Residual oil hydrogenation device yield prediction method based on machine learning

The invention discloses a method for predicting the yield of a residual oil hydrogenation device based on machine learning, which comprises the following steps of: starting from data driving, constructing a two-layer model by utilizing a machine learning algorithm, and constructing a model I for representing intrinsic dynamics by fully utilizing a large amount of small-scale experimental data obtained in the development process of a residual oil hydrogenation process; a model II representing reactor factors is constructed by considering the reactor amplification effect of a pilot plant or an industrial device; according to the method, complex reaction mechanism and reaction process simulation calculation are avoided, rapid calculation of reaction processes of residual oil hydrogenation devices of different scales can be achieved through the two-layer model, and the model is wide in application range, high in accuracy and high in practicability. 本发明公开一种基于机器学习的渣油加氢装置的收率预测方法，以数据驱动出发，利用机器学习算法构建两层模型，充分利用渣油加氢工艺开发过程