranean sites in Corsica and has been used to generate 3-hourly synthetic time series for these considered sites. The synthetic time series have been examined to determine their ability to preserve the statistical properties of the Corsican wind speed time series. In this context, using the main statistical characteristics of the wind speed (mean, variance, probability distribution, autocorrelation function), the data simulated are compared to experimental ones in order to check whether the wind speed behavior was correctly reproduced over the studied periods. The purpose is to create a data generator in order to construct a reference year for wind systems simulation in Corsica.     

Cox–Ingersoll–Ross model for wind speed modeling and forecasting

We propose a dynamic model for the squared norm of the wind speed which is a Markov diffusion process. It presents several advantages. Since the transition probability densities are in closed form, it can be calibrated with the maximum likelihood method. It presents nice modeling features both in terms of marginal probability density function and temporal correlation. We have tested the model with real wind speed data set provided by the National Renewable Energy Laboratory. The model fits very well with the data. Besides, we obtained a very good performance in forecasting wind speed at short term. This is an interesting perspective for operational use in industry. Copyright © 2015 John Wiley & Sons, Ltd.     

Estimating daily wind speed under climate change

A semi-empirical downscaling approach is presented to estimate spatial and temporal statistical properties of local daily mean wind speed under global climate change. The present semi-empirical downscaling method consists of two elements. Since general circulation models (GCMs) are able to reproduce the features of the present atmospheric general circulation quite correctly, the first element represents the large-scale circulation of the atmosphere. The second element is a link between local wind speed and large-scale circulation pattern (CP). The linkage is expressed by a stochastic model conditioned on CP types. Parameters of the linkage model are estimated using observed data series; then this model is utilized with GCM-generated CP type data corresponding to a 2 × CO₂ scenario. Under the climate of Nebraska the lognormal distribution is the best two-parameter distribution to describe daily mean wind speed. The space-time variability of wind speed is described by a transformed multivariate autoregressive (AR) process, and the linkage between local wind and large-scale circulation is expressed as a conditional AR process, i.e. the autoregressive parameters depend on the actual daily CP type. The basic tendency of change under 2 × CO₂ climate is a considerable increase of wind speed from the beginning of summer to the end of winter and a somewhat smaller wind decrease in spring.     

Stochastic simulation of hourly average wind speed sequences in Tangiers (Morocco)

Twelve years of hourly average wind speed data are used to build an autoregressive model (AR(2)) to simulate hourly average wind speed (HAWS). The model matches well the characteristics of the experimental values of the wind speed. Tests have been performed to validate the model. Comparisons have been made between generated and real series of data to check if the wind behaviour is reproductible. The model is then used to build up a reference year for Tangiers and may be used to forecast wind speed, with good results.     

Anticipatory Control of Wind Turbines With Data-Driven Predictive Models

The concept of anticipatory control applied to wind turbines is presented. Anticipatory control is based on the model predictive control (MPC) approach. Unlike the MPC method, noncontrollable variables (such as wind speed) are directly considered in the dynamic equations presented in the paper to predict response variables, e.g., rotor speed and turbine power output. To determine future states of the power drive with the dynamic equations, a time series model was built for wind speed. The time series model was fused with the dynamic equations to predict the response variables over a certain prediction horizon. Based on these predictions, an optimization model was solved to find the optimal control settings to improve the power output without incurring large rotor speed changes. As both the dynamic equations and time series model were built by data mining algorithms, no gradient information is available. A modified evolutionary strategy algorithm was used to solve a nonlinear constrained optimization problem. The proposed approach has been tested on the data collected from a 1.5 MW wind turbine.      

Practical ways of evaluating wind speed persistence

Wind speed persistence is a measure of the mean wind speed duration over a given period of time at any location. This definition implies that wind speed persistence means a positive serial correlation in time series. The wind speed persistence provides useful information about the general climatological characteristics of the wind persisting at a given location. Therefore, wind speed persistence should be taken into account in many studies such as weather forecast, site selection for wind turbines and synthetic generation of the wind speed data. On the other hand, if wind direction information is considered together with the wind speed then this type of persistence can be used for additional purposes such as forest fires, dispersion of the air pollutants, building ventilation, etc. In this study, three different methods with some modifications of the previous methods have been applied to the wind speed data obtained from the meteorology stations located at the northwest part of Turkey. These methods are based on autocorrelation function, conditional probability and the wind speed duration curves. It has been shown that the proposed methods clearly reflect the persistence properties of the wind speed in the study area.     

First and second order Markov chain models for synthetic generation of wind speed time series

Hourly wind speed time series data of two meteorological stations in Malaysia have been used for stochastic generation of wind speed data using the transition matrix approach of the Markov chain process. The transition probability matrices have been formed using two different approaches: the first approach involves the use of the first order transition probability matrix of a Markov chain, and the second involves the use of a second order transition probability matrix that uses the current and preceding values to describe the next wind speed value. The algorithm to generate the wind speed time series from the transition probability matrices is described. Uniform random number generators have been used for transition between successive time states and within state wind speed values. The ability of each approach to retain the statistical properties of the generated speed is compared with the observed ones. The main statistical properties used for this purpose are mean, standard deviation, median, percentiles, Weibull distribution parameters, autocorrelations and spectral density of wind speed values. The comparison of the observed wind speed and the synthetically generated ones shows that the statistical characteristics are satisfactorily preserved.     

Influence of local wind speed and direction on wind power dynamics – Application to offshore very short-term forecasting

Wind power time series usually show complex dynamics mainly due to non-linearities related to the wind physics and the power transformation process in wind farms. This article provides an approach to the incorporation of observed local variables (wind speed and direction) to model some of these effects by means of statistical models. To this end, a benchmarking between two different families of varying-coefficient models (regime-switching and conditional parametric models) is carried out. The case of the offshore wind farm of Horns Rev in Denmark has been considered. The analysis is focused on one-step ahead forecasting and a time series resolution of 10 min. It has been found that the local wind direction contributes to model some features of the prevailing winds, such as the impact of the wind direction on the wind variability, whereas the non-linearities related to the power transformation process can be introduced by considering the local wind speed. In both cases, conditional parametric models showed a better performance than the one achieved by the regime-switching strategy. The results attained reinforce the idea that each explanatory variable allows the modelling of different underlying effects in the dynamics of wind power time series.     

Fuzzy modeling techniques and artificial neural networks to estimate annual energy output of a wind turbine

The purpose of this article is to develop a new method to estimate annual energy output for a given wind turbine in any region which should be easy to use and has satisfactory accuracy. To do this, hourly wind speeds of 25 different stations in Netherlands, output power curve of S47 wind turbine and fuzzy modeling techniques and artificial neural networks were used and a model is developed to estimate annual energy output for S47 wind turbine in different regions. Since this model has three inputs (average wind speed, standard deviation of wind speed, and air density of that region), this model is easy to use. The accuracy of this method is compared with the accuracy of conventional methods and it is shown that this new method performs better. Thereafter, we have shown that by making some small changes to this proposed model, other pitch control wind turbines could be modeled too. As an example, we have modeled E82 wind turbine based on the model developed for S47 and it is shown that this model has still satisfactory accuracy.     

Prediction of wind speed time series using modified Taylor Kriging method

Wind speed forecasting is critical for the operations of wind turbine and penetration of wind energy into electricity systems. In this paper, a novel time series forecasting method is proposed for this purpose. This method originates from TK (Taylor Kriging) model, but is properly modified for the forecasting of wind speed time series. To investigate the performance of this new method, the wind speed data from an observation site in North Dakota, USA, are adopted. One-year hourly wind speed data are divided into 10 samples, and forecast is made for each sample. In the case study, both the modified TK method and (ARIMA) autoregressive integrated moving average method are employed and their performances are compared. It is found that on average, the proposed method outperforms the ARIMA method by 18.60% and 15.23% in terms of (MAE) mean absolute error and (RMSE) root mean square error. Meanwhile, further theoretical analysis is provided to discuss why the modified TK method is potentially more accurate than the ARIMA method for wind speed time series prediction.     

Wind generator potentiality in Somalia

The wind data of several measurement sites in Somalia have been analysed in order to characterize the wind potentiality in relation to the type of wind generators; these have been defined by a simple model of the system output. The relation between machine and local frequency distribution as to energy extraction can be defined by a parameter (“site effectiveness”), which is maximized by a suitable combination of the rated and cut-in wind speed. On this basis it is shown that Somalia is characterized by wind frequency distributions that can be exploited in the best way by relatively slow rather than fast wind machines.     

Wind energy potential assessment in Uttara Kannada district of Karnataka, India

Availability of wind energy and its characteristics at Kumta and Sirsi in Uttara Kannada District of Karnataka has been studied based on primary data collected at these sites for a period of 24 months. Wind regimes at Karwar (1952–1989), Honnavar (1939–1989) and Shirali (1974–1989) have also been analysed based on data collected from India Meteorological Department (IMD) of respective meterological observatories. Wind energy conversion systems would be most effective in these taluks during the period May to August. The monthly frequency distributions of wind speed have been analysed for Kumta and Sirsi where hourly wind speed recording is available. It is shown that two parameter Weibull distribution is a good representation of the probability density function for the wind speed. Energy Pattern Factor (EPF) and Power Densities are computed for sites at Kumta and Sirsi. With the knowledge of EPF and mean wind speed, mean power density is computed for Karwar, Honnavar and Shirali. Our analyses show that the coastal taluks such as Karwar and Kumta have good wind potential. This potential, if exploited would help local industries and coconut and areca plantations. Premonsoon availability of wind energy would help in irrigating these orchards and makes the wind energy a desirable alternative.     

Synthesis of long-term hourly wind speed time series on the basis of European wind atlas data

A method for the synthesis of annual wind speed time series with a time resolution of 1 hour is presented. It is based upon statistical information on the wind climate given in the European Wind Atlas. The synthetic time series reproduce the monthly average daily time pattern of the site. The distribution of the synthetic wind speed data shows the correct mean value of the cubed wind speed. The site-specific variance of the wind speed and the power spectrum of the wind speed fluctuations are closely approximated. Results of time step simulations for small stand-alone wind energy systems using synthetic and measured data sets as input data show a close agreement.     

A stochastic Markov chain model for simulating wind speed time series at Tangiers, Morocco

Time series of then years of Hourly Average Wind Speed, HAWS, data from Tangiers station are analysed on a statistical basis by applying the Markovian process. The limiting behaviour of the Markov chain is then examined and compared to the histogram of observed wind speed. It was found that a 12×12 transition probability matrix was necessary to generate an acceptable synthetic time series. The manner in which the Markovian model can be used to generate wind speed time series are also described.Using the transition probability matrix developed from the real wind data, the synthetic wind speed time series are generated. The comparison between the real wind speed and the synthetic one shows that the statistical characteristics of wind speed are faithfully reproduced. The synthetic HAWS may be utilised as input data for any wind energy system.     

基于时间序列模型的风电场风速预测研究

基于时间序列的方法,对风速的长期预测进行了研究,并在工程应用的基础上提出了新的预测思路:首先将风速信号分解成趋势信号和去趋势项随机信号,然后分别用滑动滤波和小波分析这2种方法对分解出的去趋势项随机信号进行数据处理并比较,再用时间序列的方法对趋势项信号和处理后的信号分别进行预测并叠加,得到最后的预测风速信号.结果表明:五项滑动滤波处理数据的方法与Daubechies小波分解法均能实现精度较高的风速长期预测;与小波分解法相比,滑动滤波方法算法的复杂性低,在工程应用上可行性更高.