基于条件建模的聚合反应过程动态优化-研究

《基于条件建模的聚合反应过程动态优化-研究》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、浙江大学硕士学位论文温操作切换的回溯模拟及产品设计动态优化，验证了方法的可行性。关键词：自由基聚合；条件建模；动态优化；灵敏度分析；互补约束AbstractPolymerisachemicalcompoundormixtureofcompoundsconsistingofrepeatingstructuralunitscreatedthroughaprocessofpolymerization．Withintherangeofwholeconversion，Typicalfreeradicalpolymerizationshavemanycomplicatedphenom

2、enasuchasgeleffect,glasseffectandcageeffectseverelyinfluencesomekeyphysicalquantities(suchasviscosity,diffusioncoefficientsandreactionkineticcoefficients)，whichresultsinconditionalmodelingproblemsinpolymerizationprocess．Conditionalmodelbasedondifferential-algebraicequations(DAE)includesno

3、n-continuousgradients，causinghugedifficultyofsolvingdynamicoptimizationproblem．Thus，itmakesagreatpartofpolymerizationprocessthathowtosolvethisdynamicoptimizationproblem．Inthispaper,conditionalmodelingoffreeradicalpolymerizationsisf'wstanalyzedthoroughly，bothsequentialandsimultaneousdynami

4、coptimizationstrategyarepresentedtosolvetheproblem．Sequentialoptimizationalgorithmincludessensitivityanalysisandmodelsmoothingstrategy，whilesimultaneousapproachisdiscretizedbasedonDAEmodelandMPCCalgorithm．Themainresearchcontentsandcontributionsofthispaperareasfollows：1．Mechanismoffreeradi

5、calpolymerizationisfirststudiedindetail，andconditionalmodelingproblemcausedbythethreeeffectsisanalyzedconsequently．Withagoodunderstandingofthecauseandeffect，asummaryofthedifficultiesontraditionalwaystosolvetheproblemisproposed．2．Sequentialdynamicstrategybasedonsensitivityanalysisandmodels

6、moothingmethodisestablished．Sensitivityanalysisalgorithmcanimprovetheaccuracyofgradients，andmodelsmoothingcontributestoproblemsolving．：Basedonwhich，adynamicproblemofproductsdesigninfreeradicalpolymerizationprocessisdevelopedandsolvedwell．Abetterresultisobtainedthanthatwithtraditionalway．B

7、esides，theefficiencyofoptimizationsolvingisimproved．3．TheoriginalconditionalDAEmodelisreformulatedtoalarge-scalenonlinearprogramming(NLP)probleminsimultaneousapproachafterdiscretization．Withcomplementarityconstraintsstrategyandmodelreformulation，MPCCalgorithmisfocus