Current Project that will hopefully give me a PhD
- My PhD project started out as an investigation into alternative ways to simulate fluid flow dynamics. Traditionally the Navier Stokes equations are used, but new methods are being developed that may have certain advantages over the incompressible Navier Stokes equations. These include the lattice Boltzmann equations and DPD (dissipative-particle-dynamics).
- I learned about the lattice Boltzmann equations and about the Immersed Boundary method (which is a must since Charlie Peskin is my academic grandfather).
- I then incorporated the lattice Boltzmann method (as the fluid solver replacing Navier Stokes) into the Immersed Boundary Method (this is done using asymptotic expansions to relate the Navier Stokes equationss to the microscopic lattice Boltzmann equations).
- Every method needs an application. Since Aaron is my advisor, it was natural for me to pick a blood flow problem. It is a well know phenomenon that platelets (which help your blood clot!) in whole blood get "pushed" toward the vessel wall under arterial flow conditions due to the presence of red blood cells. This has been seen experimentally but how and why this happens is not currently known.
- So I used my model to simulate this effect, but due to the high number of red blood cells in the simulation, the code ran way. too. slow.
- I then parallelized this code using OpenMP and then using MPI (which turns out to be even faster than OpenMP. It would be cool to use a hybrid method, but the math computers aren't set up for that).
- I use this parallel version of the code to visualize the influence red blood cells have on platelets motion in blood vessel, specifically how something called lateral platelet motion occurs and creates a near-wall excess of platelets.
- Now I'm trying to understand the physics behind this motion.