Modeling Flow/Saturation/Absorption in Fibers

Gene Dougherty

Hi. I have had quite a bit of experience doing process modeling in another job with the previous company I had worked for. However, modeling was unappreciated there, so I got downsized. Now I am with a new company. At my previous company I had been trained in and used Fluent and Gambit to solve a CFD modeling problem where I got a lot of good insights into mixing (especially spatial, compositional and temperature heterogeneities) for a special polyurethane reactor with remarkably little effort.

Sadly, my new company does not have a Fluent or Gambit license, so I am starting sort of from scratch. It is new to using modeling to solve problems. This is partly why this (consumer products) company hired me: to see what computer modeling might provide as a means of understanding how their products work and how to design new, better ones. Specifically, I am looking at a problem related to absorption of fluids by solid, hydrophilic fibers that will likely deform as a result of becoming partially saturated. Because of confidentiality, I can't get too detailed, but I can give you an idea of what I think I want to model.

Because of voids, I see this problem as a partially saturated flow problem in a porous medium, in which the medium itself (i.e. the fibers) can "flap around", reorient and deform in response to the fluid. Flow takes place in a channel, so the fibers can't go just anywhere. Depending on the fiber materials, boundaries, fiber design and orientation, etc. the fluid (with its own unique properties) could conceivably bypass the fibers, form a gel, and/or be absorbed by the fibers by either capillary action or osmotic forces, as least until the deforming fibers get saturated. We do have access to experiments to use for some sort of model validation, but the modeling we really want to do is difficult to check fully by experiment.

I've been asking software companies (e.g. CFD contacts at Fluent, Ansys/CFX, as well as FEA places, etc.) for their recommendations on how best to solve this problem. Naturally, every company recommends their own software. To further complicate matters, some CFD company licenses are very expensive and costly to maintain (for the limited amount of time we would spend using CFD). Another choice is to work with a consultant. But many consultants just want me to turn over the problem to them, and let them solve it, giving me a report in return. Being a "hands-on" modeler, I don't like to work that way.

My intuition suggests that this problem is solvable, or at least we'd learn something by modeling it, but want to get with the right partner.

Anyway, my questions are: Do you agree this problem is "addressable" through software ? What would you guys recommend ? What software is best suited to this ? What should I read to get familiar with the problem ? are there any "mentors" who could assist me with modeling, as opposed to consultants who want to take the problem away from a hands-on geek like me ?

Thanks.