[1]肖立川,薛国新.BP 神经网络算法研究及其在燃煤锅炉中应用[J].常州大学学报(自然科学版),2001,(03):13-16.
　XIAO Li-chuan,XUE Guo -xin.A Study on the Algor ithm of Neural Networks and i ts Application in a Coal -burning Boiler[J].Journal of Changzhou University(Natural Science Edition),2001,(03):13-16.
点击复制

BP 神经网络算法研究及其在燃煤锅炉中应用()
分享到：

常州大学学报(自然科学版)[ISSN:2095-0411/CN:32-1822/N]

卷:
期数:
2001年03期
页码:
13-16
栏目:
出版日期:
2001-09-25

文章信息/Info

Title:
A Study on the Algor ithm of Neural Networks and i ts Application in a Coal -burning Boiler
作者:
肖立川;  薛国新
江苏石油化工学院机械工程系, 江苏 常州 213016
Author(s):
XIAO Li-chuan ;  XUE Guo -xin
Department of Mechanical Engineering , Jiangsu Institute of Petrochemical Technology , Changzhou 213016 , China
关键词:
BP 神经网络; 算法; 燃煤锅炉
Keywords:
BP neural netw orks ; algo rithm ; coal -burning boiler
分类号:
T P 393
文献标志码:
A
摘要:
神经网络的种种特性使得它们在动力工程中得到广泛的应用。根据系统的特点研究使用合理的神经网络算法已成为广受 关注的课题。BP 神经网络是一种最基本的神经网络。国际国内学者围绕如何避免训练陷入局部极小和加速收敛, 提出了多种 措施。但以往方法主要基于如何选择合理的学习率或加入适当的动量项, 未能充用利用控制系统的特点。通过分析具有不同极 点分布的控制系统BP 神经网络的误差曲面形状, 提出在总体误差的下降方向和其相反方向上同时进行一维搜索的方法。算法 较多地考虑了有关的细节性问题。此方法被用于某燃煤锅炉, 结
Abstract:
Neural networks have found wide application in power engineering because of their special functions. The use of reasonable algorithms of neural networks for the control of a sy stem acco rding to its characteristics has been a subject for study at tracting wide attention .A BP neural network is one of the basic type .Researchers at home and abroad have proposed a lot of methods to overcome the difficulty of local minimum point s encountered in the training process of a BP neural netw ork and to accelerate the training speed .These methods are based on the choice of a bet ter learning rate o r appending of a more appropriate momentum term .They didn′t take advantage of the characteristics of control systems .Analy zing the appearances of error cambers of the BP neural netw orks of control systems with different limit points distributions , this paper present s an improved algori thm which performs o ne dimension searching in the descent di rection and the opposite direction at the same time . The improved algorithm takes much consideration on related details .The application to a coal -burning boiler shows a great success.

参考文献/References:

[1] Rohani S , Haeri M , Wood H C .Modeling and Control of a Continuous Crystallization Process :Part 1 .Linear and Non -linear Modeling [J] .Comput ers & Chemical Engineering , 1999 , 23 : 263 -277 .
[2] Voge T , Wangi s I .Accelerating and Convergence of the Back Propagat ion M ethod [J] .Biol Cyberneti cs , 1998 , 59 :257 - 261 .
[3] Baldi P , Hornik K .Neural Netw orks and Principal C omponent Analysis :Learning f rom Examples Without Local Minimum [J] . Neural Netw orks , 1989 , 2 :53 -58 .

相似文献/References:

[1]周诗岽,王树立,赵会军.BP 神经网络预测管道运行费用研究[J].常州大学学报(自然科学版),2006,(02):30.
　ZHOU Shi- dong,WANG Shu- li,ZHAO Hui- jun.Investigation on Pipeline Transportation Expense Forecast[J].Journal of Changzhou University(Natural Science Edition),2006,(03):30.
[2]王　珂,许　波.一个反对称矩阵乘法的快速算法[J].常州大学学报(自然科学版),2001,(02):52.
　WANG Ke,XU Bo.A Fast Algor ithm of Anti -symmetric Matrix Multiplication[J].Journal of Changzhou University(Natural Science Edition),2001,(03):52.
[3]王其红,史国栋,薛国新,等.基于BP 神经网络的化工过程建模研究[J].常州大学学报(自然科学版),2000,(02):45.
　WANG Qi -hong,SHI Guo -dong,XUE Guo -x in,et al.A Study on the Modeling of Chemical Processes with BP Neural Networks[J].Journal of Changzhou University(Natural Science Edition),2000,(03):45.

备注/Memo

备注/Memo:
基金项目:江苏省科技厅2000 年度国际合作基金项目(编号BS2000730) 作者简介:肖立川(1946 -), 男, 江苏靖江人, 教授, 硕士, 主要从事热能工程方面的研究。
更新日期/Last Update: 2001-09-25