改进BP神经网络预测Ni/Al2O3催化CH4-CO2重整反应

doi:10.16085/j.issn.1000-6613.2016-2126

化工进展 ›› 2017, Vol. 36 ›› Issue (07): 2393-2399.DOI: 10.16085/j.issn.1000-6613.2016-2126

改进BP神经网络预测Ni/Al₂O₃催化CH₄-CO₂重整反应

付柯, 谢良才, 闫雨瑗, 李波, 贺改, 徐龙, 马晓迅

陕北能源先进化工利用技术教育部工程研究中心, 陕西省洁净煤转化工程技术研究中心, 西安市能源高效清洁化工利用工程实验室, 西北大学化工学院, 陕西西安 710069

收稿日期:2016-11-17 修回日期:2017-01-05 出版日期:2017-07-05 发布日期:2017-07-05
通讯作者: 徐龙,教授。
作者简介:付柯(1991-),男,硕士研究生,研究方向为甲烷重整。E-mail:fuke916@163.com。
基金资助:
国家自然科学基金重点项目（21536009）及西安市科技计划项目（CXY1511（4））。

Predicting model of CH₄-CO₂ reforming on Ni/Al₂O₃ catalyst by improved back propagation（BP）neural network

FU Ke, XIE Liangcai, YAN Yuyuan, LI Bo, HE Gai, XU Long, MA Xiaoxun

Chemical Engineering Research Center of the Ministry of Education for Advanced Use Technology of Shanbei Energy, Shaanxi Research Center of Engineering Technology for Clean Coal Conversion, Xi'an Engineering Laboratory for Energy Efficient and Clean Chemical Utilization, College of Chemical Engineering, Northwest University, Xi'an 710069, Shaanxi, China

Received:2016-11-17 Revised:2017-01-05 Online:2017-07-05 Published:2017-07-05

摘要/Abstract

摘要： CH₄-CO₂重整制合成气的反应可以实现CO₂的减排和C₁资源的高效利用。这一反应受反应温度、原料气比例、催化剂种类等诸多因素影响，如果考察每种因素影响，势必大大增加实验的工作量。人工神经网络以其超强容错性、并行处理、可学习和自适应等优点，在非线性预测方面具有明显的优势。本文基于Ni/Al₂O₃催化CH₄-CO₂重整反应过程，采用改进的BP神经网络建立了关于CH₄转化率、CO₂转化率和H₂/CO比的预测模型，结果表明，采用改进BP神经网络在此研究中具有更快的收敛速度和更好的网络稳定性，其收敛次数仅为BP神经网络的58.86%；改进BP神经网络模型的敏感度分析表明，输入因素对反应结果影响的顺序为：反应温度 > Ni负载量 > 平均孔径≈比表面积≈孔体积。

关键词: CH₄-CO₂催化重整, BP神经网络, 模拟, 优化, 预测

Abstract: CH₄-CO₂ reforming reaction can produce synthesis gas,which is an ideal way both for the reduction of CO₂ emission and the efficient utilization of C₁ resources.This reaction is affected by many factors,such as reaction temperature,ratio of raw material gas,catalyst type and so on.If each of factors were investigated,it would greatly increase the workload of the experiment.Artificial neural network (ANN) has obvious advantages in nonlinear prediction because of its superior fault tolerance,parallel processing and adaptive learning.The prediction model about CH₄-CO₂ reforming reaction catalyzed by Ni/Al₂O₃ was built based on artificial neural network.This model was trained by back propagation (BP) algorithm and improved BP algorithm,respectively.It was found that the improved BP model was much better than the BP model in view of the stability and convergence speed.Compared with the BP algorithm,the improved BP algorithm reduced the number of convergence times greatly,which was only 58.86% of that in BP model.By sensitivity analysis of the models,it showed that the reaction temperature was the most important factor on the reaction indexes (CH₄ conversion,CO₂ conversion,and H₂/CO ratio) among five input factors,followed by Ni loading.In addition,the average pore size,the specific surface area,and the pore volume had relatively small effects on reaction indexes within the experimental range.

Key words: CH₄-CO₂ catalytic reforming, BP neural networks, simulation, optimization, prediction

中图分类号:

TE646

付柯, 谢良才, 闫雨瑗, 李波, 贺改, 徐龙, 马晓迅. 改进BP神经网络预测Ni/Al₂O₃催化CH₄-CO₂重整反应[J]. 化工进展, 2017, 36(07): 2393-2399.

FU Ke, XIE Liangcai, YAN Yuyuan, LI Bo, HE Gai, XU Long, MA Xiaoxun. Predicting model of CH₄-CO₂ reforming on Ni/Al₂O₃ catalyst by improved back propagation（BP）neural network[J]. Chemical Industry and Engineering Progress, 2017, 36(07): 2393-2399.

参考文献

[1] ZHANG X H,HU R J,LIU Y F,et al.The performance of Co/ZrO₂-Al₂O₃ composite oxide catalyst for CH₄/CO₂ reforming reaction[J]. International Journal of Hydrogen Energy,2015,40(32):10026-10032.
[2] SONG C S. Global challenges and strategies for control conversion and utilization of CO₂ for sustainable development involving energy,catalysis,adsorption and chemical processing[J]. Catal.Today,2006,115(1-4):2-32.
[3] ASHCROFT A T,CHEETHAM A K,GREEN L H,et al. Partial oxidation of methane to synthesis gas-using carbon-dioxide[J]. Nature,1991,352:225-226.
[4] LIU Y F,HE Z H,ZHOU L,et al. Simultaneous oxidative conversion and CO₂ reforming of methane to syngas over Ni/vermiculite catalysts[J]. Catalysis Communications,2013,42:40-44.
[5] SUN N N,WEN X. Catalytic performance and characterization of Ni-CaO-ZrO₂ catalysts for dry reforming of methane[J]. Applied Surface Science,2011,257(21):9169-9176.
[6] QI J Z,SUN Y P,XIE Z L,et al. Development of Cu foam-based Ni catalyst for solar thermal reforming of methane with carbon dioxide[J]. Journal of Energy Chemistry,2015,24(6):786-793.
[7] SUN N N,WEN X,WANG F,et al. Effect of pore structure on Ni catalyst for CO₂ reforming of CH₄[J]. Energy Environ.Sci.,2010,3:366-369.
[8] XU L L,SONG H,CHOU L J. Carbon dioxide reforming of methane over ordered mesoporous NiO-Al₂O₃ composite oxides[J]. Catal.Sci.Technol.,2011(1):1032-1042.
[9] LUAN A E C,IRIARTE M E. Carbon dioxide reforming of methane over a metalmodified Ni-Al₂O₃ catalyst[J].Appl. Catal,A:Gen.,2008,343(1/2):10-15.
[10] LI Y,HE D,ZHU Q,et al. Effects of redox properties and acid-base properties on isosynthesis over ZrO₂-based catalysts[J]. Catal.,2004, 221(2):584-593.
[11] 何玉彬,李新忠. 神经网络控制技术及其应用[M]. 北京:科学出版社,2000:2-35. HE Y B, LI X Z. Shen jing wang luo kong zhi ji shu ji qi ying yong[M]. Beijing:Science Press,2000:2-35.
[12] 张敬玲. BP神经网络的应用[J]. 石家庄技术学院报,2015,27(4):34-35. ZHANG J L. Application of BP neural network[J]. Journal of Shijiazhuang Vocational Technology Institute,2015,27(4):34-35.
[13] 王文新,潘立登,李荣,等. 常减压蒸馏装置双模型结构RBF神经网络建模及其应用[J]. 北京化工大学学报,2004,31(4):91-94. ` WANG W X,PAN L D,LI R,et al. Development of RBF neural network with double model structure and its application to atmospheric and vacuum distillation units[J]. Journal of Beijing University of Chemical Technology,2004,31(4):91-94.
[14] JUAN G S,JOSE D,MARTIN G,et al. Neural networks for analyzing the relevance of input variables in the prediction of tropospheric ozone concentration[J]. Atmospheric Environment,2006,40:6173-6180.
[15] 苏鑫,裴华健,吴迎亚,等. 应用经遗传算法优化的BP神经网络预测催化裂化装置焦炭产率[J]. 化工进展,2016,35(2):389-396. ` SU X,PEI H J,WU Y Y,et al. Predicting coke yield of FCC unit using genetic algorithm optimized BP neural network[J]. Chemical Industry and Engineering Progress,2016,35(2):389-396.
[16] 李晨,张海涛,应卫勇,等. 钴基催化剂F-T合成的人工神经网络模拟[J]. 计算机与应用化学,2006,23(10):963-966. LI C,ZHANG H T,YING W Y,et al. Aritncial neural network simulation of cobalt-based catalyst F-T synthesis[J]. Computers and Applied Chemistry,2006,23(10):963-966.
[17] 黄玮,丛玉凤,郭大鹏. 基于BP神经网络的石蜡催化氧化反应的研究[J]. 石油化工高等学校学报,2012,25(6):30-38. HANG W,CONG Y F,GUO D P. Catalytic oxidized reaction of paraffin wax based on BP neural network[J]. Journal of Feterochemcial Universities,2012,25(6):30-38.
[18] XU L L,ZHAO H H,SONG H L. Ordered mesoporous alumina supported nickel based catalysts for carbon dioxide reforming of methane[J]. Hydrogen Energy,2012,37(9):7497-7511.
[19] CHEN X Y,CHAU K W,BUSARI A O. A comparative study of population-based optimization algorithms for downstream river flow forecasting by a hybrid neural network model[J]. Engineering Applications of Artificial Intelligence,2015,46:258-268.
[20] BARCENAS O P,OLIVAS E S,MARTIN J D. Unbiased sensitivity analysis and pruning techniques[J]. Ecoogical Modelling,2005,182(2):149-158.
[21] MURIEL G,IOANNIS D,SOVAN L. Review and comparison of methods to study the contribution of variables in artificial neural networks models[J]. Ecological Modelling,2003,160(3):249-264.
[22] SOVAN L,MARITXU G,GIRAUDEL J L. Predicting stream nitrogen concentration from watershed features using neural networks[J]. Water Research,1999,33(16):3469-3478.
[23] SZECOWKA P M,SZCZUREK A,LICZNERSKI B W. On reliability of neural network sensitivity analysis applied for sensor array optimization[J]. Sensors and Actuators B,2011,157:298-303.
[24] NIKHIL R P. Soft computing for feature analysis[J]. Fuzzy Sets and Systems,1999,103:201-221.
[25] PODOLAK I. Feedforward neural network's sensitivity to input data[J]. Computer Physics Conunication,1999,117:181-188.
[26] PARK Y S,JORGE R,SOVAN L. Sensitivity analysis and stability patterns of two-species pest models using artificial neural networks[J]. Ecological Modelling,2007,204:427-438.

改进BP神经网络预测Ni/Al₂O₃催化CH₄-CO₂重整反应

Predicting model of CH₄-CO₂ reforming on Ni/Al₂O₃ catalyst by improved back propagation（BP）neural network

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	李梦圆, 郭凡, 李群生. 聚乙烯醇生产中回收工段第三、第四精馏塔的模拟与优化[J]. 化工进展, 2023, 42(S1): 113-123.
[2]	张瑞杰, 刘志林, 王俊文, 张玮, 韩德求, 李婷, 邹雄. 水冷式复叠制冷系统的在线动态模拟与优化[J]. 化工进展, 2023, 42(S1): 124-132.
[3]	王太, 苏硕, 李晟瑞, 马小龙, 刘春涛. 交流电场中贴壁气泡的动力学行为[J]. 化工进展, 2023, 42(S1): 133-141.
[4]	孙继鹏, 韩靖, 唐杨超, 闫汉博, 张杰瑶, 肖苹, 吴峰. 硫黄湿法成型过程数值模拟与操作参数优化[J]. 化工进展, 2023, 42(S1): 189-196.
[5]	王福安. 300kt/a环氧丙烷工艺反应器降耗减排分析[J]. 化工进展, 2023, 42(S1): 213-218.
[6]	陈匡胤, 李蕊兰, 童杨, 沈建华. 质子交换膜燃料电池气体扩散层结构与设计研究进展[J]. 化工进展, 2023, 42(S1): 246-259.
[7]	崔守成, 徐洪波, 彭楠. 两种MOFs材料用于O₂/He吸附分离的模拟分析[J]. 化工进展, 2023, 42(S1): 382-390.
[8]	王正坤, 黎四芳. 双子表面活性剂癸炔二醇的绿色合成[J]. 化工进展, 2023, 42(S1): 400-410.
[9]	徐若思, 谭蔚. C形管池沸腾两相流流场模拟与流固耦合分析[J]. 化工进展, 2023, 42(S1): 47-55.
[10]	张凤岐, 崔成东, 鲍学伟, 朱炜玄, 董宏光. 胺液吸收-分步解吸脱硫工艺的设计与评价[J]. 化工进展, 2023, 42(S1): 518-528.
[11]	孙玉玉, 蔡鑫磊, 汤吉海, 黄晶晶, 黄益平, 刘杰. 反应精馏合成甲基丙烯酸甲酯工艺优化及节能[J]. 化工进展, 2023, 42(S1): 56-63.
[12]	郭强, 赵文凯, 肖永厚. 增强流体扰动强化变压吸附甲硫醚/氮气分离的数值模拟[J]. 化工进展, 2023, 42(S1): 64-72.
[13]	邵博识, 谭宏博. 锯齿波纹板对挥发性有机物低温脱除过程强化模拟分析[J]. 化工进展, 2023, 42(S1): 84-93.
[14]	许友好, 王维, 鲁波娜, 徐惠, 何鸣元. 中国炼油创新技术MIP的开发策略及启示[J]. 化工进展, 2023, 42(9): 4465-4470.
[15]	李春利, 韩晓光, 刘加朋, 王亚涛, 王晨希, 王洪海, 彭胜. 填料塔液体分布器的研究进展[J]. 化工进展, 2023, 42(9): 4479-4495.