Abstract
Wind power generation increases rapidly. The availablewind energy depends on the wind speed, which is a randomvariable. For the wind-farm operator, this poses difficulty in thesystem scheduling and energy dispatching, as the schedule of thewind-power availability is not known in advance. In this paper,we propose an intelligent technique for forecasting wind speed ofwind turbine. This technique is based on artificial neuralnetwork (ANN). The Back propagation (BP) neural network isthen supplied with the data to establish the relationship betweenthe inputs and the output. The model based on the neuralnetwork demonstrated a good agreement and produced the windforecast with the accuracy of 90% and above.Index Terms-- Wind power generation, Wind speed, Windturbine, Artificial neural network.
I. INTRODUCTION
The energy is a vital input for the social and economicdevelopment of any nation. With increasing agriculturaland industrial activities in the country, the demand forenergy is also increasing. Formulation of an energy modelwill help in the proper allocation of widely availablerenewable energy sources as solar, wind, bio-energy andhydropower in meeting future energy needs. Wind energy isone of the best resources contributed to solution of globalwarming because of it is completely pollution free. Due toUncertainty in Nature, the wind production, this mainlydepends on the wind speed, is frequently changing. Therefore,it will be more beneficial for the producers to forecast therenewable power production according to some crucialparameters. It is important for the power industry to have thecapability to estimate this changing power. The characteristicsof wind power generation are studied and a neural network isused to estimate it. There are a few techniques available forwind speed prediction, which require a minimum number ofinput parameters. Four different statistical techniques, viz.,curve fitting, Auto Regressive Integrated Moving AverageModel (ARIMA), extrapolation with periodic function andArtificial Neural Networks (ANN) are employed to predictwind speed. These methods require wind speeds of previoushours as input. It has been found that wind speed can bepredicted with a reasonable degree of accuracy using twomethods, viz., extrapolation using periodic curve fitting andANN The paper describes an overview on ANNs in wind powersystems with prediction of wind speed. In this paper, thetheoretical background is reviewed in Part-2. TheMathematical Model artificial neural network is in Part-3,where different mathematical model of ANN are discussedand their performance summarized. At last the discussionabout forecasting results and the conclusions of this paper ispresented in the Part-4 & Part-5
Why Neural Network
The power generated by electric wind turbines changesrapidly because of the continuous fluctuation of wind speedand direction. Wind speed is non-linear fluctuation. Soforecasting is very difficult in normal method. As thetechnique of solving a nonlinear problem there is a methodusing the intelligent engineering represented by a neuralnetwork, a genetic algorithm, a chaos fractal, etc? Thesetechniques are already adopted as numerical prediction,prediction of the weather, etc., and the practicality is proved.So we predict wind speed to adopt artificial neural network(ANN), which shows a high performance about patternrecognition and the prediction problem of each field, and asaddition, determinate generation output from wind speed touse the characteristic curve of a wind generator. Because itwas impossible to get the data of output generated windturbine, which almost belong to the power company,generation output is predicted. In contradiction to this, it iseasy to buy wind speed data from meteorological office. ANNlearns the correlation with the desired output made into theinformation and ideal of an input, and If a strange input is putinto the network of the result, the approximation solutioncorresponding to it to will be calculated. When carrying outwind speed prediction using ANN.
II. OVERVIEW OF ARTIFICIAL NEURAL NETWORKS
Artificial neural networks (ANN) have been developed tomatch the powerful thinking, remembering, and problemsolvingcapabilities of the human brain. Artificial neuralnetworking is made up of simple highly inter connectedprocessing units called neurons each of which perform twofunctions: aggregation of its inputs from the externalenvironment and generation of an output from the aggregatedinputs.


Download full report
http://googleurl?sa=t&source=web&cd=2&ve...348160.pdf%3Farnumber%3D5348160&ei=4irBTZiYNMrJrAeJyLzgAw&usg=AFQjCNH46zhAYqkkT6zmbtQcRC2AaMzhMg&sig2=6f4enXAz8B78S7BP50z_kQ