| Hydrogen energy as one of the cleanest energies has been widely used in many fields such as industry,medicine and so on.Hydrogen energy can also be used in fuel cell as a source of power for new energy vehicles,which is one of the important clean renewable energy in the future.The hydrogen production,transportation,storage and purification restrict the large-scale application,especially the hydrogen purification,not only closely related to other procedures,but also has an important significance and impact on the safety,efficiency,reliability and sustainable use of the whole hydrogen energy system.Pressure swing adsorption(PSA)as an important gas separation technology can remove impurities from hydrogen mixtures.In this work,the heat and mass transfer models for multicomponent mixture were established based on the mass conservations,energy conservations,adsorption isotherm equations and mass transfer equations.The characteristics of PSA for hydrogen purification were also studied.In this work,the adsorption isotherms of single-component(H2,CH4,CO,N2,CO2)were validated on activated carbon,zeolite 5A and UTSA-16 respectively.The calculation results were then compared with the experimental results of the extended Langmuir-Freundlich model and Dual-site Langmuir model.After validated the adsorption isotherm models,the breakthrough curves of multicomponent mixture were established and validated.Furthermore,the PSA cycle models were built,the parametric study were discussed to investigate the effect of parameters on the purification performance of PSA(hydrogen purity,recovery and productivity).In order to study the effect of the number of adsorbent layers on the purification performance of PSA cycle,the single-layered bed and double-layered bed were firstly studied.Then,a new kind of adsorbent of metal organic framework(MOF),UTSA-16 were introduced to build a three-layered adsorbent bed with activated carbon and zeolite 5A.The effects of adsorption time,P/F ratio and adsorption pressure on the purification performance were studied.Different optimization strategies were discussed to optimize the purification performance of PSA cycle.By using the Box-Behnken Design(BBD)response surface optimization method for experimental design and analysis.The PSA cycle parameters of adsorption time,pressure equalization time and P/F ratio as the factors in the BBD method,and the purification performance of purity,recovery and productivity as three responses,the quadratic regression equations can obtain for the responses of the system.Different optimization conditions and results were obtained using the BBD method.The three-layer back propagation(BP)neural network,consisted of an input layer,a hidden layer and an output layer,was constructed to optimize the PSA cycle process.The adsorption time,P/F ratio and adsorption pressure as the input parameters,the hydrogen purity,recovery and productivity as the output parameters,the performance of BP neural network was validated to be able to apply to the prediction and analysis of PSA cycle.Then the genetic algorithm was used to optimize the BP neural network model,the results of testing error and training error obtained are compared with the traditional BP neural network model.As the results show,the BP neural network optimized by genetic algorithm has better generalization performance.According to the interior point optimization algorithm,under certain constraints,the objective function is simplified and optimized by linear weighting method,which shows that the BP neural network optimized by genetic algorithm is feasible for the optimization of PSA cycle process.In this dissertation,the effects of hydrogen mixture component,PSA operation parameters and adsorption bed structure on hydrogen purification performance were studied.Based on different optimization strategies,the optimization design criteria of PSA system for hydrogen purification were proposed,which is of great theoretical significance and practical application value for improving PSA purification performance. |
PDF Full Text Request