Font Size: a A A

Application Of BP Neural Network In Methane Chemical-Looping Reforming Reaction

Posted on:2020-06-01Degree:MasterType:Thesis
Country:ChinaCandidate:Y Y YanFull Text:PDF
GTID:2381330590957231Subject:Energy and Chemical Engineering
Abstract/Summary:PDF Full Text Request
Methane chemical-looping reforming reaction is one of the most valuable chemical utilization method for natural gas and the ceria-based oxides are its most promising oxygen carriers.In this paper,BP neural network and its improved neural networks were introduced to the simulation of chemical-looping reforming reaction which is catalyzed by ceria-based composite oxides,providing intelligent advices for system analysis,result prediction,process optimization and so on.Based on 419 groups experimental methane chemical-looping reforming reaction data,this paper established 7 BP neural network models to associate 9 input variables and 3 output variables.The 9 input variables were composite metal oxide species,molar ratio of ceria oxide,molar ratio of composite oxide,preparation method,calcination temperature,calcination time,reaction temperature,reaction time and cycle numbers,while the 3 output variables were CH4conversion,CO selectivity and H2 selectivity.The 7 BP neural network models included a normal BP neural network which was noted as BP-dlp and six improved BP neural networks which were named BP-dlp2,BP-dxtt,GA-BP,PSO-BP,CS-BP and DE-BP.And the PSO-BP neural network had the best performance.Using the PSO-BP neural network model,the experimental results under a certain condition were predicted and the minimum relative error between the predicted data and the actual data was 0.118%.Meanwhile,the preparation conditions of the composite oxygen carrieres and the reaction conditions of the fuel reactor when the reaction results were optimal could be found by the PSO-BP model,and the obtained preparation and reaction conditions were consistent with the actual experimental conditions of the methane chemical-looping reforming reaction.The best reaction results would be obtained when the preparation method was method NO.2,the composite metal oxide is ferric oxide,the molar ratio of ceria-iron composite oxides was0.7/0.3,the temperature and time of calcination was 800°C and 6 h,the reaction temperature and time was 850°C and 13 min,and the number of cycles was 0.
Keywords/Search Tags:Methane Chemical-Looping Reforming, Ceria-based oxides, BP neural network, Improved BP neural network
PDF Full Text Request
Related items