Thesis - Simulation and Optimization of Wind Farm Operations

《Thesis - Simulation and Optimization of Wind Farm Operations》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、SIMULATIONANDOPTIMIZATIONOFWINDFARMOPERATIONSUNDERSTOCHASTICCONDITIONSADissertationbyEUNSHINBYONSubmittedtotheOceofGraduateStudiesofTexasA&MUniversityinpartialfulllmentoftherequirementsforthedegreeofDOCTOROFPHILOSOPHYMay2010MajorSubject:IndustrialEngineeringSim

2、ulationandOptimizationofWindFarmOperationsunderStochasticConditionsCopyright2010EunshinByonSIMULATIONANDOPTIMIZATIONOFWINDFARMOPERATIONSUNDERSTOCHASTICCONDITIONSADissertationbyEUNSHINBYONSubmittedtotheOceofGraduateStudiesofTexasA&MUniversityinpartialfulllmentof

3、therequirementsforthedegreeofDOCTOROFPHILOSOPHYApprovedby:ChairofCommittee,YuDingCommitteeMembers,NatarajanGautamLewisNtaimoChananSinghHeadofDepartment,BrettA.PetersMay2010MajorSubject:IndustrialEngineeringiiiABSTRACTSimulationandOptimizationofWindFarmOperationsu

4、nderStochasticConditions.(May2010)EunshinByon,B.S.,KoreanAdvancedScienceandTechnology;M.S.,KoreanAdvancedScienceandTechnologyChairofAdvisoryCommittee:Dr.YuDingThisdissertationdevelopsanewmethodologyandassociatedsolutiontoolstoachieveoptimaloperationsandmaintenanc

5、estrategiesforwindturbines,helpingreduceoperationalcostsandenhancethemarketabilityofwindgeneration.Theintegratedframeworkproposedincludestwooptimizationmodelsforenablingdecisionsupportcapability,andonediscreteevent-basedsimulationmodelthatcharacterizesthedynamico

6、perationsofwindpowersystems.TheproblemsintheoptimizationmodelsareformulatedasapartiallyobservedMarkovdecisionprocesstodetermineanoptimalactionbasedonawindturbine'shealthstatusandthestochasticweatherconditions.Therstoptimizationmodeluseshomogeneousparameterswitha

7、nassumptionofstationaryweathercharacteristicsoverthedecisionhorizon.Wederiveasetofclosed-formexpressionsfortheoptimalpolicyandexplorethepolicy'smonotonicity.Thesecondmodelallowstime-varyingweatherconditionsandotherpracticalaspects.Consequently,theresultingstrateg

8、yareseason-dependent.Themodelissolvedusingabackwarddynamicprogrammingmethod.Thebenetsoftheoptimalpolicyarehighlightedviaacasestudythatisbaseduponelddatafromt