Simulation Of Delayed Coking Adiabatic Reaction Process With Reaction Kinetics Based On Structure Oriented Lumping

Optimizing delayed coking process based on molecular refining is an urgent requirement to improve the heavy oil utilization efficiency. In this paper, combined with the adiabatic reactor, a molecular scale reaction kinetics model of delayed coking process was built using structure oriented lumping (SOL) method to simulate the delayed coking process of residual oil. The corresponding software was programed.21 structural increments including hydrocarbon groups, heteroatom groups and metal groups were selected to build a molecule set with 2791 typical molecules derived from 55 seed molecules. Combining hydrocarbon composition information and average molecular structural parameters, the composition calculation model of residual oil based on structure oriented lumping using simulation annealing algorithm was established to calculate molecular composition matrix of residual oil. The comparison between prediction and experimental results of the macroscopic properties of residual oils showed that this matrix could describe the composition of residual oils quantitatively.According to the reaction mechanism of delayed coking,38 reaction rules including reactant selection rules and product generation rules were used to establish the delayed coking reaction network, which could be transformed into kinetic differential equations by Matlab programming. Combining the reaction rate constants and the reaction heats by quantum chemistry calculation, delayed coking reaction kinetics model was constructed based on structural oriented lumping method. The model could calculate the conversion rate in infinitesimal reaction time and temperature change of reaction system. Combining the adiabatic reactor model of delayed coking tower, the total molecular composition changes of delayed coking process was gained through loop computation. So the product distributions of delayed coking process could be predicted. Delayed coking SOL adiabatic reactor model could simulate the reaction process for different residual oils at different reaction temperatures and recycle ratios. The prediction errors of gas, gasoline, diesel, wax oil and coke yields were less than 1.6% and the effects of temperatures and recycle ratios were consistent with the experimental results. The prediction errors of individual molecules in products did not exceed 2%.