Simulating microtransport in realistic porous media
Promotion date: September 27, 2012
Promotor: Prof. dr. ir. Bernard J. Geurts
Simulations in porous media widely adopt macroscopic models of transport phenomena. These models are computationally efficient as not all geometrical details at the pore scale are accounted for. Generally, these models require closure relations for effective transport parameters, where the parameters are used to convey information from the pore scale to the macroscale. Proper closure can only result from detailed knowledge of the characteristic behavior of pore-scale transport.
My thesis details the development of a numerical algorithm for simulating the pore-scale flow of an incompressible fluid with conjugate heat transfer using spatially periodic geometric models of porous media. The simulations use geometric data on the pore network extracted from detailed images taken using X-ray computed tomography.
Subsequent processing of the pore-scale results (i.e. velocity, pressure and temperature), yields predictions for effective transport parameters. The primary focus is on the transport parameters that quantify flow resistance and the rate of interphase heat exchange: the permeability and the interfacial heat-transfer coefficient respectively.
The numerical approach is described as is its validation in simplified geometric models of porous media. These are compared directly to data from a physical experiment, thereby establishing the effectiveness of the method.
How was your research positioned on a scale of fundamental versus application oriented?
The numerical algorithm was used to describe an industrial process which resembles the way an incompressible fluid flows inside a porous medium. The flow is laminar in structure, but the geometry of the actual porous medium is really complex. A simulation technique was developed to describe what really happens as process parameters and/or geometry parameters change.
The strength of the algorithm lies in its ability to quickly and easily transform data on the cross-sectional information of a porous medium – using micro-computed tomography – and build a suitable computational mesh of the domain. Assessing changes to the geometry and its behavior on the flow characteristics can therefore easily be investigated. This technique could ultimately lead to a tool that can support product development.
Did the results resemble reality?
Yes, very much so. With our simulation method we were able to replicate the exact same conditions used in a physical experiment of flow through a tube filled with a complex porous medium. We were able to clearly represent the porous structure and to simulate the way the fluid travelled through the pores. For us, this was the first time that we successfully validated our method against a real-world problem.
Most of the simulations were performed on the high performance computing facilities at Sara, in Amsterdam. As a member of a NCF/NWO funded program we were able to get vital computing time.
In what way did you develop personally as a researcher and a scientist?
I learned not to lose myself in all the details that come with doing scientific research. Being a perfectionist, this was very difficult for me. But, just being able to "step back", and gain some perspective on your problem is the most productive way of obtaining results.
I also learned how to communicate my results, both written and oral, to fellow-researchers and to laymen. My promoter encouraged me to do so. The University of Twente offers the great opportunity to present oneself at conferences, thereby providing ample opportunity to improve one's communication skills and to build your network.
Did you manage to have some nice publications?
Yes, one article was published in Computers and Fluids and the other in the International Journal of Heat and Fluid Flow. One more article is on its way.
What are your future plans?
At the moment I work as a contracted software development engineer in Switzerland for the company that funded my research.
What in your opinion is important to help Mesa+ remain successful in the future?
I like the way company and academic research can go hand in hand. In my case, a large industrial company provided a challenging problem onto which I could apply my skills in a multidisciplinary way. Some features of my research were fundamental, but there has always been a clear application for our developments. I guess, this can be the strength of the Mesa+ institute in the future. I believe great research can be done if universities and industrial partners can work harmoniously together. Mesa+ should definitely play a role in bringing these parties closer together.