Research On Smart Grid Load Forecasting Platform Based On Cloud Computing

Along with the deve lop ment o f Smart Grid, a large number of inte lligenttermina l equip me nts are used in power grid. The environme nt, the mea ns, and thetargets o f load forecast ing will been greatly c hanged. The load forecasting s yste mshave o utsta nd ing heteroge neous prob le ms. Forecasting mode ls cannot contro l thedivers ity and comp le xity, intermittent o f the load. Forecasting process needs toweigh the rationa lity o f the mathe matica l model and the real-time of the calc ulatio nfreq uently. And servers running are at low utilizatio n. But load forecasting is highlyrequired by S mart Grid Forecasting Load (SGFL). The de ve lopme nt o f loadforecasting will enter a new sta ge, inc lud ing mass data ma nage me nt o f loadforecasting, the e merge nce of fluctuating loads, and increas ingly higher de mandabout the dens ity and accurac y of load forecasting and so on, whic h ha ve importa ntimpact on impro ving the le ve l o f load forecasting. With the emergence o f c loudcomputing, power c loud generates. Clo ud comp uting will be ab le to realize the massdata processing, and achie ve reasonable reso urce allocatio n. Clo ud comp uting hasbeen wide ly used in med ical, educatio n, and e lectronic co mmerce, etc., and alsosuccessfully used in power syste m. There fore, clo ud co mputing will be adopted inSGLF in the near future.Thro ugh a na lyzing the deve lop ment of load forecasting domestic a nd abroad,the ma in proble ms o f load forecasting in S mart Grid, and applicatio n o f c loudcomputing in power s yste m, this paper forms the methods o f research o n fra meworkof SGLF p latform based on c lo ud co mputing. The ma in co ntents of this paper are asfo llows:Firstly, this paper ana lyzes the re latio ns hip between S mart Grid and loadforecasting. Through ana lyzing the shortco ming o f load forecasting and the newdemand o f S mart Grid, this paper puts forward the ma in characteristics, the ma infra mework, and the ma in forecasting tec hnolo gies o f SGLF.Second ly, this paper introduces the basic concept of c loud comp uting,productio n and dep loyme nt mode ls o f power c loud. It adopts virtua lized C loudStackand KVM to build private c loud about load forecasting a nd exp la ins load forecastingclo ud mecha nis m. This paper stud ies the resource scheduling in clo ud environme nt.A series o f optimizatio n a lgorithms can be used under the private c loud o f load forecasting.Third ly, according to the defects a nd s hortcomings o f trad itio na l loadforecasting s yste ms, this paper adopts the structure of Hadoop to build p latform o fclo ud comp uting a nd introduces key techno logies, s uc h as Vis ua lizatio n Techno logy(VT), paralle l pro gra mming model tec hno lo gy, mass data distrib utio n, storagetechno lo gy, mass data ma na ge ment techno lo gy, Service-Orie nted Arc hitecture (SOA)fra mework a nd so on. It arra nges c lo ud pla tform o f SGLF, forms its majorfra mework.Fina lly, this paper presents S mart Grid Load Services (SGLS) based onfundame nta l load ana lys is service, load forecasting as a service based on loadana lys is, and load ma nage me nt as a service about De mand Side Mana ge ment (DSM),multis tage coordination optimizatio n as a service. It is proved multis tage coordinatio noptimizatio n as a service is practica l by bas ic coordination mode l.Thro ugh mak ing research o n Fra mework of S mart Grid Load ForecastingPlatform about Clo ud Comp uting, this paper will improve the le ve l o f loadforecasting, propose a clear d irection for research about SGLF, and deepen theapp lication o f c loud co mputing in power syste m. It is certainly that the research la ysa solid foundation for S mart Grid dispatc hing work, and provides a powerfulguarantee for stability ana lys is o f power grid.