基于相似日搜索的改進LMD與ESN相結合的短期電力負荷預測模型
發(fā)布時間：2019-02-17 10:17
【摘要】：短期電力負荷容易受到自然因素及社會因素的影響,這使得負荷預測比較困難.為了提高短期負荷的預測精度,提出了基于相似日搜索的改進局部均值分解(ILMD)和回聲狀態(tài)網(wǎng)絡(ESN)相結合的短期電力負荷預測模型.首先用模糊聚類分析將與預測日最相似的多個日期篩選出來.然后把這些相似日的整點負荷數(shù)據(jù)按照時間先后排成一組數(shù)據(jù)序列,用改進的LMD進行分解,對分解出的各個分量分別建立一個ESN網(wǎng)絡,對每一個網(wǎng)絡分別訓練并進行預測.最后把每個網(wǎng)絡的預測結果累加起來就是最終的預測值.實驗證明此方法能有效提高預測精度.
[Abstract]:Short-term power load is easily affected by natural and social factors, which makes load forecasting more difficult. In order to improve the accuracy of short-term load forecasting, a short-term load forecasting model based on similarity day search based on improved local mean decomposition (ILMD) and echo state network (ESN) is proposed. Firstly, fuzzy clustering analysis is used to screen out the most similar dates to the predicted date. Then the load data of these similar days are arranged into a set of data sequence according to the time sequence, then decomposed by improved LMD, each component is decomposed into a ESN network, and each network is trained and predicted separately. Finally, the sum of the forecast results of each network is the final prediction value. Experiments show that this method can effectively improve the prediction accuracy.
【作者單位】： 河南省電力勘測設計院;鄭州大學電氣工程學院;河南省工程咨詢公司;
【基金】：河南省青年骨干教師項目(2015GGJS-148) 河南省產(chǎn)學研合作項目(152107000058) 河南省重點科技攻關項目(152102210036)
【分類號】：TP183;TM715
本文編號：2425064
[Abstract]:Short-term power load is easily affected by natural and social factors, which makes load forecasting more difficult. In order to improve the accuracy of short-term load forecasting, a short-term load forecasting model based on similarity day search based on improved local mean decomposition (ILMD) and echo state network (ESN) is proposed. Firstly, fuzzy clustering analysis is used to screen out the most similar dates to the predicted date. Then the load data of these similar days are arranged into a set of data sequence according to the time sequence, then decomposed by improved LMD, each component is decomposed into a ESN network, and each network is trained and predicted separately. Finally, the sum of the forecast results of each network is the final prediction value. Experiments show that this method can effectively improve the prediction accuracy.
【作者單位】： 河南省電力勘測設計院;鄭州大學電氣工程學院;河南省工程咨詢公司;
【基金】：河南省青年骨干教師項目(2015GGJS-148) 河南省產(chǎn)學研合作項目(152107000058) 河南省重點科技攻關項目(152102210036)
【分類號】：TP183;TM715
【相似文獻】
相關期刊論文 前2條
1 楊博;程振波;鄧志東;;面向匹配決策問題的漏整合神經(jīng)元稀疏ESN網(wǎng)絡[J];北京科技大學學報;2012年01期
2 ;[J];;年期
相關碩士學位論文 前3條
1 李莎莎;基于數(shù)據(jù)分解和ESN網(wǎng)絡的短期電力負荷預測模型[D];鄭州大學;2015年
2 陳穎;基于ESN的無線傳感網(wǎng)絡位置指紋定位方法研究[D];蘭州交通大學;2016年
3 于永兵;基于改進ESN的混沌時間序列預測方法的研究[D];遼寧科技大學;2012年
，本文編號：2425064
本文鏈接：http://www.sikaile.net/kejilunwen/zidonghuakongzhilunwen/2425064.html
最近更新
教材專著
