The Research On Application Of Coupling Models And Modifying Algorithms Towards The Design Of Columns For The Separation Of Compounds

Column is a separation unit that is stable,effective and preferable in the separation engineering for compounds.Under some certain backgrounds regarding the industrial separation,numerous factors that are related to the column size,the packed mediums,the cost on the column,and so on,should be taken into consideration while designing column is our particular focus.Optimizing every column using experimental method,however,is not only labor-intensive,but also is time-consuming.Furthermore,in the applied research on the design of separation columns,the covered academic subjects mainly are chemistry,mass transfer science,separation science and the optimization of the capital cost on the feasible technology.Accordingly,both the intersection among these subjects and the strength of the industrial practice determined the importance of employing the powerful automation capability into the design of columns,especially in the era of large-scale computation.For carrying out the calculation regarding the column design,expressing the physiochemical process using the appropriate model is the first and foremost step.Subsequently,searching the reasonable algorithm to solve the complicated model about column process is for guaranteeing the success of the whole calculation.The innovative points in our thesis are represented by applying the coupling models and the modified algorithm to study how to separate three compounds（Taxifolin,Oridonin and Hydrogen Sulfide（H₂S））using columns.To isolate taxifolin from the extracts of the Larix gmelini sawdust,a framework used for predicting the equilibrium time（te）was established.The characteristic existed in this framework was that the physical chemistry models,which were the modified Langmuir model,the reformed second-order kinetic models and intra-particle diffusion model in which the initial adsorption factor（Ri）was defined,were coupled on the basis of some specific parameters.Before coupling these models,model identification was done through fitting the experimental data.Through the related verification experiments,the best numerical range of this framework was determined.This chapter described the physical and chemical mechanism on adsorbing taxifolin using AB-8 MARs in detail.Moreover,in the aspect of computation,the calculating scale of the above framework,although,is small,integrating the theoretical models in the form of flowchart with the input-output style can offer the future researchers,who will work on the large-scale optimization,the supporting information.A series of results of the desorption experiments that followed the statistic adsorption experiments,in which the operating time was set to be equal to the equilibrium time（te）calculated by the predict strategy,showed that the optimal eluting solvent was independent on the adsorption conditions.Then,the best dynamic conditions determined by analyzing all known adsorption equilibrium time（te）were found to be same as the correspondingly experimental results.Then,we concluded that this framework possesses the applied meaning.In our proposed technology about combing AB-8 MARs column and pre-HPLC to separate taxifolin,keeping the total amount of taxifolin existed in the sample solutions that would be injected into the pre-HPLC column equal to 40mg is an effective way to get taxifoin with the purity of 95%.Additionally,narrowing the time range of the sample collection in the pre-HPLC unit can make the final purity of taixifolin further improved to 98%.Herein,the scale-up correlation existed among the columns packed with the certain mediums was represented by choosing the operating conditions of the pre-HPLC according to the theoretical parameters of the analytical HPLC chromatograms.With the industrial requirement was taken into consideration,the pre-HPLC unit about separating taxifolin was replaced by crystallization.The so-called coupling model method was reflected by how to combine the selected separation units together.For designing every separation unit,the equations,in which the retention time of the reference compounds and the polarity parameter of the solvent were chosen as the independent variables,were established to calculate the best volume ratio of the eluting solvent in water.On the basis of the current design,AB-8 MARs column coupling crystallization improved the purity of taxifolin to over 98%.Meanwhile,the technology,which was consisted of HP-20 MARs column unit,two-phase extraction based on 3 kinds of solvents（ethyl acetate,methanol and water）and crystallization,was used for the separation of oridonin.Oridonin with the purity of 98%was the result of the above technology.In our thesis,studying how to build technology to separate nature compounds through taking advantage of the model coupling method and choosing the column as the center unit was represented in the separation industry in the first time.The separation characteristic of removing H₂S is different from the one of improving the purity of taxifolin or oridonin.In addition,this difference was reflected by whether the total number of all components existed in the original system can be accounted.In the design of the technology used for removing H₂S gas,the zeolite column was chosen as the only one separation unit.Taking the fact that the recovery of H₂S gas component can be kept equal to be larger than 90%at the same time of its purity was over 95%in pressure swing process by changing cycles and arranging the separation steps into account,we decided to choose the selection of the relatively appropriate zeolites from 199 kinds of zeolites as the main goal for removing H₂S.There were 3 steps,which are the ZEOMIC software computes based on geometry,the grand canonical Monte Carlo（GCEMC）simulation and the process simulation,included into the whole screening work.Moreover,the model coupling method was exactly shown by combing these 3 steps together.With the data generated by the geometric computation and the GCEMC simulation as the cut-off reference,the finally process screen was aided by the overall model.When the finite volume algorithm was employed to solve the overall model,the errors caused by setting the improper iteration step,which has relationship with the operating conditions,into the basically Ronge-Kutta algorithm encapsulated in the function ode 23s contained in the software Matlab7.0,was found.Using the optimization method as the basis,the mapping existed between the operating conditions of the pressure swing process and the purity and the recovery of H₂S,thus,was established for guaranteeing the success of the screen work.Several two-component gas systems,which were CO₂/H₂S,H₂S/CO₂,H₂S/CH₄ and H₂S/N₂,were studied here.At last,two separating configures related to the system containing H₂S and CO₂ were proposed.And the sequence of the adsorption ability of the regarding gas components on the appropriated zeolites was concluded as follows:CO₂>H₂S>CH₄>N₂.In this chapter,both model coupling method and algorithm modifying method with the extended meaning were represented.Referencing to the separation principal shown through varying the separation step for the application of the pressure swing process,we choose the break-through point,which was the sign indicating that the compound has started leaking from the MARs column,as the predicted index.If the traditional idea that the sample solutions should be stopped into the MARs column while the break-through point has been reached was ignored,it is no doubt that the number of the variables regarding the MARs column will be increased and the design of the MARs column will be more flexible.In this case,the applied meaning of the prediction of the break-through point will be highlighted.Both the water solutions containing taxifolin（98%purity）prepared by the pre-HPLC and the sample solutions containing the crude extracts of taxifolin were studied for obtaining the equilibrium data.Due to that both HPD-100 MARs and HP-20 MARs were chosen as the adsorption medium,all equilibrium data were divided into 2 groups.Combining a constrain condition and the average expression of multi-component Langmuir model was determined as the new method to fit every group of equilibrium data.Then,it was found that the competitive adsorption phenomenon happened among all components contained in the sample solutions should not be ignored.For saving the cost on the purchase of the standards of every compound,a simplified framework was built through deleting the part related to all unknown compounds from the multi-component isotherm equation and adding the transformed time into the original algorithm loaded into the multi-component framework in turn.The numerical variation caused by the deleted part was made up by choosing the appropriate transformed time at the same time of varying the numerical rule of the key parameters.The sensitivity analysis of the key parameters occurred in the sing-component framework served as the reference for studying the numerical rules of the key parameters in this simplified framework.Finally,the dynamic adsorption results about these two kinds of MARs column identified that applying the simplified framework to predict the break-through point of taxifolin contained in the sample solution was accurate.Furthermore,it can be concluded that lowing the industrial cost of the column by modifying algorithm is a promising way.In this thesis,the technologies used for preparing three different compounds with high purity were proposed.The research on application of coupling models and modifying algorithms towards the design of columns for the separation of compounds is helpful to accelerate the development of the separation engineering.The relatively calculated data will offer the future researchers the theoretical reference with numerical characteristic and guidance.