No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Short-term wind power forecasting with WRF-ARW model and genetic programming
Abstract
Forecasting wind power in the short-term usually involves the use of numerical weather prediction models. These models need to run at very high resolutions to provide the best forecasts possible. Producing high resolution forecasts is resource and time consuming, which can be a problem when the forecasts need to be available for the grid operator on the day-ahead. This paper introduces a novel approach for short-term wind power prediction by combining the Weather Research and Forecasting - Advanced Research WRF model (WRF- ARW) with genetic programming, using the latter one for final downscaling and prediction technique, estimating the total hourly power output on the day ahead at a wind farm located in Galicia, Spain.
... Some studies involved models that are a combination of physical and statistical methods, also referred to as hybrid methods, to predict wind characteristics. For instance, Che et al. [19] utilized the Kalman filter, and Martínez-Arellano et al. [20] employed a genetic algorithm to clean out the physical model output for wind farms in Awaji, Japan, and Satavento, Galicia, respectively. These studies include the forecasting time horizon ranging from 24 to 72 h however did not consider the wake effects of wind turbines while forecasting wind characteristics. ...
... 1-12, 14-15, 33) of wind turbines is less influenced by the upstream wind farms wakes than the Set B (turbines no. 13,[16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32]. Three wind turbines from each row in both sets were selected to analyze the wind speed and power variation experienced over seven days in June and January. ...
Forecasting skills for a wind farm would significantly degrade if the complex wake effects of the upstream wind farms are excluded, especially when they are spatially close to each other. In this study, the Weather Research and Forecasting (WRF) model has been used to predict wind speed and power for a wind farm in Pakistan in the presence of wake interference from neighboring wind farms for two different seasons. Forecasting is done for two different cases i.e. without and with inter-farm wake effects, and different statistical error parameters were evaluated based on the real observations. A significant reduction in errors was observed in the latter case. For instance, the mean absolute errors in wind speed prediction were reduced by 7.7% and 14% in June (summer) and January (winter) respectively, by the inclusion of inter-farm wake effects. Similarly, an improved forecast of power output was obtained by incorporating the interaction of upstream wind farms i.e. a reduction of 15% and 26% in the normalized mean absolute error in power output values was observed for June and January, respectively. However, the prediction accuracy of power output substantially deteriorated in the winter season.
... Some recent studies have been developed about the shortterm predictability of wind speeds with the use of dynamic, mathematical and statistical tools using Numerical Weather Prediction (NWP) (Cheng et al., 2017;Martinez-Arellano and Nolle, 2013), stochastic (Monteiro and Souza, 2013) and hybrid (Camelo et al., 2018b,a;Krishnaveny et al., 2017) models. Ramos et al. (2013) investigated the prediction of hourly wind speeds at 30 m above ground with the atmospheric model WRF (Weather Research Forecasting) for the State of Alagoas-Brazil. ...
The predictability of wind information in a given location is essential for the evaluation of a wind power project. Predicting wind speed accurately improves the planning of wind power generation, reducing costs and improving the use of resources. This paper seeks to predict the mean hourly wind speed in anemometric towers (at a height of 50 m) at two locations: a coastal region and one with complex terrain characteristics. To this end, the Holt–Winters (HW), Artificial Neural Networks (ANN) and Hybrid time-series models were used. Observational data evaluated by the Modern-Era Retrospective analysis for Research and Applications-Version 2 (MERRA-2) reanalysis at the same height of the towers. The results show that the hybrid model had a better performance in relation to the others, including when compared to the evaluation with MERRA-2. As such, the hybrid models are a good method to forecast wind speed data for wind generation.
... Modifying coefficients from the Kain-Frisch convective scheme in the model can improve the precipitation forecast in a tropical cyclone [6]. Wind speed forecasts can be improved using GP to perform a symbolic regression from a set of past forecasts obtained from the WRF-ARW grid [7]. Venkadesh et al. [8] use genetic algorithms to determine the optimal duration and resolution of prior data for weather variables that was considered a potential input for an ANN model. ...
The effective short-term forecasting of wind speeds for wind farm applications requires the tailored application of a collection of various forecasting tools. Physical forecasting methods generally form the basis of these forecasting tools, coupled with regression and statistical methods. In this work, the Weather Research and Forecasting (WRF) model is utilized to generate an ensemble spread for a targeted site. The optimal deterministic forecast is subsequently determined using two techniques: a novel application of the mean-variance portfolio theorem; as well as a two-layered feed forward neural network. The mean-variance portfolio optimization technique is assessed for its suitability for weighted ensemble selection by comparing its optimized deterministic forecast accuracy with a neural network approach.
Bibliografia recuperada de la Revisión Sistemática de Literatura de pronóstico de velocidad de viento para generación eólica desde el año 2012 hasta el año 2017
Thesis to receive the degree of doctor in philosophy. The thesis is about the application of machine learning techniques such as Genetic Programming and Fuzzy Logic to forecast wind power and characterise ramp events at a wind farm in Galicia, Spain.
Two new postprocessing methods are proposed to reduce numerical weather prediction's systematic and random errors. The first method consists of running a postprocessing algorithm inspired by the Kalman filter (KF) through an ordered set of analog forecasts rather than a sequence of forecasts in time (ANKF). The analog of a forecast for a given location and time is defined as a past prediction that matches selected features of the current forecast. The second method is the weighted average of the observations that verified when the 10 best analogs were valid (AN). ANKF and AN are tested for 10-m wind speed predictions from the Weather Research and Forecasting (WRF) model, with observations from 400 surface stations over the western United States for a 6-month period. Both AN and ANKF predict drastic changes in forecast error (e.g., associated with rapid weather regime changes), a feature lacking in KF and a 7-day running-mean correction (7-Day). The AN almost eliminates the bias of the raw prediction (Raw), while ANKF drastically reduces it with values slightly worse than KF. Both analog-based methods are also able to reduce random errors, therefore im-proving the predictive skill of Raw. The AN is consistently the best, with average improvements of 10%, 20%, 25%, and 35% with respect to ANKF, KF, 7-Day, and Raw, as measured by centered root-mean-square error, and of 5%, 20%, 25%, and 40%, as measured by rank correlation. Moreover, being a prediction based solely on observations, AN results in an efficient downscaling procedure that eliminates representativeness dis-crepancies between observations and predictions.
The Technical Note series provides an outlet for a variety of NCAR manuscripts that contribute in specialized ways to the body of scientific knowledge but which are not suitable for journal, monograph, or book publication. Reports in this series are issued by the NCAR Scientific Divisions; copies may be obtained on request from the Publications Office of NCAR. Designation symbols for the series include: EDD: IA:
We present a computational framework for integrating a state-of-the-art numerical weather prediction (NWP) model in stochastic unit commitment/economic dispatch formulations that account for wind power uncertainty. We first enhance the NWP model with an ensemble-based uncertainty quantification strategy implemented in a distributed-memory parallel computing architecture. We discuss computational issues arising in the implementation of the framework and validate the model using real wind-speed data obtained from a set of meteorological stations. We build a simulated power system to demonstrate the developments.
Wind speed and the related electrical power of wind turbines are forecasted. The work is focused on the operation of power systems with integrated wind parks. Artificial neural networks models are proposed for forecasting average values of the following 10 min or 1 h. Input quantities for the prediction are wind speeds and their derivatives. Also, spatial correlation of wind speeds and its use for forecasting, are investigated. The methods are tested using data collected over seven years at six different sites on islands of the South and Central Aegean Sea in Greece.
This paper presents an advanced statistical method for wind power forecasting based on artificial intelligence techniques. The method requires as input past power measurements and meteorological forecasts of wind speed and direction interpolated at the site of the wind farm. A self-organized map is trained to classify the forecasted local wind speed provided by the meteorological services. A unique feature of the method is that following a preliminary wind power prediction, it provides an estimation of the quality of the meteorological forecasts that is subsequently used to improve predictions. The proposed method is suitable for operational planning of power systems with increased wind power penetration, i.e., forecasting horizon of 48 h ahead and for wind farm operators trading in electricity markets. Application of the forecasting method on the power production of an actual wind farm shows the validity of the method
Wind power forecasting system is useful to increase the wind energy penetration level. Latest statistics show that China has been the biggest wind energy market throughout the world. However, few studies have been published to introduce the wind energy forecasting technologies in China. This paper presents the performance evaluation and accuracy enhancement of a novel day-ahead wind power forecasting system in China. This system consists of a numerical weather prediction (NWP) model and artificial neural networks (ANNs). The NWP model is established by coupling the Global Forecasting system (GFS) with the Weather Research and Forecasting (WRF) system together to predict meteorological parameters. In addition, Kalman filter has been integrated in this system to reduce the systematic errors in wind speed from WRF and enhance the forecasting accuracy. The numerical results from a real world case are proven the effectiveness of this forecasting system in terms of the raw wind speed correction and wind power forecasting accuracy. The Normalized Root Mean Square Error (NRMSE) has a month average value of 16.47%, which is an acceptable error margin for allowing the use of the forecasted values in electric market operations. This forecasting system is profitable for increasing the wind energy penetration level in China.
This paper introduces support vector machines (SVM), the latest neural network algor-ithm, to wind speed prediction and compares their performance with the multilayer percep-tron (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 pre-dicted data. These results are confirmed for a system with order 1 to system with order 11.
Predicting wind power is considered as one of the most important tasks for the large-scale integration of intermittent wind-powered generators into power systems. Given the cubic relationship between wind speed and wind power, accurate forecasting of wind speed is important for the estimation of future wind power generation output. This paper presents a short-term wind speed forecasting technique using a hybrid intelligent algorithm that utilizes a data filtering technique based on wavelet transform (WT) and a soft computing model (SCM) based on fuzzy ARTMAP (FA) network. The effectiveness of the proposed hybrid WT+FA model is evaluated by comparing it with various other SCMs as well as hybrid models. The test results show that a significant reduction in forecast error of an individual FA network by more than 40% through the application of a combined FA and WT. The forecasting performance of the proposed WT+FA is not only robust and more effective than that of individual FA network but also it shows superiority over other considered SCMs. The forecasting techniques were tested using the real data from the North Cape wind farm located in PEI, Canada.
This paper proposes a wavelet support vector machine (WSVM)-based model for short-term wind power prediction (WPP). A new wavelet kernel is proposed to improve the generalization ability of the support vector machine (SVM). The proposed kernel has such a general characteristic that some commonly used kernels are its special cases. Simulation studies are carried to validate the proposed model with different prediction schemes by using the data obtained from the National Renewable Energy Laboratory (NREL). Results show that the proposed model with a fixed-step prediction scheme is preferable for short-term WPP in terms of prediction accuracy and computational cost. Moreover, the proposed model is compared with the persistence model and the SVM model with radial basis function (RBF) kernels. Results show that the proposed model not only significantly outperforms the persistence model but is also better than the RBF-SVM in terms of prediction accuracy.
Despite the major progress made by Numerical Weather Prediction (NWP) in the last decades, meteorological models are usually unable to provide reliable surface wind speed forecasts, especially in complex topography regions, because of shortcomings in horizontal resolution, physical parameterisations, initial and boundary conditions. In order to reduce these drawbacks, one of the most successful approaches is the Kalman filtering technique, which combines recursively observations and model forecasts to minimise the corresponding biases. In meteorology, Kalman filters are widely used to improve the prediction of variables characterised by well-defined cyclicities, whereas the evolution of wind speed is usually too irregular. In the present paper, the Kalman filter is analysed in order to find the best configuration for wind speed and wind power forecast. The procedure has been tested, in a hindcast mode, with 2-year-long data sets of wind speed provided by a NWP model and two anemometric stations located in the eastern Liguria (Italy). It is shown that, tuning time step and forecast horizon of the filter, this methodology is capable to provide significant forecast improvement with respect to the wind speed model direct output, especially when used for very short-term forecast. In this configuration, Kalman-filtered wind speed data have been used to forecast the wind energy output of the nearby wind farm of Varese Ligure. After 2 years of testing, the percentage error between simulated and measured wind energy values was still very low and showed a stable evolution.
Wind power generation differs from conventional thermal generation due to the stochastic nature of wind. Thus wind power forecasting plays a key role in dealing with the challenges of balancing supply and demand in any electricity system, given the uncertainty associated with the wind farm power output. Accurate wind power forecasting reduces the need for additional balancing energy and reserve power to integrate wind power. Wind power forecasting tools enable better dispatch, scheduling and unit commitment of thermal generators, hydro plant and energy storage plant and more competitive market trading as wind power ramps up and down on the grid. This paper presents an in-depth review of the current methods and advances in wind power forecasting and prediction. Firstly, numerical wind prediction methods from global to local scales, ensemble forecasting, upscaling and downscaling processes are discussed. Next the statistical and machine learning approach methods are detailed. Then the techniques used for benchmarking and uncertainty analysis of forecasts are overviewed, and the performance of various approaches over different forecast time horizons is examined. Finally, current research activities, challenges and potential future developments are appraised.Highlights► This paper presents an in-depth review of current methods and advances in wind power forecasting. ► We discuss numerical wind prediction from global to local scales, ensemble forecasting, upscaling and downscaling processes. ► We detail statistical, machine learning approach techniques for benchmarking and uncertainty analysis of forecasts. ► We examine the performance of various approaches over different forecast time horizons. ► We appraise current research activities, challenges and potential future developments in wind power forecasting.
Seven adaptive approaches to post-processing wind speed forecasts are discussed and compared. Forecasts of the wind speed over 48 h are run at horizontal resolutions of 7 and 3 km for a domain centred over Ireland. Forecast wind speeds over a 2 year period are compared to observed wind speeds at seven synoptic stations around Ireland and skill scores calculated. Two automatic methods for combining forecast streams are applied. The forecasts produced by the combined methods give bias and root mean squared errors that are better than the numerical weather prediction forecasts at all station locations. One of the combined forecast methods results in skill scores that are equal to or better than all of its component forecast streams. This method is straightforward to apply and should prove beneficial in operational wind forecasting. Copyright © 2011 Royal Meteorological Society
We present a new method of reducing the error in predicted wind speed, thus enabling better management of wind energy facilities. A numerical weather prediction model, COSMO, was used to produce 48 h forecast data every day in 2008 at horizontal resolutions of 10 and 3 km. A new adaptive statistical method was applied to the model output to improve the forecast skill. The method applied corrective weights to a set of forecasts generated using several post-processing methods. The weights were calculated based on the recent skill of the different forecasts. The resulting forecast data were compared with observed data, and skill scores were calculated to allow comparison between different post-processing methods. The total root mean square error performance of the composite forecast is superior to that of any of the individual methods. Copyright © 2010 John Wiley & Sons, Ltd.
This paper will describe a system which predicts the expected power output of a number of wind farms. The system is automatic and operates on-line. The paper will quantify the accuracy of the predictions and will also give examples of the performance for specific storm events. An actual implementation of the system will be described and the robustness demonstrated.
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.
In this article we have used the ARMA (autoregressive moving average process) and persistence models to predict the hourly average wind speed up to 10 h in advance. In order to adjust the time series to the ARMA models, it has been necessary to carry out their transformation and standardization, given the non-Gaussian nature of the hourly wind speed distribution and the non-stationary nature of its daily evolution. In order to avoid seasonality problems we have adjusted a different model to each calendar month. The study expands to five locations with different topographic characteristics and to nine years. It has been proven that the transformation and standardization of the original series allow the use of ARMA models and these behave significantly better in the forecast than the persistence model, especially in the longer-term forecasts. When the acceptable RMSE (root mean square error) in the forecast is limited to 1.5 m/s, the models are only valid in the short term.
This paper describes a model for prediction of the power produced by wind farms connected to the electrical grid. The time frame is from 0 to 36 h ahead. The goal is to develop a model that can be integrated in the dispatching system at a utility. The physical basis of the model is the predictions generated from forecasts from the high-resolution limited area model (HIRLAM) of the Danish Meteorological Institute. These predictions are then made specific for individual sites (wind farms) by applying a matrix generated by the submodels of WASP (Wind Atlas Application and Analysis Program). To verify the model one year's worth of data from 17 wind farms have been used. The farms are located in Denmark on the Zealand (14) and Bornholm (3) islands and are all controlled by the Danish utility ELKRAFT/SK Power.
Hyper-parameters estimation in regression Support Vector Machines (SVMr) is one of the main problems in the application of this type of algorithms to learning problems. This is a hot topic in which very recent approaches have shown very good results in different applications in fields such as bio-medicine, manufacturing, control, etc. Different evolutionary approaches have been tested to be hybridized with SVMr, though the most used are evolutionary approaches for continuous problems, such as evolutionary strategies or particle swarm optimization algorithms. In this paper we discuss the application of two different evolutionary computation techniques to tackle the hyper-parameters estimation problem in SVMrs. Specifically we test an Evolutionary Programming algorithm (EP) and a Particle Swarm Optimization approach (PSO). We focus the paper on the discussion of the application of the complete evolutionary-SVMr algorithm to a real problem of wind speed prediction in wind turbines of a Spanish wind farm.
Wind forecasts for wind power generation using Eta model
- L Lazi´lazi´c
- G Pejanovi´pejanovi´c
- M Andandˇandživkovi´andživkovi´c
Lazi´Lazi´c, L. and Pejanovi´Pejanovi´c, G. andandˇandŽivkovi´andŽivkovi´c, M., Wind forecasts for wind power generation using Eta model,
Renewable Energy, Vol.35, No.6, 2010, pp.1236-1243.
World wind energy half-year report
WWEA, World wind energy half-year report 2012, Tech. rep.World Wind Energy Association, 2012.
- S Salcedo-Sanz
- E G Ortiz-García
- A Portilla-Figueras
- L Prieto
- D Paredes
Salcedo-Sanz, S. and Ortiz-García, E. G. and Portilla-Figueras, A. and Prieto, L. and Paredes, D.Renewable
Energy, Vol.34, No.6, 2009, pp.1451-1457.
Genetic Programming Theory and Practice VII
- M E Kotanchek
- E Y Vladislavleva
- G F Smits
Kotanchek, M. E. and Vladislavleva, E. Y. and Smits, G. F., Genetic Programming Theory and Practice
VII, Springer US, 2010.
A Description of the Advanced Research WRF Version 2
- W C Skamarock
- J B Klemp
- J Dudhia
- O Gill
- D Barker
- D M Wang
- W Powers
Skamarock, W.C. and Klemp, J.B. and Dudhia, J. and O. Gill, D. and Barker, D. M. and Wang, W. and
Powers, J. G., A Description of the Advanced Research WRF Version 2, AVAILABLE FROM NCAR,
Vol.88, 2001, pp. 7-25.
Wind forecasts for wind power generation using Eta model
- L Lazić
- G Pejanović
- M Andživković
Lazić, L. and Pejanović, G. andŽivković, M., Wind forecasts for wind power generation using Eta model,
Renewable Energy, Vol.35, No.6, 2010, pp.1236-1243.