k-Wave

1. rodrigo.silva0494
   

   Member
   Joined: May '19
   Posts: 3

   Dear Mr. Treeby,

   First of all, thank you so much for developing this incredible tool.
   I've been writing a code for a 3D heat transfer simulation by a HIFU, and I realized it was possible to create a temperature field by using the kWaveDiffusion class without defining the boundary conditions. Then, I read in another topic that, by default, the thermal simulation uses a periodic boundary condition.
   I was wondering if there was a simple way to specify the boundary conditions I need for my simulation. My region consists in a 50mmx50mmx50mm cube, and the temperature at the six surfaces must be constant equals to 37 degC.

   Thank you for you attention.
   Rodrigo Silva

   Posted 4 years ago #

2. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,640

   Hi Rodrigo,

   Are your simulations in a homogeneous medium? If so, then this is possible with a few undocumented options, and some extra codes to calculate discrete trigonometric transforms. Let me know. If your medium is heterogeneous, unfortunately, only periodic boundary conditions are supported at this time. One practical way to counteract this is to make the domain larger, so the temperature on the boundary is maintained at 37.

   Brad.

   Posted 4 years ago #

3. rodrigo.silva0494
   

   Member
   Joined: May '19
   Posts: 3

   Hello, Brad.

   Thank you very much for your answer.
   I've developed my own code for the thermal simulation, so I am no longer worried about changing the boundary conditions when using kWaveDiffusion. Instead of that, I am having troubles while solving the acoustic problem. When I try to calculate the acoustic pressure field by using the extractAmpPhase, I recieve a message informing that I'd need about 190GB of memory to finish the simulation. My medium is heterogeneous (layers of water, skin, fat and muscle), and the grid is 256x256x256.
   Could you please give me any adivice for dealing with that?

   Thank you.
   Rodrigo

   Posted 4 years ago #

4. rodrigo.silva0494
   

   Member
   Joined: May '19
   Posts: 3

   Hello, Mr. Treeby.

   Acctually my problem was in recording the acoustic pressure without setting any start index, which made me record a lot of information that I don't really need. I set the start index from a point after reaching the steady state - also close to the final time -, and the simulation worked perfectly. Sorry for my mistake.
   Now I am indeed trying to use the extractAmpPhase function, but I recieve another error message informing that the data can't have more then 3 dimensions. How can I use this function if, since I'm running a 3D simulation, the input is a 4D matrix (3D + time)?

   Thank you.
   Rodrigo Silva.

   Posted 4 years ago #

5. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,640

   Hi Rodrigo,

   This problem will be solved using the upcoming release (V1.3). In the meantime, you could reshape your data to 2D (space, time) before calling extractAmpPhase:

   data_rs = reshape(data, [], size(data, 4));

   Hope that helps,

   Brad.

   Posted 4 years ago #

RSS feed for this topic

Reply

You must log in to post.