|本期目录/Table of Contents|

 混沌粒子群优化的RBF神经网络在热油管道仿真中的应用(PDF)

《纳米技术与精密工程》[ISSN:1672-6030/CN:12-1351/O3]

期数:
2017年3期
页码:
181-186
栏目:
精密测量
出版日期:
2017-05-15

文章信息/Info

Title:
 Hot Oil Pipeline Simulation Based on Chaotic Particle Swarm
Optimized RBF Neural Network
文章编号:
1672-6030(2017)03-0181-06
作者:
 李树杉1 张宇1 周明12 靳世久1
 (1.精密测试技术及仪器国家重点实验室(天津大学),天津 300072;
2.中国石油化工股份有限公司管道储运分公司,徐州 221008)
Author(s):
 Li Shushan1 Zhang Yu1 Zhou Ming12 Jin Shijiu1
 (1. State Key Laboratory of Precision Measuring Technology and Instruments(Tianjin University), Tianjin 300072, China;
2. SINOPEC Pipeline Transport & Storage Company, Xuzhou 221008, China)
关键词:
 热油管道仿真混沌粒子群算法RBF神经网络预测模型
Keywords:
 hot oil pipeline simulation chaotic particle swarm optimization algorithm RBF neural network forecast model
分类号:
TP183
DOI:
10.13494/j.npe.20150127
文献标识码:
A
摘要:
 针对加热输油管道仿真中管道内油流温度和压力变化,本文基于神经网络算法,利用管道数据采集与监控(SCADA)系统获取的500组历史运行数据,建立了管道沿线温度和压力的预测模型.提出的预测模型采用混沌粒子群改进的RBF(CPSO-RBF)神经网络算法.对RBF神经网络的参数(中心和宽度)、连接权重进行优化,通过与其他方法对比可知提出的CPSO-RBF预测模型具有精度高、收敛快等特点.在日照仪征热油管道实际运行方案中验证了提出的CPSO-RBF预测模型的可行性.
Abstract:
Regarding the variation of temperature and pressure in hot oil pipeline simulation, a forecast model of temperature and pressure along the pipeline based on neural network is established using 500 sets of historical operation data from supervisory control and data acquisition (SCADA) system. The forecast model is proposed based on chaotic particle swarm optimized radial basis function (CPSO-RBF) neural network algorithm in this paper, through which parameters (the center and the width) and connection weight are optimized. Compared with other algorithms, CPSO-RBF based forecast model shows higher precision and faster convergence speed. Its application in RizhaoYizheng hot oil pipeline verifies the feasibility of CPSO-RBF based forecast model.

参考文献/References

备注/Memo

备注/Memo:
收稿日期: 2016-01-22.
基金项目: 天津市应用基础与前沿技术研究计划(青年项目)资助项目(14JCQNJC04800);天津市科技支撑计划资助项目(14ZCZDGX00003).
作者简介: 李树杉(1989—),男,硕士研究生,592023279@qq.com.
通讯作者: 张宇,讲师,zhangyutju@gmail.com.
更新日期/Last Update: 2017-07-11