CAS OpenIR  > 研究所(批量导入)
神经网络用于流化床反应器的故障诊断
俞江萍
Subtype硕士
Thesis Advisor罗保林
1998-06-01
Degree Grantor中国科学院研究生院
Abstract本文针对丙烯腈合成流化床反应器,采用人工神经网络模型和流化床数学模型相结合的方法,来进行流化床反应的故障诊断。首先基于丙烯氨氧化反应动力学实验结果以及反应动力学行为分析,建立了丙烯腈合成流化床反应器的数学模型。编制了具有友好用户界面的Windows 程序,同时对网络的结构和训练参数进行了讨论。提出了判断局部最小的规则,采用遗传算法来跳出局部最小。训练后的网络回响结果远胜于非线性拟合,对未经训练的样本预测结果很精确。在程序诊断功能块,加入流化床反应器数学模型作为专家系统,用来指导诊断过程,并据此反馈调节网络的输入,直到完成反应器的调优操作。对含有故障的样本进行诊断,结果表明,神经网络能根据数学模型导出的经验规则,准确辩认故障类型,并作出相应的正确调节。
Other AbstractIn the present work, artificial neural networks (ANN) combined with mathematical model of reactor is used to diagnose operation faults of fluidized bed reactor for synthesizing acrylonitrile. First, regressing experimental and industrial operating data of ammoxidation of propylene and analyzing kinetic behaviour, a mathematical model of fluidized bed reactor for synthesizing acrylonitrile was proposed. The program, which runs in Windows 95 system, is designed using Visual C++ language. The structure and training parameters of ANN were discussed. and the criterion of local minimum was proposed, and genetic algorithm (GA) is adopted to overstep local minimum. Results of reechoing and predicting are better than the results of nonlinear regression. The mathematical model of fluidized bed reactor is used as expert system to guide diagnosing faults and implement feedback-adjusting input of ANN until outputs of ANN become regular and optimum adjusting is finished. The diagnosis of faults sample indicts ANN can recognize types of fault and make corresponding adjusting properly.
Pages75
Language中文
Document Type学位论文
Identifierhttp://ir.ipe.ac.cn/handle/122111/8317
Collection研究所(批量导入)
Recommended Citation
GB/T 7714
俞江萍. 神经网络用于流化床反应器的故障诊断[D]. 中国科学院研究生院,1998.
Files in This Item:
File Name/Size DocType Version Access License
神经网络用于流化床反应器的故障诊断.pd(5523KB) 限制开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[俞江萍]'s Articles
Baidu academic
Similar articles in Baidu academic
[俞江萍]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[俞江萍]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.