基于改进RBF网络的乙烯纯度软测量建模方法

A Soft-Sensing Modeling Approach to Ethylene Purity Based on Improved RBF Neural Network

对兰州某石化厂乙烯精馏塔产品质量进行了软测量建模研究,针对经典RBF神经网络建模时存在的问题,提出了一种改进的RBF网络结构,增加了从输入直接到输出的线性环节,应用最近邻聚类和最陡下降法训练改进后的网络,实验结果表明改进后的网络具有较好的网络性能,适用于乙烯精馏塔产品质量的软测量建模和预测。

To the problems in soft-sensing modeling of ethylene distillation column products with traditional RBF neural network, an improved RBF network structure was proposed, in which linear connections between input and output layers were introduced, and nearest neighbor-clustering algorithm and steepest descent algorithm were used to train the network. Simulation results show that this modified RBF network has much better performance, which can be used in the modeling and prediction of ethylene purity of ethylene distillation column.