(47f) Software Tools for Developing Molecular-Level Kinetic Models of Large, Complex Chemical Systems

Modeling energy and refinery processes can pose a challenge when the feed complexity or reactive intermediates require large numbers of reactions and/or species. Traditional lumping approaches cannot capture the detail in these systems and often overlook the impact of feed and process condition variations on the effluent. A C++ based kinetic modeling framework called Dynamic Model Builder (DMB) has been developed to dynamically model both large and small systems with molecular resolution. Advances in computer software and hardware have allowed for the development of the DMB framework to function as a standalone application on an average desktop computer. However, large scale models are cumbersome in each of the model building, solving, and parameter estimation phases, so the framework can also be compiled on a high-performance cluster to further reduce solution times. In conjunction with the systematic network building tool and composition modeling tool in the Kinetic Modeler’s Toolbox developed previously in the Klein research group, DMB was used to create a heavy oil hydroprocessing network with O(10³) species and O(10⁴) reactions. The model includes the detailed hydroprocessing chemistry (saturation, hydrocracking, ring opening, hydrodesulfurization, hydrodenitrogenation, and hydrodeoxygenation) and composition modeling based on experimental bulk properties. DMB options to model a catalytic, multi-reactor packed-bed system with side-by-side reaction and vapor-liquid equilibrium were used to make the kinetic model. The final optimized model represents a real refinery reactor simulation over a range of process conditions like space velocity, temperature, pressure, and hydrogen flow. The DMB model can then be run from a user-friendly application for easy simulation and visualized in the DMB user interface for network and reaction flux analysis to aid in the comprehension of these large molecular systems.