Page - 318 - in Short-Term Load Forecasting by Artificial Intelligent Technologies

energies Article Short-TermLoadForecastingforElectricBus ChargingStationsBasedonFuzzyClusteringand LeastSquaresSupportVectorMachineOptimized byWolfPackAlgorithm XingZhang DepartmentofEconomicManagement,NorthChinaElectricPowerUniversity,Baoding071003,China, 51851719@ncepu.edu.cn Received: 22May2018;Accepted: 1 June2018;Published: 4 June2018 Abstract: Accurate short-term load forecasting is ofmomentous significance to ensure safe and economic operation of quick-change electric bus (e-bus) charging stations. In order to improve the accuracy and stability of loadprediction, this paperproposes ahybridmodel that combines fuzzy clustering (FC), least squares support vectormachine (LSSVM), andwolf pack algorithm (WPA).On thebasis of loadcharacteristics analysis for e-bus charging stations, FC is adopted to extract samplesonsimilardays,whichcannotonlyavoid theblindnessof selectingsimilardays byexperience,butcanalsoovercometheadverseeffectsofunconventional loaddatacausedbya suddenchangeof factorsontraining. Then,WPAwithgoodglobalconvergenceandcomputational robustness isemployedtooptimize theparametersofLSSVM.Thus,anovelhybrid loadforecasting model for quick-change e-bus charging stations is built, namely FC-WPA-LSSVM.Toverify the developedmodel, twocasestudiesareusedformodelconstructionandtesting. Thesimulationtest resultsprove that theproposedmodelcanobtainhighpredictionaccuracyandideal stability. Keywords: short-termloadforecasting;electricbuschargingstation; fuzzyclustering; least squares supportvectormachine;wolfpackalgorithm 1. Introduction In recentyears, low-carboncitieshavebecomeacommonpursuit aroundtheworld,which is facedwith increasingenergycrises andenvironmentalproblems [1]. Electricbuses (e-buses)have developedquicklywith theburgeoningconstructionof low-carboncities [2].As important supporting facilities, charging stationsbringnewchallenges tooptimaldispatchingandsafe operationof the powergriddue togreatvolatility, randomnessand intermittenceof the load [3]. Therefore, it is of great significance toconduct researchonloadcharacteristicsanalysisandshort-termloadforecasting. Ononehand, this contributes to theoptimalcombinationofgeneratorunits in termsofpowersystem, economical dispatch, optimal powerflowandelectricitymarket transactions. On the other hand, it provides adecisionbasis for constructionplanning, energymanagement, orderly charging and economical operation for charging stations, which can guarantee and promote the development of low-carbon cities. Therefore, research on short-term load forecasting for quick-change e-bus chargingstationshasbeenconductedtoprovidedatasupportandatheoreticalbasis for the large-scale constructionofchargingstations. Nowadays, scholars have conducted a large amount of research on load forecasting for charging stations. The predictionmethods are primarily divided into two categories: traditional forecasting approaches, such as time series [4], regression analysis [5], and fuzzy prediction [6], andartificially intelligentalgorithms.Conventionalpredictionmethodsaimingat load forecasting for e-bus charging stations aremainly established on the foundation of probability and statistics Energies2018,11, 1449;doi:10.3390/en11061449 www.mdpi.com/journal/energies318