关键词: 渣油加氢, 连续集总模型, 轴向扩散模型, 信赖域优化算法

Abstract: With the crude oil getting heavier and more inferior continuously, residue hydrogenation process has gradually become the mainstream technology for the treatment of heavy oil. In the residue hydrogenation reaction process, impurities such as sulfur, nitrogen, nickel and vanadium are effectively removed from the residue, and the thick aromatic hydrocarbons of large molecules are hydrocracked into light small molecules, providing cleaner and more suitable raw materials for the subsequent process. However, due to the extremely complex molecular composition of residue and its hydrogenation reaction system, accurate simulation of residue hydrogenation reaction process is a key and difficult point, and is also a current research focus. In order to solve the problem of steady state simulation, a dynamic parameter model based on gamma distribution and a continuous lumped axial diffusion model are proposed in this paper. The complexity of the model is simplified and the generality of the model is improved by gamma distribution function. The axial diffusion model is used to simulate the flow mass transfer process of fixed bed residue hydrogenation, and the trust region optimization algorithm is used to correct the model parameters by using industrial data, which greatly improves the computational efficiency and convergence rate of the model correction compared with other optimization algorithms. In addition, two other residue hydrogenation process conditions were used to verify the model and the optimized model parameters, and the results showed that the prediction accuracy and prediction efficiency were high, indicating that the simplified mechanism model based on gamma distribution had high prediction accuracy, calculation efficiency and calculation potential.

Key words: residue hydrotreating, continuous lumped model, axial dispersion model, trust region optimization algorithm