Application of BP neura1 network in the gel point prediction of blend crude oil

作者：Hou L^*; Xu X Y; Liu X L

来源：Journal of Petrochemical Universities, 2009, 22(3): 86-88+98.

DOI：10.3696/j.issn.1006-396X.2009.03.021

• Blend crude oil
• BP neura1 network
• Gel point
• Prediction

The BP neura1 network and the tested gel points of the blend crude oil were studied. A prediction model based on the BP neura1 network was established to calculate the gel point of the blend crude oil. The back-propagation artificial neural network has the satisfactory nonlinear prediction ability. The gel points of the component crude oil and their proportions were taken as the input vector and the gel points of the blend crude oil as the output one. The relationship between the gel points of the blend crude oil and the component crude oil was researched. The results show that the gel point prediction model based on the BP neura1 network is better than other gel point prediction models. The maximum average deviation of the predicted gel point is only 0.128°C and the average absolute deviation is 0.022°C.

• 出版日期2009
• 单位中国石油大学（华东）