基于ANN方法预测柴油调和的凝点和冷滤点

Predicting Solidifying Point And Cold Filter Plugging Point of Blended Diesel Oil with ANN Method.

根据南京炼油厂所提供的柴油调和凝点和冷滤点基础数据,用人工神经网络(ANN)的反向传播BP算法对凝点和冷滤点进行预测。提出了适宜的人工神经网络拓扑结构,讨论了BP算法中学习速率、动量系数及过拟合现象对网络的影响,通过实验数据的检验,证明了用ANN方法建立的柴油调和模型能有效地给出预测信息。研究表明,ANN方法比常用的调和系数模型、凝点指数模型、凝点换算因子模型等更能准确地关联和预报调和柴油的凝点和冷滤点。

According to the date base of Solidifying Point(SP)and Cold Filter Plugging Point(CFPP) of the blending diesel oil that Institute of Nanjing Oil Refinery offered,the artificial neural network(ANN) was used for prediction SP and CFPP with back-propagation(BP) method. The appropriate topology of ANN was obtained.The learning rate,the momentum factor and overfitting phenomena in BP network were discussed. It was proved that ANN model can successively give the predicting information of blending diesel oil.The results showed that ANN method can predict SP and CFPP of blending diesel oil with much more accuracy than regression mathematics model such as blending coefficient model,index correlating model,SP conversion factor model.