Learning Forum

[Joe]

Hi. I have a simulation where I want to utilise the Runtime DFT capabilities to give me Fourier coefficients in the domain for different discrete frequencies. I am using the single tone runtime DFT option for static pressure at a single frequency (for testing purposes). For reasons I'm not clear on, following sufficient run time, the DFT results are zero for both magnitude and phase. 

I suspect this may be due to the fact that when I run the calculation, I have a boundary condition that updates at every time step (following each time step, a new pressure profile is defined at one boundary). Would this implementation be a reason that the Runtime DFT option would not work correctly? Surely, if the data is still in memory (i.e. the simulation isn't closed during the runtime required), the DFT should still work. I'm wondering if anyone has any experience or suggestions?

Thanks in advance,

Joe

[Federico]

How are you specifying the pressure profiles at each time step?

[Joe]

Hi, Thanks for your message. 

After each time step, I define a profile at the pressure outlet. In my journal file, this reads as 

"file/read-profile "location/*.prof"
 /define/boundary-conditions pressure-outlet outlet_boundary yes yes no pressure p no 284 no yes no no yes no 10 yes yes no no no

 solve/dual-time-iterate 1 10
 define/profiles/delete-all".

I repeat these steps for the length of my simulation. The *.prof file is set up as specified in the example in the user guide. 

I have also tried running the simulation for many time steps (longer than specified in the DFT setup menu) but still do not get data out of the Runtime DFT. 

Thanks in advance,

Joe

[Federico]

Instead of reading .prof files, can you test the same case with a time varying Named Expression for this outlet_boundary? Just to check if this has any impact

[Joe]

Hi - I have now tried this but still get zero's for all Runtime DFT output. Do you have any suggestions? Thanks

[Joe]

I will try. I have tried running the same case as above without the profile boundary (just a single-valued pressure outlet) for many time-steps and still get no data. This was a continuation of a time-varying simulation so will have included oscillations I'd expect to get data for.