Computational fluid dynamic (CFD) modelling of dense particle clouds and strands
Many important processes depend on the behaviour of dense particle clouds and strands. Such dense clouds and strands will behave quite differently to the way individual particles would in the same fluid. Numerical simulation of this is very difficult, demanding proper modelling of two-way coupling between fluid and particles, and of the particle-particle and particle-wall interaction.
We study this at different levels of detail:
Eulerian-Eulerian modelling, where the particles are treated as a pseudo fluid, interpenetrating and interacting with the carrier fluid. This is suitable for simulations of larger systems.
Eulerian-Lagrangian modelling, where each particle is treated as a point, and tracked in the flow field, using models for their interaction with the fluid and with other particles and containing walls
Direct numerical simulation, where the particles are large relative to the computational grid, so that the fluid flow equations can be solved around them, giving the forces acting on them directly.
In recent work we have extended the classical hard-sphere model to include adhesion and cohesion so that a collision may result in deposition on a wall or agglomeration of particles. This makes it possible to study a wealth of industrially relevant processes, e.g. for flow assurance, directly by Eulerian-Lagrangian simulations. Two articles about this are in print, one in Phys. Rev. E and one in Chem. Eng. Sci.
Animation of a particle collision using a standard hard-sphere collision model
Animation of the same particle collision using the modified hard-sphere collision model with cohesion, the Hamaker constant is high enough in this example to cause agglomeration
In our fuel cell activities, sponsored by the NFR and carried out in collaboration with CMR Prototech AS, we are looking at the production and characterization of nano-particle-based raw materials for oxide fuel cell (SOFC) components and the formation and testing of cells from these raw materials.
Specifically we are
producing and characterizing a wide variety of nanoparticles by a new, patented, variant of the sol-gel process for use as precursor powders for SOFC components.
Producing and testing thin layers of controlled porosity from the nanoparticles to produce solid oxide fuel cells with improved efficiency, e.g. with electrophoretical deposition
Nanoparticles of yttrium-stabilized zirconia
SEM image of complete cell produced during the project
Stochastic models for particle transport in fluidized beds
Most mathematical models in chemical engineering are deterministic, based on conservation equations. For a complex system such as the transport of particles in fluidized beds, this approach, however, often does not lead directly to a solvable model. In this study a new modeling approach, more consistent with the inherently random particle transport, based on Markov chains is used. There are many attractive features of this approach, for instance the models are easy to formulate based on physical ideas about the motion of a single particle, and they are computationally easy to handle with a matrix oriented package.
Although there is no project running in this field right now, this is a continuing interest in the group. Some of the things we have been looking at are:
Slugging fluidized beds, modeling the displacement of particles in the system as a result of the formation and rise of slugs.
Formulating models for the gas and particle flow in the regeneration vessel of the FCC process, investigating the effect on the performance of the unit of changing different operational and geometrical parameters.
Residence time distributions for particles or fluid elements in processes
Cohesive particles and the buildup of plugs and deposits in process equipment and pipelines
Understanding the build-up of solid-particle deposits is crucial for the processing industry to avoid deposition and plugging by knowledge-based design. This study aims to observe and model the build-up of particle deposits in transport and processing equipment. Experimentally the build-up on solid surfaces is studied by micro-imaging of a surface exposed to a particle-laden flow. Theoretically CFD simulations, making use of the above-mentioned collision model, are carried out.
In another activity we look at the surface properties of hydrate particles in various environments using classical molecular dynamics (MD) simulations, and determine the forces between to opposing surfaces, to determine the cohesivity of the particles, and the most important factors influencing cohesivity.
CFD simulation of particles for process safety applications
Some of our activities are carried out in collaboration with the group Process Safety. In the field of dust explosions the behaviour of dense particle clouds and strands needs to be modelled precisely to assess the danger of formation of explosive clouds and to model the progression of an explosions once it has been initiated.
In spite of the fast integration of CFD into academia and industry, many fundamental problems still need to be resolved, particularly in relation to multiphase systems. Simulations, that at first glance look plausible, may, in fact, have little to do with reality if the interaction between the phases is not properly modelled, and this may be critical when assessing the danger of accidents taking place.
Comparison of dust lifting from a smooth (top) and a rough (bottom) wall behind a passing shockwave
Solid Separation from Highly Viscous Liquids by Cyclone Technology (CLEANSAND)
The aim of this project is to design innovative hydrocyclones to extend the use of hydrocyclone technology within the oil and gas exploration industries.
Two experimental techniques are being used, one is to test the separation efficiency of liquid cyclones on new types of particles and fluids and relate the results to know models for cyclone efficiency with the aim of designing cyclone-based separators for new and exotic applications. The other is the study the actual separation using the technique referred to as PEPT (positron emission particle tracking), where a radioactive particle is followed as it flows through the separator in a PET (positron emission tomography) camera.
Grade-efficiency curves obtained in a dedicated experimental rig, the size analyses required to generate these curves were carried out using a sedimentation-based technique, eliminating a number of potential errors arising when other sizing techniques are used.
Animation of tracer particle moving through a hydrocyclone measured using the PEPT technique. The size of the particle is exaggerated.
The project is carried out in collaboration with BKK. The aim is to model the working of Pelton turbines economically and precisely, gleaning on the information that is relevant to determine the efficiency of the turbine. The task is challenging, mainly due to the free jets impacting on solid surfaces in such turbines
The project is in its start-up phase. The aim will be to investigate diffusion of gaseous species in networks of micro-and nanopores experimentally and theoretically. Applications are solid-catalysed reactions of gaseous species, e.g. in fuel cells.