3
$\begingroup$

This question might seem a bit off-topic, but I guess there are a lot of people here that know about optics, telescopes, etc.

I would like to simulate solar thermal systems that focus solar radiation, e.g. heliostats or dish Stirling. Usually they are simulated using raytracing or Gaussian cone optics.

Basically, I would like to simulate the optical system and get the resulting image of the sun on some surface.

In the end I would like to calculate the "flux distribution" (~brightness) of the reflected radiation on some sort of receiver. The receiver could be a plate with heat pipes beneath or a Stirling motor. I would like to create images like http://solarenergyengineering.asmedigitalcollection.asme.org/data/journals/jseedo/929668/sol_136_03_031013_f002.png

So far, I read that Zernike polynomials can only be used to simulate optical systems with circular apertures. This is not the case for my application. Currently I am evaluating the optical system using Monte Carlo raytracing (GPU based) , which yields the very high accuracy that is needed.

We have not done any work on this so far, but for very large mirror arrays, e.g. 25000 heliostats, the raytracer takes quite a while to converge. And calculation time is a very critical point when in comes to optimizing the positions of the mirrors to avoid the shading and blocking amongst them.

This is why we are looking for faster ways to evaluate the optical performance of these systems while still getting highly accurate results.

Do you think it is possible to simulate them using Zernike polynomials? The simulation needs to consider surface errors of the mirrors, astigmatism and non-ideal focal lengths. Can one use Zernike polynomials only for arbitraty optical systems?

$\endgroup$
  • $\begingroup$ Do you want to say 'approximate the surface using Zernike polynomials'? Else I can't make sense of what you're writing. $\endgroup$ – AtmosphericPrisonEscape Jan 10 '18 at 16:53
  • $\begingroup$ Sorry, I added more info to the question. $\endgroup$ – Joe Jan 10 '18 at 17:18
  • 1
    $\begingroup$ I think it would be great if optics questions could have their own SE site, but there isn't one. However, optics questions sometimes get excellent answers in Physics SE. It's helpful to formulate the question carefully and show as much prior work as you can. That helps define the level at which the answer is written. Also, the question shouldn't be just "is it possible" because an answer of "yes" may not be very helpful. Explain what you are trying to do, what level of math you are comfortable with, and what level of accuracy you need. Zernike polynomials may be overkill for a concentrator. $\endgroup$ – uhoh Jan 11 '18 at 1:56
  • 1
    $\begingroup$ I also posted the question in physics, but no response so far. I will add some more info to the question. $\endgroup$ – Joe Jan 11 '18 at 7:13
  • 1
    $\begingroup$ My guess is that ray tracing with GPUs is the only way to include all of the real-world effects that you need to include like pointing errors and deviations of all of the individual elements. Analytical methods will severely limit what bits of reality you can include. If you are computationally limited, maybe look for "short cuts" that are used in photorealistic rendering technology for movies. Possibly you can check out the Blender SE site. $\endgroup$ – uhoh Jan 12 '18 at 5:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.