Page 77 - 2023年第54卷第6期
P. 77

[18] YINJN,TSAIFTC,LU CH.Bi - objectiveextraction - injectionoptimizationmodelingforsaltwaterintrusion
                       controlconsideringsurrogatemodeluncertainty [J].WaterResourcesManagement,2022,36(15):6017 - 6042.
                [19] NANTC,WU JC.Random WalkPathsolutiontogroundwaterflowdynamicsinhighlyheterogeneousaquifers
                       [J].JournalofHydrology,2018,563:543 - 559.
                [20] NANTC,LIKX,WUJC,etal.Assessmentofgroundwaterexploitationinanaquiferviarandomwalkongrid
                       (WOG)method:acasestudyatOrdos,China[J].HydrogeologyJournal,2018,26:1669 - 1681.
                [21] ASHERM J, CROKE B FW, JAKEMAN A J, etal.A review ofsurrogatemodelsandtheirapplicationto
                       groundwatermodeling[J].WaterResourcesResearch,2015,51:5957 - 5973.
                [22] LANGEVINCD,SHOEMAKERW B,GUO W.MODFLOW- 2000,theUSGeologicalSurveyModularGround -
                       WaterModel - DocumentationoftheSEAWAT - 2000VersionwiththeVariable - DensityFlowProcess (VDF)andthe
                       IntegratedMT3DMSTransportProcess (IMT)[R].Open - FileReport2003 - 426,USGeologicalSurvey,2003.
                [23] 李小兰,曾献奎,王栋,等.基于优化- 自适应稀疏网格替代模型的地下水模拟参数不确定性分析[J].吉
                       林大学学报(地球科学版),2022,52(4):1234 - 1243.
                [24] LUC,CAOH,MAJ,etal.Aproof - of - conceptstudyofusingalesspermeableslicealongtheshorelinetoin
                       creasefreshgroundwaterstorageofoceanicislands :Analyticalandexperimentalvalidation[J].WaterResources
                       Research ,2019,55:6450 - 6463.
                [25] ZHAOZX,HAOYH,WANGW K,etal.Ageneralizedanalyticalsolutionofgroundwaterheadresponsetodu
                       altideinamulti - layeredislandleakyaquifersystem [J].HydrologicProcesses,2022,36:e14548.
                [26] DEUTSCH CV,JOURNELAG.GSLIBGeostatisticalSoftwareLibraryandUser’sGuide.[M].secondedition
                       Oxford :OxfordUniversityPress,1998.
                [27] GUOW,LANGEVINCD.User’sguidetoSEAWAT:AComputerProgramforSimulationofThree - DimensionalVariable -
                       DensityGround - WaterFlow[R].TechniquesofWater - ResourcesInvestigations6 - A7,USGeologicalSurvey,2002.
                [28] 顾文龙,卢文喜,张宇,等.基于贝叶斯推理与改进的 MCMC方法反演地下水污染源释放历史[J].水利
                       学报,2016,47(6):772 - 778.
                [29] 张将伟,卢文喜,曲延光,等.基于 MonteCarlo方法的地表水地下水耦合模拟模型不确定分析[J].水利
                       学报,2018,49(10):1254 - 1264.




                  Random walk - basedphysics - informedsurrogatemodelingforgroundwaterflowinislands

                                                                    1,2
                                                                                  4
                                                                                               1,2,3
                                    1,2
                                                 2,3
                       NANTongchao ,PANShuo ,ZHANGJiangjiang ,WUJichun,LUChunhui
                   (1.StateKeyLaboratoryofHydrology - WaterResourcesandHydraulicEngineering,HohaiUniversity,Nanjing 210098,China;
                             2.YangtzeInstituteofConservationandDevelopment,HohaiUniversity,Nanjing 210098,China;
                          3.CollegeofWaterConservancyandHydropowerEngineering,HohaiUniversity,Nanjing 210098,China;
                               4.SchoolofEarthSciencesandEngineering,NanjingUniversity,Nanjing 210023,China)
                  Abstract:Groundwaterisanimportantfreshwatersourceinislands.Toprotectandutilizeislandgroundwateref
                  fectively ,itiscrucialtounderstandtheeffectsofspatialvariabilityofaquifers,precipitationrecharge,tidalfluc
                  tuations ,pumpingandotherfactorsonthegroundwaterdynamicsinislands.Inthispaper,asurrogatemodeling
                  approachviarandomwalkalgorithmforgroundwaterflowinislandsbasedongoverningequationsisproposed ,and
                  surrogatemodelsforartificialenhancedaquifers ,stratifiedleakyaquifers,andrandom heterogeneousaquifersin
                  islandsareestablishedsuccessivelybyrandomwalkalgorithmwithGPUparallelization.Thesesurrogatemodelsare
                  comparedwithcorrespondinganalyticalornumericalsolutionstoverifytheirapplicabilityandreliabilityindifferent
                  islandenvironments.Theeffectsofthetidaldynamicboundaryonthestratifiedaquifersoftheislandareanalyzed.
                  Besides ,theeffectsofthenumberofrandomwalkersontheerrorsandcomputationalcostsinbuildingsurrogates
                  areevaluatedbeforeconcludingthepaper.
                  Keywords:island;aquifer;tides;randomwalk;surrogatemodel;physics - informed

                                                                                    (责任编辑:于福亮)


                                                                                                —  7 0 5 —
   72   73   74   75   76   77   78   79   80   81   82