| 柴油调合人工神经网络模型研究 |
| |
| 引用本文： | 李为民,华贲.柴油调合人工神经网络模型研究[J].炼油技术与工程,2001,31(4):44-46. |
| |
| 作者姓名： | 李为民 华贲 |
| |
| 作者单位： | 1. 江苏石油化工学院
2. 华南理工大学 |
| |
| 基金项目： | 教育部强化传热与过程节能重点实验室研究基金资助!项目编号 :5 2 10 83 0 |
| |
| 摘 要： | 根据某炼油厂所提供的柴油调合凝点和冷滤点基础数据 ,用人工神经网络的反向传播方法进行建模预测。提出了人工神经网络适宜的拓扑结构 ,通过生产数据的检验 ,证明了用该方法建立的柴油调合模型能有效地给出预测信息。研究表明 ,凝点预测绝对值平均误差为 1.0 5℃ ,冷滤点预测绝对值平均误差为 0 .91℃ ,较常用的调合系数模型、指数关联模型和凝点换算因子模型 ,能更准确地预报调合柴油的凝点和冷滤点。
|
| 关 键 词： | 人工智能 神经网络控制 柴油 调合 凝点 冷滤点 降凝剂 |
| 修稿时间： | 2000年12月20 |
STUDY ON DIESEL BLENDING MODELS BASED ON ARTIFICIAL NEURAL NETWORK METHOD |
| |
| Li Weimin,Hua Ben.STUDY ON DIESEL BLENDING MODELS BASED ON ARTIFICIAL NEURAL NETWORK METHOD[J].Petroleum Refinery Engineering,2001,31(4):44-46. |
| |
| Authors: | Li Weimin Hua Ben |
| |
| Abstract: | Based on the data of solidifying point(SP) and cold filter plugging point(CFPP) of blending diesel oil from a refinery,diesel blending models were established with back-propagation method of an artificial neural network(ANN).An appropriate topological structure of ANN was put forward.It was proved that the predicted information about blending diesel oil can be successfully obtained.The average errors of predicted SP and CFPP are 1.05℃ and 0.91℃ respectively.It was shown that the ANN model can predict both SP and CFPP more accurately than the blending coefficient model,index correlation model and the SP conversion factor model. |
| |
| Keywords: | artificial intelligence neural network control diesel blending solidifying point cold filter plugging point pour point dipressant |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
