基于人工神經(jīng)網(wǎng)絡(luò)的電力短期負(fù)荷預(yù)測(cè)系統(tǒng)研究.pdf

1、霹飄分類號(hào)一密級(jí)一——～U[】C艮礬理了犬學(xué)碩士學(xué)位論文基主厶王盟經(jīng)圓終毆電力縋塑魚煎亟型丞筮盟壅研究生姓名陳浩指導(dǎo)教師姓名、職稱——奎薹壟!熬塑!學(xué)科專業(yè)管理科學(xué)與工程研究方向厶三盤絲論文工作起止日期2004年6月3日論文提交日期2005年4月10E1AbstractNowadays，ourcourttryisbeinginaseriouselectricpowerscarceperiodTheelectricpowershortag

2、einnation—widescopenotonlYhasseriouslyaffectedtheindustrialproductionandbroughtgreatinconveniencetopeople‘slife，butalsohascausedaseriesofsocialproblemTosolvetheproblemofelectricpowershortage，wemustdeepenonthereformofelec

3、tricpowermarketcontinuouslyThus，itrequiresthePowerenterprisestoexcavatePotentialofthemselvesunderexistingconditions，reasonablymanagepowerresource，andreducethepecuniarylossbroughtbytheelectricpowershortageasbestasonecanTh

4、eshort—termelectricloadforecastisaveryimportantlinkintheelectricpowermanagement，resultedinagreatimprovementintheestablishmentoftheproductionplanofgenerators，thereasonabledeployforfireandwatergeneratee】ectricityandfuelthe

5、short。termmaintenanceofsafeanalysisequipmentsandthespreadoftheenergyofelectricalnetworkInadaythatinformativeconstructionareinvigorouSdevelopment，traditionalartificialforecastCannotalreadySatiSfytheneedsofelectriCpowerind

6、ustrialdevelopmentmoreandmoreHowever，theIlewforecastmethodofrepresentative，whichiSartificialneuralnetwork，showsitShugesuperiorityAfterstudyingplentyofdocumentsandanalyzingvariousimportantfactorsofelectricpowerload，theBPn

7、etworkwiththree—layerstructurehasbeenconstructed，andthenetworkwastrainedwithhistoricalelectricpoWerloaddataresuRedsncaassfulshoft—termelectriCpowerloadforecastAndtheprecisionhasrisentoahigherlevelcomparedwiththeresultfor

8、ecastedbytraditionaImethods，whichprovesthevalidityofartificialnervenetworkinthepowerloadforecastfieldofshortperiodAftertheartifieialnel1ralnetworkmodelofshortperiodpowerloadfecasthadbeen’established，andguidedbytheobjcot—