資源描述:
《基于小波分析和GM-ARIMA模型的月度售電量預(yù)測.pdf》由會員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在行業(yè)資料-天天文庫。
1、第42卷第4期華北電力大學(xué)學(xué)報Vo1.42.No.42015年7月JournalofNoahChinaElectricPowerUniversityJu1.,2015doi:10.3969/j.ISSN.1007—2691.2015.04.17基:f:dx波分析和GM—ARIMA模型的月度售電量預(yù)測樊嬌,馮昊,牛東曉,王筱雨,劉福炎(1.華北電力大學(xué)經(jīng)濟(jì)與管理學(xué)院,北京102206;2.國網(wǎng)浙江省電力公司經(jīng)濟(jì)技術(shù)研究院,浙江杭州310008)摘要:月度售電量直接反映電力企業(yè)的經(jīng)營效益,準(zhǔn)確的電量預(yù)測對于電力企業(yè)合理
2、安排購售電方案、確定融資缺口具有重要意義。鑒于各地區(qū)月度售電量時間序列不但有隨時間逐漸增長的趨勢,還受節(jié)假日、氣溫等因素的影響存在隨機(jī)項(xiàng),為了提高電力企業(yè)月度售電量的預(yù)測精度,采用小波分析理論將月度售電量時間序列分解為近似序列和細(xì)節(jié)序列,并通過對分解后子序列的特征進(jìn)行分析,分別采用相匹配的GM(1,1)模型和ARIMA模型對子序列進(jìn)行預(yù)測,然后通過序列重構(gòu)得到月度售電量的預(yù)測值。經(jīng)實(shí)際算例驗(yàn)證,該組合預(yù)測方法的平均誤差率為3.7%,與神經(jīng)網(wǎng)絡(luò)等常用單一預(yù)測方法相比能明顯提高預(yù)測精度,具有較強(qiáng)的適應(yīng)能力。關(guān)鍵詞:小波
3、分析;分解序列;灰色模型;ARIMA中圖分類號:TM614文獻(xiàn)標(biāo)識碼:A文章編號:1007—2691(2015)04—0101—05MonthlyElectricitySalesForecastBasedonWaveletAnalysisandGM.ARIMAModelFANJiao,F(xiàn)ENGHao,NIUDongxiao,WANGXiaoyu,LIUFuyan(1.SchoolofEconomicsandManagement,NoahChinaElectricPowerUniversity,Beijing1022
4、06,China;2.EconomicResearchInstituteofZhejiangElectricPowerCompany,StateGrid,Hangzhou310008,China)Abstract:Monthlyelectricitysalescanreflectthebusinessprofitofpowerplantsdirectly.Thustheaccurateforecastofelectricitysalescanhelpthoseplantsmakereasonablearrangem
5、entoftheelectricitypurchasingschemeanddeter—minethefinancinggap.Thetimeseriesofmonthlyelectricitysalesnotonlytendstograduallyincreasewithtime,butalsocomesundertheinfluenceofthefactorsliketemperatureandholidays.InordertOimprovethepredictionaccuracyofthemonthlys
6、aleofpowerplants,thispaperemployedthetheoryofwaveletanalysistodecomposethetimeseriesofmonthlyelectricitysalesintoapproximationsequenceanddetailsequences.Afteranalyzingthecharacteristicsofde—compositionsequencesrespectively,matchingmethods——GM(1,1)andARIMAmodel
7、,wereemployedtopredictthedecomposedsubsequences.Thenthroughthereconstructionofsubsequences,thefinalpredictionofelectricitymonthlysaleswiththeaverageerrorrateof3.7%wereobtained.Finally.itwasverifiedbyexamplesthatthecombi—nationforecastingmethodcouldobviouslyimp
8、rovethepredictionaccuracy,withastrongabilitytoadapt.Keywords:waveletanalysis;decompositionsequence;greymodel;ARIMA建立與發(fā)展,電力企業(yè)運(yùn)營的經(jīng)濟(jì)性成為重要指0引言標(biāo)。對于電力企業(yè)而言,主要的現(xiàn)金流入來源于售電收入,在我國目前實(shí)行電價管制的情況下,