Projektmenu

Introduction

Recently, Lagrangian simulation models have come more and more into focus as they are perfectly suited for modelling particle motion and fluid particle interaction at a micro-scale. They allow capturing effects such as segregation, cavity formation or interaction with flow inserts, which are important in metallurgical industry.

  • Develop dedicated micro- and meso-scale Langrangian models in a CFD-DEM framework
  • Embed detailed simulations in larger scale simulations for industrial application
  • Development of a comprehensive coarse grained discrete particle model to simulate industrial moving bed reactors
  • Gain an in depth understanding of metallurgical processes at particle scale

Methodology

In Project 4.3 existing Lagrangian particle models will be further developed with special focus on industrial application in the field of moving bed processes. This objective requires a combination of mathematical model improvements and efficient hardware-aware implementation.

Resolved simulations (grid size << particle size) are accurate, but not feasible for application to real-world applications. A two-fluid approach (grid size >= particle size), on the other hand, does not provide direct access to the behaviour of individual particles. In this project both worlds are combined. The dynamics of solid particles are computed using the Discrete Element Method (DEM) using the open source DEM code LIGGGHTS. Motion of the continuous phase, e.g. a gas or liquid, is simulated using the open source CFD package OpenFOAM. Here, momenta and changes in local volume fractions are exchanged between LIGGGHTS and OpenFOAM using the coupling code CFDEMcoupling. In this framework all codes are open source and be readily be adapted to the needs of this project. On the meso-scale, this framework will be extended to describe chemical reactions in the gas phase (OpenFOAM), the particle phase (LIGGGHTS) and conversions between the two as well as heat transfer transport of fines/dust, deposition and influence on the flow field.

In order to go to the size of industrial plants, approximations have to be imposed, e.g. coarse graining (instead of single particles, parcels of particles and their trajectories are simulated) or two-fluid models. In regions of particular interest, fine grained simulations can be embedded in the coarse grained surroundings.

Results and application

This project assembles three scientific partners contributing specific knowledge. In the field of particle flows and coarse graining Johannes Kepler University Linz (Department of Particulate Flow Modelling), the topic of gas-solid reaction modelling is covered by Technische Universität Wien (Institute of Chemical Engineering) and for description and modelling of combustion processes and NOX formation Montanuniversität Leoben (Chair of Thermal Processing Technology) has competences. By joining forces these groups develop a comprehensive coarse grained discrete particle model for the simulation of reactive, industrial-scale moving bed reactors.

The direct injection of alternative reducing agents into the raceway cavity of blast furnaces, computed based on finite volume approaches and Lagrangian tracking schemes, will be coupled to simulation tools that compute the movement of particulate matter of the coke bed in the presence of fluid phases applying discrete element methods. Consequently, the shape of the raceway cavity shall be computed explicitly, accounting for combustion processes, blast momentum and solids pressure effects resulting from the load in the furnace shaft. The gasification and combustion of injected material (e.g. oil, tar, plastics, pulverized coal etc.) and their influence on the raceway structure and geometry will be studied systematically.

By means of project P 4.3 (Discrete Particle Models) we will gain an in depth understanding of metallurgical processes at particle scale. Similar to project P4.2 these findings will enable industrial partners to optimize existing plants or to develop new ones. One the one hand, coarse-graining methods will enable the simulation of whole processes at industrial-scale. One the other hand these discrete particle based models can also be embedded into P4.2 Bulk Solid Models, resulting in a discrete particle based ‘magnification lens’ of interesting sub-regions of a bulk-solid reactor. The modelling methods developed in Project 4.3 are integrated into the K1-MET Simulation Platform of Project 4.1.