I have reach a difficult teorical point to resolve in my project.
I designed a pseudo-equatorial solar tracking system. The system runs in a Arduino Mega.
It moves south/north and west/east using DC actuators, which I control using 4 relays. To know where the solar panel is I use an Inclinometer (+-90º in both axes) (Murata Scat100d2)
I have written the code using Hannes Hassler`s SolarTrack4Arduino which gives all the solar information I need (Solar elevation, azimuth and solar hour angle)
All these works good.
But I do not really know how to integrate that information so I can get the correct angles to position the solar panel. I know how to move the solar panel to a known angle, but I do not know how to calculate the angle so the panel faces the sun.
At my first trial I made:
To calculate correct North/South axis: 90º-Sun elevation.
To calculate correct West/east axis: Same as the hour angle, but in the other direction. For example at the morning when solar hour angle is -30º tilt the sun panel 30º east.
As I could prove, it does not work. I think my approach is not correct, and I would need to get vectors involved. Is that correct? Do you know where I could get more information? (I have search in google but I coundn´t find any usefull information)
Thank you very much for your help! All ideas are welcome and sory for my broke english!
Why not just use two pairs of LDRs to tell you when the panel is pointing at the sun. Just move the panel a little towards the brighter LDR and then check if another movement is required. Probably 20 lines of code.
Robin2:
Why not just use two pairs of LDRs to tell you when the panel is pointing at the sun. Just move the panel a little towards the brighter LDR and then check if another movement is required. Probably 20 lines of code.
...R
solar power can be pulled from a cloudy sky that would not have the panels rotate by an LDR.
besides, why not learn how to do a thing ?
dave-in-nj:
note that a heliostat moves partially,
the heliostat is a half-way point to hold a reflection onto a fixed object.
you want to track the sun directly.
Looking in that webpage. . . lots of code, not directly what I need but maybe I can take some ideas out of it!
Thank you very much!
Robin2:
Why not just use two pairs of LDRs to tell you when the panel is pointing at the sun. Just move the panel a little towards the brighter LDR and then check if another movement is required. Probably 20 lines of code.
...R
I though about that, plus there are lots of examples and tutorials already made, but I have two big problems:
1.- My solar panels are built in fresnel lenses (concentration type) and so they need +-3º accuracy to work
2.- There is a huge white wall and the reflection of the sun in the wall confused the sensors
Anyway if this approach exceeds my ability I will be doomed to that system
I admire your attempt to get an algorithm to run, will watch with interest.
My solar panels are built in fresnel lenses (concentration type) and so they need +-3º accuracy to work
Even more reason to use a real-time light sensing system.
By the way the sensors are not tracking the sun, they are tracking to follow where the greatest solar energy is coming from, big difference in some cases.
There is a huge white wall and the reflection of the sun in the wall confused the sensors
Hi,
If the white wall is the wall in the background of your picture, it looks like it will be "behind" the array, I don't think it wold be a problem, you would be putting tubes over the sensors to help get your direction accuracy, so they wouldn't see any glare from the wall.
