论文部分内容阅读
基于ANN(artificialneuralnetwork)的分馏塔侧线质量指标动态在线检测系统 ,简称DMS (dynamicon -linemonitoringsystem) ,成功地应用于炼油厂粗汽油干点、柴油凝点、Reid蒸汽压、闪点、冰点等的实时、在线、动态检测 ,达到很高的预测精度 ,极大地改善了分馏塔的控制性能 ,实现了分馏塔的操作优化 .本文提出一种新的网络训练方法———在线训练法 ,提高了在线检测系统的预测精度 .实际应用表明 ,这种方法可以有效克服人工神经网络离线训练的不足 ,提高动态在线检测系统的可靠性 ,并且省时省力
Based on Artificial Neural Network (ANN), dynamic online detection system of distillate lateral quality index (DMS), referred to as DMS (Dynamicon-line monitoring system), has been successfully applied to the real time of crude oil drying point, diesel pour point, Reid vapor pressure, flash point, , Online and dynamic detection, to achieve a high prediction accuracy, which greatly improves the control performance of the fractionation tower and optimizes the operation of the fractionation tower.This paper proposes a new network training method --- online training method to improve online Which can effectively overcome the shortage of offline training of artificial neural network, improve the reliability of dynamic on-line inspection system, and save time and effort