資源描述：
《基于半?yún)?shù)分析的電力需求預(yù)測(cè)算法.pdf》由會(huì)員上傳分享，免費(fèi)在線(xiàn)閱讀，更多相關(guān)內(nèi)容在行業(yè)資料-天天文庫(kù)。
1、第16卷第8期中國(guó)安全科學(xué)學(xué)報(bào)Vol.16No.82006年8月ChinaSafetyScienceJournalAug.2006基于半?yún)?shù)分析的電力需求預(yù)測(cè)算法1,2211趙林峰工程師韓國(guó)燕李勇王印紅(1中國(guó)礦業(yè)大學(xué)(北京)管理學(xué)院,北京1000832國(guó)電長(zhǎng)源沙市熱電廠(chǎng),荊州434001)學(xué)科分類(lèi)與代碼:620.20中圖分類(lèi)號(hào):X924.4文獻(xiàn)標(biāo)識(shí)碼:B摘要市場(chǎng)經(jīng)濟(jì)對(duì)電力需求預(yù)測(cè)提出了更高的要求。在簡(jiǎn)要介紹和評(píng)論以往國(guó)內(nèi)外常用電力需求預(yù)測(cè)方法優(yōu)缺點(diǎn)基礎(chǔ)上,為了提高電力需求預(yù)測(cè)的準(zhǔn)確度及
2、預(yù)測(cè)方法的適應(yīng)性,建立半?yún)?shù)回歸模型預(yù)測(cè)電力需求。半?yún)?shù)回歸模型分為線(xiàn)性和非線(xiàn)性?xún)刹糠?。線(xiàn)性部分反映了負(fù)荷預(yù)測(cè)可知的部分規(guī)律,非線(xiàn)性部分反映了負(fù)荷預(yù)測(cè)的不確定因素的影響。用偏殘差方法估計(jì)半?yún)?shù)回歸模型,估計(jì)結(jié)果由兩步得到。首先估計(jì)線(xiàn)性參數(shù)部分,然后再估計(jì)非線(xiàn)性參數(shù)部分。通過(guò)示例計(jì)算,半?yún)?shù)回歸模型對(duì)電力需求預(yù)測(cè)的估計(jì)誤差數(shù)值小于二元線(xiàn)性回歸法對(duì)電力需求預(yù)測(cè)的估計(jì)誤差。結(jié)果表明:利用半?yún)?shù)回歸法預(yù)測(cè)電力需求是一種預(yù)測(cè)精度高、計(jì)算容易、普適性強(qiáng)的算法。關(guān)鍵詞電力需求;預(yù)測(cè);半?yún)?shù);回歸;核函數(shù);窗寬PredictiveCalculationTechniquef
3、orElectricPowerDemandBasedonAnalysisofSemiparametricRegression1,2211ZHAOLinfeng,EngineerHANGuoyanLIYongWANGYinhong(1SchoolofManagement,ChinaUniversityofMining&Technology(Beijing),Beijing100083,China2ThermalelectricPlant,ChangyuanShashiCorporationofChinaPowerGroup,Jingzhou434001,C
4、hina)Abstract:Marketeconomyhighlyrequiresmoreprecisepredictionofelectricpowerdemand.Theadvantageanddisadvantageofpredictionmodelsprevioususedbothathomeandabroadwerebrieflyintroduced.Inordertomakethepredictionmorepreciseandfeasible,asemiparametricmodelforpredictingtheelectricpowerdeman
5、dcomposedoflinearparameterandnonlinearparameterwasestablished.Linearparameterreflectedtheunderstandableregularityofloadpredictionwhilethenonlinearparameterreflectedtheimpactofuncertainfactorsoftheloadprediction.Thesemiparametricmodelwasestimatedbythemethodofpartialresidualdeviation.E
6、stimationwasobtainedbytwosteps.Thepartoflinearparameterwasestimatedfirstfollowedbytheestimationofthepartofnonlinearparameter.Theexemplifiedcalculationshowedthattheestimationdeviationbysemiparametricmodelregressionisrathersmall,andevensmallerthanthatobtainedfrombinarylinearregression.
7、Theresultshowedthatthismethodismoreprecise,andeasytocalculatewithgoodadaptability.Keywords:electricdemand;prediction;semiparametric;regression;corefunction;band文章編號(hào):1003-3033(2006)08-0017-05;收稿日期:2006-03-31;修稿日期:2006-06-27中國(guó)安全科學(xué)學(xué)報(bào)第16卷!18!ChinaSa
