原油闪点预测模型的研究

针对原油这类结构组成复杂、差异性大、可燃的复杂混合体系,选取各地区共计101种原油的恩氏蒸馏温度、20℃密度、20℃粘度作为输入变量,建立原油闪点预测模型。采用主成分分析法对输入变量进行降维,除去恩氏蒸馏系列数据中的信息冗余,分别采用多元线性回归(MLR)、BP神经网络、RBF神经网络三种方法建模,并对模型的预测结果进行对比,RBF神经网络模型的预测准确度与稳定性均为最优,绝对误差期望为2.94℃,相对误差期望为3.45%,BP神经网络模型的准确性优于多元线性回归模型,稳定性不如MLR模型。

Research on flash point prediction model of crude oil

Aiming at the complex mixture system of crude oil with complex composition, great difference in properties and flammability, a flash point prediction model was established with the ASTM distillation temperature and the density and viscosity at 20℃ of 101 crude oils from different regions as input variables. Principal component analysis(PCA) is used to reduce the dimension of the input variables and eliminate the information redundancy in the series data of ASTM distillation. Three methods, multiple linear regression(MLR), BP neural network and RBF neural network, are used to model respectively, and the prediction results of the models are compared. The prediction accuracy and stability of the RBF neural network model are the best, with an absolute error expectation of 2.94℃, and a relative error expectation of 3.45%. The accuracy of BP neural network model is better than that of MLR model, and the stability is not as good as MLR model.