四川电力技术
四川電力技術
사천전력기술
SICHUAN ELECTRIC POWER TECHNOLOGY
2013年
5期
5-8
,共4页
周专%姚秀萍%王维庆%任华%申盛召
週專%姚秀萍%王維慶%任華%申盛召
주전%요수평%왕유경%임화%신성소
帝国主义的竞争算法-神经网络%数值天气预报%短期风功率预测%风电场
帝國主義的競爭算法-神經網絡%數值天氣預報%短期風功率預測%風電場
제국주의적경쟁산법-신경망락%수치천기예보%단기풍공솔예측%풍전장
imperialistic competitive algorithm -neural network%numerical weather prediction%short-term wind power pre-diction%wind farm
随着风电大规模的接入电网,风电对电网的影响越来越大。由于风电出力具有随机性、间歇性和不可控性,导致风电对电网调度运行带来巨大的挑战。为了充分利用风电,必须将风电由未知变为基本已知,提高对风电出力的预测精度。提出一种基于帝国主义竞争算法的神经网络( ICA-NN)方法来提高短期风功率预测的精度。在该方法中,首先,建立一个基于多层感知器( MLP)人工神经网络的风速预测模型,然后,用帝国主义竞争算法优化神经网络中的权值。将该预测方法应用于新疆某风电场,验证了该方法应用于短期风功率预测的有效性,证明了该方法可以提高短期风功率预测的精度。
隨著風電大規模的接入電網,風電對電網的影響越來越大。由于風電齣力具有隨機性、間歇性和不可控性,導緻風電對電網調度運行帶來巨大的挑戰。為瞭充分利用風電,必鬚將風電由未知變為基本已知,提高對風電齣力的預測精度。提齣一種基于帝國主義競爭算法的神經網絡( ICA-NN)方法來提高短期風功率預測的精度。在該方法中,首先,建立一箇基于多層感知器( MLP)人工神經網絡的風速預測模型,然後,用帝國主義競爭算法優化神經網絡中的權值。將該預測方法應用于新疆某風電場,驗證瞭該方法應用于短期風功率預測的有效性,證明瞭該方法可以提高短期風功率預測的精度。
수착풍전대규모적접입전망,풍전대전망적영향월래월대。유우풍전출력구유수궤성、간헐성화불가공성,도치풍전대전망조도운행대래거대적도전。위료충분이용풍전,필수장풍전유미지변위기본이지,제고대풍전출력적예측정도。제출일충기우제국주의경쟁산법적신경망락( ICA-NN)방법래제고단기풍공솔예측적정도。재해방법중,수선,건립일개기우다층감지기( MLP)인공신경망락적풍속예측모형,연후,용제국주의경쟁산법우화신경망락중적권치。장해예측방법응용우신강모풍전장,험증료해방법응용우단기풍공솔예측적유효성,증명료해방법가이제고단기풍공솔예측적정도。
With the large-scale access of wind power grid , the wind power has more and more influence on power grid .Be-cause the wind power output is random , intermittent and uncontrolled , it brings the huge challenge to power grid dispatching and operation .In order to make full use of wind power , the wind power must be changed from unknown to known , and the prediction accuracy of wind power output should be improved .A neural network method based on imperialist competitive algo-rithm (ICA -NN) is presented to improve the accuracy of short -term wind power prediction.In this method, first of all, a prediction model of wind speed is established based on artificial neural network with multi -layer perceptron ( MLP), and then, the weight values of neural network are optimized with the imperialist competitive algorithm .The prediction method has been applied to a wind farm in Xinjiang , which verifies its effectiveness in short -term wind power prediction , and proves that the proposed method can improve the accuracy of short -term wind power prediction .