Hi, All:

Anybody had experience in using CFD to simulate the blood flow in a capillary tube. Please recommend some references. Thanks advance!

Check out the work of Professor Klienstrauer (sp?) at NC State. The did work on flow through arterial grafts.

Jeff

Have a look at the work of Discacciati, Quarteroni, Deparis & Rozza at the Ecole Polytechnique Fédérale de Lausanne and at the Politecnico di Milano.

Cheers

Edi.

I have developped a specialized module for blood flows (including non newtonian effects) that I use for consulting projects. Would you be interested in testing it ?

Thanks. Jean-marie. In your module, did you include porous effect? In my case, I have a porous layer that will reduce the blood fill-in time. Thanks!

Hi, Tim! I use a general porous boundary condition at the outlet (to represent the capillary tree of arteries downstream my model), it could be adapted for the inlet side. Provide me with some data : radius, length, pressure drop, viscosity and I can prepare a mesh and input parameters. Thanks

To:jean-marie

I am doing a project on Flow Investigation of blood flow in LVAD, can I use your specialized module for blood flows? Regard

To: Richard Have you got a mesh (tetradra or hexaedra) ready, or can you prepare it with a meshing tool ? Then indicate the type of boundary conditions (pressure pulse ?, type of action of the LVAD :impeller ?) so that I can suggest an input data set. Then, for an academic project, I could let you have a few runs to test the numerical model. If there are 'industrial' prospects and fundings, we could then discuss on commercial basis ... Can you mail a sketch and a description of how the system works?

For LVADs, depending on the complexity of the geometry involved (and thus the complexity of the flow associated with it), there are far more important priorities in the simulation than the ability to model blood itself as a non-Newtonian fluid, especially in light of my own observation that there are no universally valid models for blood viscosity.

For LVADs (1) you want a code that can simplify mesh generation while maintaining solution accuracy, especially keeping in mind that you have a complex geometry that is potentially time-varying, (2) even more importantly, you need a code that can handle laminar-transitional-turbulent flow (if your LVAD produces swirling flow, as in most cases, then simple turbulence models will do very poorly). This latter is indirectly related to your gridding capability as well, because if you have numerical diffusion due to poor gridding (quality) you won't be able to get the correct vortical structures in the LVAD.

If you have more specific questions I'd be glad to help you out ...

Adrin Gharakhani

For Adrin,

I quite agree that the ability to compute on deforming grids is essential. The ALE scheme I developped on deformable finite volumes with constant connectivity is fully conservative, and uses a higher order interpolation in space with partial upwindind to reduce numerical diffusion. Would you be interested in envisaging a collaboration, I especially need to introduce a proper turbulence modelling for larger cavities such as this LVAD. My consultancy works to date deal with bounded jets (anastomoses, narrowed lumen coronary flows). I could mail to you a recent paper at a EUROMECH conference on BioFluids Flows. Jean-Marie

Jean-Marie,

I am generally open to collaborative efforts, but I have long ago moved away from grid-based computing to Lagrangian gridless solution strategies. Especially for geometries that experience significant change in shape as a function of time, gridless methods are ideal. We have successfully simulated 2D DNS and 3D LES of mechanical heart valve opening process (from fully closed to fully open) under physiologic flow conditions, which is clearly laminar-transitional-turbulent. Given the very wide range of length and flow scales involved in this problem, it would be very difficult to obtain good results with grid based methods. Nevertheless if you have any particular ideas in mind drop me a note and we'll discuss it.

Adrin Gharakhani

> I especially need to introduce a proper turbulence modelling for larger cavities such as this LVAD

I don't know anything about "this LVAD", but generally for the Reynolds number range involved in biofluid problems, k-omega is used in the literature. The list of reasons why any of the traditional turbulence models would be inappropriate for such problems is beyond the scope of this message. The most reliable model is LES with capability to account for transition (this excludes standard smagorinsky, for example)

Adrin Gharakhani