Multiphysics and multiscale modeling for the intensification of chemical engineering processes

点击翻译按钮获取中文摘要

Process intensification (PI) is a strategy aimed at making striking improvements in yield and cost effectiveness by developing innovative equipments and technologies to reach full-scale process performance at the pilot scale. For most unit operations, this is achieved by greatly enhancing mass and heat transfer. The time-to-market of intensified processes must be reduced. Experimental approaches are expensive and do not provide enough insights into velocity, concentration, and temperature profiles to increase the application of PI. Simulations that use multiphysics computational fluid dynamics models have been pinpointed by researchers as the key technology that needs to be leveraged to unlock the potential of PI of unit operations. High-order computational fluid dynamics (CFD) is a relatively new simulation approach in which the accuracy of the models can be controlled not only by altering the mesh but also by increasing the order of the scheme. High-order CFD has the capacity to provide the same accuracy as traditional methods with, however, a much shorter computational time (10x). With its low numerical dissipation, it is suited for the prediction of the turbulent and transitional flows that occur in intensified processes by large eddy simulation (LES). However, the use of high-order CFD has thus far been confined to the aeronautics industry. This is due to a lack of available commercial or accessible open source software to leverage the capacity of high-order schemes to simulate the incompressible flows encountered in chemical engineering. The combination of high-order CFD and LES would make it possible to develop more accurate and robust models for the reactive and multiphase flows that occur in chemical processes. The overall long-term objective of this research program is to develop open source high-order high-performance multiphysics and multiscale CFD models for single phase and solid-fluid flows. With these models, researchers, industry and, other stakeholders will be able to use process intensification. Our short-term objectives (next 5 years) are: 1) To generate high-order unresolved CFD-DEM (Discrete Element Method) models to simulate turbulent flows in solid-fluid contactors of arbitrary geometry 2) To develop large-eddy and direct numerical simulation models to simulate catalyzed reactions and to predict species and temperature profiles in reactors 3) To elaborate multiscale modeling strategies that use deep learning to bridge the particle/pore and process scales The results of this work will make PI possible for single and solid-fluid processes such as reactors. PI can reduce plant size dramatically (>100x), provide energy savings ranging from 20% to 80%, and decrease reagent inventories up to 10- to 1000-fold. Achieving our objectives will provide chemical engineers with validated models and computational tools that will enable them to use PI to design new highly efficient unit operations or to optimize existing ones.