Keynote Talk: Long-Time Simulation of MTO Reactors with EMMS-Based Discrete Particle Method

The methanol to olefins (MTO) is an important method is to produce light olefins. However, the hydrodynamic behaviors and the reacting properties cause some difficulties in designing and optimizing the MTO reactor. In MTO reaction, the coke will deposit on the catalysts and the coke content will greatly affect the reacting behavior and the final products. However, the process of coke deposition is very slow and it usually takes hours of physical time for the catalysts to reach the optimal level of coke content. So, to accurately model the whole process of MTO reaction in a reactor, both the accuracy and efficiency of the simulating method are required. The discrete particle method (DPM) is adopted to accurately model the catalysts at the particle scale. A coarse-grained particle method is developed based on the energy-minimization multiscale (EMMS) model to reduce the computational cost of the particles. Furthermore, the computation of particles is accelerated with the graphics process unit (GPU), and the computational ability could be extended to hundreds of millions of computational particles. The simulation of 8 hours of physical time of a MTO reactor is achieved. The final productions are well accordance with the experimental results. The whole process of the coke deposition and the details of the movements of catalytic particles are obtained, which provides more information for scaling-up MTO reactors.