梯级水电站群优化调度的群体搜索算法

打开文本图片集

WANGKechao’，CAI Ruoyi，SHENJianjian1*

nstituteofHydopower&Hydroiformatics，DalianUiversityofcholog，Dan6O4，Chia;2.PearliverCopsie Center，PearlRiver WaterResourcesCommissionof the Ministryof WaterResources，Guangzhou51O611，China）

Abstract：Thecoronavirus herd immunityoptimizer（CHIO）algorithmwasintroduced toaddresstheoptimal scheduling problemof cascade hydropower plants，andanimproved algorithmcaledLCHO was proposed.The CHIOalgorithmwas enhancedbyadopting twostrategies：theLatinhypercubestrategyforgenerating initialpopulationsolutionsanddynamicsocialdistanceandadaptive thresholdstrategies.Theseimprovementsbalancedglobal searchandlocal search，preventing prematureconvergence while accelerating theconvergencespeed.SimulationresultsforfourcascadehydropowerplantsinaregioninsouthwestChinashowthatthe difference between the results of CHIO and LCHIO and the dynamic programming POA-DDDP is less than 1% ，and thevariation coefficients are 0.13% and 0.11% ，which verify the rationalityand effectivenessof CHIO to solve the cascade reservoir group schedulingproblemandtheimprovementefectofLCHOalgorithmonCHIO.Furtheranalysisofmultipleobjeetivefunctionmodels fortheequivalent gridrevealsisuessuchasthecoexistenceofpowershortagesandcurtailment，dificultiesinrenewableenergy absorption，andweakgenerationcapacityduringdryperiods.Thesechalengeshighlightthenedforthedevelopmentofpumped storage.

KeyWords：cascadehydropowerplantscheduling;coronavirusherdimmunityoptimizeralgorithm;Latinhypercubestrategy;socia distance;adaptive threshold

随着新能源的快速并网，其发电过程所呈现出的短期随机性与中长期波动性等特征愈发明显，大大增加了剩余负荷的峰谷差、加剧电力系统调峰压力[1-3]。（剩余14910字）