An investigation of the lack of separation of scales effect in sub-grid modeling of rapid gas-solid flows

• Grant number: 11/06604-0
• Support Opportunities: Scholarships abroad - Research
• Start date: November 01, 2011
• End date: October 31, 2012
• Field of knowledge: Engineering - Chemical Engineering - Chemical Process Industries
• Principal Investigator: Fernando Eduardo Milioli
• Grantee: Fernando Eduardo Milioli
• Host Investigator: Sankaran Sundaresan
• Host Institution: Escola de Engenharia de São Carlos (EESC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Institution abroad: Princeton University, United States

Abstract

Circulating fluidized bed reactors are widely used in numerous industrial processes of great relevance, including large scale applications such as coal combustion/gasification and catalytic cracking of petroleum. A very relevant area of research in this context is that of mathematical modeling of the rapid gas-solid flows that develop in these reactors, which profoundly affect the concerning reactive processes. An important and up-to-date line of research in this field concerns the application of continuum Eulerian (or two-fluid) formulations in highly resolved simulations to generate sub-grid correlations, ultimately expecting to apply them as closures in large scale simulations of real size reactors. In view of the lack of knowledge regarding the complex physical phenomena involved, the highly resolved simulations also assume a character of computational experimentation, aiming for the formulation of new theories as well as closure models. This line of research has been practiced at the University of Princeton (Princeton-NJ, USA) since the 90's, under the command of Professor Sankaran Sundaresan. About a decade ago we joined this effort; our related studies are reported in several scientific articles published from 1999 to date (Annex 1 of the project). The evolution of our research has allowed us to productively interact with the international scientific community. Besides the contacts with Professor Sankaran Sundaresan, we also mention interactions with researchers of the U.S.A. Department of Energy: Thomas O'Brien and Sofiane Benyahia from the National Energy Technology Laboratory (NRTL); and Sreekanth Pannala from the Oak Ridge National Laboratory (ORNL). All of those researchers are assiduous in our bibliographic reference lists. From those interactions, and from our own growing experience, we formulated the present research proposal to be developed together with Professor Sundaresan's group, at the University of Princeton, starting from August 2011. The research project is intended to investigate the effect of the lack of separation of scales in sub-grid modeling of rapid gas-solid flows in circulating fluidized beds. Besides multiphase, those flows are highly heterogeneous, turbulent, reactive and multi-scale. All the scales of the flows intensely interact with each other, deeply affecting the transport processes of mass and energy, and chemical reaction rates. Those characteristics impose refined mathematical models, strongly based on computational fluid mechanics. The computational processing capabilities both present and in the foreseeable future, are far from allowing the solution of all the scales of the concerning flows. Therefore, large scale simulations (LSS) are performed in very coarse numerical grids, which require sub-grid models to recover filtered information. Sub-grid models can be generated from sub-grid simulations (SGS) applying two-fluid modeling in reduced domains under appropriate boundary conditions. There are here several challenges to overcome, such as the formulation of accurate descriptions for granular thermodynamics, rheology and drag that are appropriate for the micro-scale of the flows, and the difficulties related to the lack of separation of scales. In this research project it is proposed to investigate, through computational experimentation with two-fluid modeling, the effect of the lack of separation of scales over the results of sub-grid computational simulations. From the conclusions of these studies it is intended, in future works, to consider the proposition of new formulations and/or modeling procedures that take into account the interactions among the scales modeled in SSM and resolved in LSS. (AU)