Spectrum data simulation in Radiance

I am doing research on simulation of solar cells. Since solar cells have their own absorption distribution, we want to simulate the light spectrum data with Radiance. However, I didn’t find where I should look at.

Welcome to the forum!

Radiance has 3 color channels to work from, and these can mean whatever you want them to mean rather than the normal RGB. For example, you could put all the radiant flux for the sun in the green channel and all the flux for the sky in the blue channel, then use surface material channels that reflect each component according to their reflectivities integrated over the corresponding spectra.

After calculating the irradiance at each PV array, you then multiply the absorption for the sun and sky spectra against the G and B channels then add them to get the total absorption.

The other option is to repeat the calculation N times to get N*3 spectral bands, but that takes N times as long and probably wouldn’t be any better in terms of accuracy.


Hi, Greg,

Thanks for your reply. Actually, we are working on the visible light range from 380nm to 780nm. When the interval is 10 nm, we can get 40 spectrum data values. So, I want to set the light source and material reflectance with 40 spectral values. In this way, the whole model will be more realistic. Is it workable? Thanks.

There are two Radiance based tools for spectral lighting simulation:

  • Lark works by running multiple Radiance simulations with three spectral channels per simulation.
  • Alfa works by tracking ray reflection paths and performing an external multiplication with spectral reflectance values.

Both are targeted at human circadian simulation, but it might be possible to use them for your purpose as well.

However, I’m not sure what you mean by “more realistic”.

1 Like

Hi, Nathaniel,

Thanks. Alfa is similar to what I want to achieve. “More realistic” means the light spectrum consists of multiple wavelengths instead of {R, G, B} only.