摘要
为了解决工业中动态流量测量困难的问题,引入软测量方法对动态流量进行测量.考虑BP算法建立软测量模型时收敛速度慢,易陷"局部极小"等不足,提出一种经遗传算法优化的BP网络进行软测量建模,用遗传算法先确定BP网络的网络结构和参数,将训练一定次数后得到的连接权值作为遗传计算的初始值,再用遗传算法确定BP网络的最优连接权值,最后把用BP算法训练得到的网络用于建模.文中对在燕山大学液压实验室采集的数据进行仿真,实验结果表明这种改进的建模方法在模型的训练速度和精度上有了较大的改善.
Dynamic flow plays a key role in monitoring the condition of hydraulic systems since it is a rich carrier of condition information. However, flow signals were usually difficult to be measured on-line due to the limitations such as high cost and time delay. For solving this problem, a novel soft sensor is presented. In order to overcome the disadvantages such as slow convergent speed and local minimum , the backpropagation (BP) was used to build the model of soft sensor. The GA algorithm was applied to optimize parameters and the structure of BPNN, thus tO improve the performance of the model. The data produced by collection were a- dopted for the sake of simulation. The result verifies the validity of this method.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2009年第1期242-244,共3页
Journal of Harbin Institute of Technology
基金
国家自然科学基金资助项目(60374042)
河北省自然科学基金资助项目(F2006000267)
关键词
动态流量
软测量
BP
遗传算法
dynamical flow
soft sensor
BP (back propagation)
genetic algorithm