Assign Useful Magnitudes to Rays and Apply Them

[kalebphipps]

Is your feature request related to a problem? Please describe.
With the solution to Issue #58, Rays in ARITST now have a magnitude attribute. However, these magnitudes are currently always set to a default value of 1 and we do not take advantage of these magnitudes.

Describe the solution you'd like
We should consider:

• How we want to set the magnitude, i.e. what factors influence the initial magnitude.
• How we want to manipulate the magnitude, i.e. what factors can change the magnitude.
• How we want to consider the magnitude when calculating distribution strengths etc.

[faezs]

Something like solTrace or Solarpilot uses the direct normal irradiance in W/m^2 to the heliostat surface and then has a multiplicative reflectance < 1 for each reflective surface. Given that the atomic surfaces here are facets, there should be a DNI correction via each facet’s angle to the sun. I’m trying to model a beam down topology so can take a look but would be helpful if you have more context from your more direct experience.

[faezs]

Shading from surfaces that are in the way and a distribution over clouds could be a nice thing to model…

[faezs]

There’s a fairly standard formula for geolocated temporal DNI estimates
https://www.pveducation.org/pvcdrom/properties-of-sunlight/solar-radiation-on-a-tilted-surface
But also there’s an NREL database that has historical data available for each location for a year or so. Imo one of the advantages of the GPU approach is not only data parallel multi heliostat but also data parallel day of the year. Feels pretty obvious that the time is a Functor in the sense that a simple map function should suffice but you probably want multi heliostats to be a parallel scan because shading might be a thing.

[kalebphipps]

Hi @faezs,

Thanks for the ideas.

As you might have guessed, we are currently prioritizing other aspects of development. Therefore, it will probably be a while before we are ready to implement the magnitude in the rays. Shading from surfaces in the way and blocking are (as you mentioned) issues that we are interested in modeling - but they are also a little ways down our developmental roadmap.

We can keep you in the loop as we get closer to fixing this issue!