Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Help  Simulate intravascular blood flow
#1
Sonrisa 
Hello! 

I want to simulate blood flow in real human vessels, the problem is, real vessels are very thin, the radius maybe less than 1 mm in some places.

So, can obi work in that small-scale environment? Any suggestion?

Thank you ~

PS: In my scene, the vessels are full-body vessels, not parts of them. And the vascular network is so complex that it can't be seen clear at all.

This is one part's image:
   
Reply
#2
(25-02-2020, 03:53 AM)mypcluna Wrote: Hello! 

I want to simulate blood flow in real human vessels, the problem is, real vessels are very thin, the radius maybe less than 1 mm in some places.

So, can obi work in that small-scale environment? Any suggestion?

Thank you ~

PS: In my scene, the vessels are full-body vessels, not parts of them. And the vascular network is so complex that it can't be seen clear at all.

This is one part's image:

Hi,

Particle-based simulations are not ideal for this use case. Fluid resolution is not adaptive, which means all of your fluid must have the finest resolution you need to work with. So even in thick veins, you'd need particles < 1 mm in size, just like in finer veins. In addition to that, coupling between fluid and vein walls is quite problematic, veins are very thin and you'd quickly get tunneling issues.

Imho, I'd steer clear of particle-based methods. If I were to implement a blood flow simulation from scratch, I'd use an eulerian approach on a tetrahedral grid. No tunneling issues, trivial coupling, and very fine, hand-crafted adaptivity. There's no commercial middleware that does this realtime, as far as I know, but I'm not really into the medical field so I might be wrong.
Reply
#3
(25-02-2020, 08:53 AM)josemendez Wrote: Hi,

Particle-based simulations are not ideal for this use case. Fluid resolution is not adaptive, which means all of your fluid must have the finest resolution you need to work with. So even in thick veins, you'd need particles < 1 mm in size, just like in finer veins. In addition to that, coupling between fluid and vein walls is quite problematic, veins are very thin and you'd quickly get tunneling issues.

Imho, I'd steer clear of particle-based methods. If I were to implement a blood flow simulation from scratch, I'd use an eulerian approach on a tetrahedral grid. No tunneling issues, trivial coupling, and very fine, hand-crafted adaptivity. There's no commercial middleware that does this realtime, as far as I know, but I'm not really into the medical field so I might be wrong.

I see, thank you for your suggestion.
Reply