[ashtonJ]

Hello everyone,

I am working on developing an in silico model to mimic complex and dynamic nature of coronary arteries. To do so, I need to combine different computational modules, as listed below, in one computer model:

1. Computational fluid dynamics (CFD) for modelling haemodynamics
2. Fluid structure interaction (FSI) for modelling the fluid and vessel wall interaction as the result of blood pressure waveform
3. Moving-boundary method for modelling cardiac and respiratory motions
4. Multiphase flow for modelling blood plasma and morphological elements of blood
5. Particle transport model for modelling mass transport of low density lipoproteins and other blood species to simulate genesis and progression of diseases like atherosclerosis

As far as I know using ANSYS CFX (or any other closed source computational tool) it is not actually possible to use all of the above modules together and have them in one computer model.

I am just wondering if it would be possible using open source software like OpenFOAM to incorporate all of the above modules in one computer model.

Can anyone please enlighten me on this, and also let me know if there are other computational modules that I also need to take into consideration.

Thanks
AshtonJ

[ghorrocks]

This is going to be an extremely challenging simulation, no matter what software you use. Has a simulation of this complexity even been done before by anybody?

OpenFOAM? Good luck

CFX is one of the strongest softwares available for running complex and interacting multiphase models. So your modules 1, 2, 4 and 5 I would be choosing CFX for. CFX is not so strong in moving mesh stuff, so 3 will be a challenge.

So I would be recommending CFX, but of course I am biased. Also I would be seriously thinking about whether doing all the multiphase models you describe and a moving mesh model of a heart is even possible.

[ashtonJ]

Thanks Glenn.
I do not think any one has done this before.
Is it possible to use modules 1,2,4,5 at the same time in CFX ?

[ghorrocks]

There are many, many multiphase models in CFX. Most of them are OK to run together but there are some exclusions. Until you decide exactly what multiphase models you are using you cannot be sure.

Also, I do not know what you want to do with multiphase modelling of blood plasma and blood morphology, lipoproteins and disease modelling. Until you decide how you are going to model these processes you cannot be sure whether they will run together.

[ashtonJ]

Actually my aim is to develop a comprehensive computer model which can mimic coronary arteries in every aspect.

Some questions that may be answered by having such computer model are:
The effect of moving walls on multiphase flow?
Do RBCs move differently through the vessel if there is a lot of compliance?

I've used all the above-mentioned techniques separately as part of my PhD study and during my postdoc. My focus is on coronary artery disease.

Actually I am planning to apply for a postdoctoral fellowship for which I am trying to find a fundamental research topic in the area of computational modelling with potential application in human disease. I thought that such comprehensive model can be used not only for medical application but also in many other engineering applications.

[ghorrocks]

This sounds very useful, yes. It also sounds fiendishly difficult.

If you are a world expert in blood flow and physiology modelling, CFD and numerical methods then I wish you the best of luck. If not then you will probably be wasting your time because you will never get it working.

I don't know whether this applies to your case or not, I don't know your background and I don't know what you are trying to achieve. But I have seen this so many times I have to say it:
I have heard many times people say, when faced with a complex, interacting system that we will "just model everything". Of those people, 100% of them failed to achieve much useful at all. Not one of them produced a model which did anything constructive. The ONLY way to approach complex physics problems is to work on the physics components one at a time. You validate and verify and make sure that the physics you are modelling is captured correct. And when you look at complex interacting physics models a far more success approach is usually reduced order methods.

I don't know whether those comments apply to you. But you should know that this is one of the common traps in CFD work.

[Lance]

Originally Posted by ashtonJ

2. Fluid structure interaction (FSI) for modelling the fluid and vessel wall interaction as the result of blood pressure waveform
3. Moving-boundary method for modelling cardiac and respiratory motions

Why do you think that motion and/or FSI are significant for your application? HeartFlow (https://www.heartflow.com/) computes coronary blood flow based on a single CT time frame and they do not have any motion in their models. Granted, they (only) compute pressure differences, and you seem to be going down the 'progression of atherosclerosis'-way, but consider developing a model without motion first.