Maximum all-directional radiative slice and device implementation?

[Patrik Král]

Hi all,

Is there any device and slice file option for maximum radiative flux assessed from all directions? The value of the radiative flux depends on the direction of the device and sometimes it can be trictky to choose the best angle of the device to pick the worst radiative flux.

However, I can imagine why it may have not been implemented yet as it is surely very CPU expensive - to calculate the flux for all angles within the rad angle steps set in FDS, this for each direction of the device and from this to choose the maximum value, and this for all volumes in the slice file; it is a lot of calculations.

Something similar is done by Firewind software (http://members.optusnet.com.au/~firecomp/), but it is calculated only for a small number of radiators, not radiation from all angles (sources) to all possible angles of the receiver, and not in time.

Thanks for answers.

[dr_jfloyd]

FDS, by default, doesn't update all radiation angles in a single timestep.  It updates a fraction of the angles  with the solution ultimately updated over ANGLE_INCREMENT * TIME_STEP_INCREMENT timesteps (defaults of 5 and 3 respectively). This means we don't need to store all angles in the gas phase, only ANGLE_INCREMENT values.  For a slice file you would have to store all angles everywhere. This would be very expensive in terms of memory.  

The heat flux at a location is the integral over the two pi angles centered on the surface normal of the incident flux times the flux direction cosined with the normal direction. The maximum flux would require doing that integration over all possible normal directions and storing the maximum value. This would not be a cheap calculation for a device and extremely expensive for a slice. 

In most cases, I expect you would have some idea of rough direction or range of directions where the maximum flux would exist. You could add multiple flux devices at a single location with each having a different ORIENTATION and then use a MAX control function to pick the maximum value and output that maximum with a CONTROL VALUE device. 

[Patrik Král]

I understand the first paragraph, I might have not expressed myself correctly. Yet I might disagree with the last sentence – I don't think storing the values for calculations for a single DEVC would be so dramatic (for RAM, and CPU-cache-RAM data transfers), and the result is then still just a single value. However, as I mentioned in my original question, I am aware of all the steps and necessary calculations and how extremely expensive it would be for a higher number of devices or a slice. It was just a naive wish of mine to which I wasn't giving many chances.

About the last paragraph, I thought of this too.

Thank you.

Dne pondělí 16. května 2022 v 23:17:15 UTC+12 uživatel dr_jfloyd napsal: