I'd like to see a comparison with a dumb as a box of hammers controller like PI or extremum seeking on the local gradient. I can't imagine that it is learning much more than a simple strategy like this, but maybe that's my lack of imagination...
I built an open-source, AI-powered solar panel that's 95% optimal
11–20 of 143 posts
Re: I built an open-source, AI-powered solar panel that's 95% optimal
#12This is very cool, and a neat application / exploration of RL. Am I correct in understanding that the results are only from the results of simulation? If so, it'd be cool to see this work (both a time-lapse video and a chart) in the real world, over a day or over a week!
Re: I built an open-source, AI-powered solar panel that's 95% optimal
#13I'll just leave this here...
Re: I built an open-source, AI-powered solar panel that's 95% optimal
#14Given any GPS coordinates, couldn't one simply calculate the location of the sun in the sky to point directly toward it?
Re: I built an open-source, AI-powered solar panel that's 95% optimal
#15Sorry for the off-topic. I think if just 10% of the efforts were put on improving mini/micro/pico hydro electric generator rather than on solar systems, most of the rural areas will probably better off now with reliable power supply than relying solely on the intermittent solar power[1], [2].
[1]Micro hydro:
https://en.m.wikipedia.org/wiki/Micro_hydro
[2]Micro hydro power with turgo generator:
Re: I built an open-source, AI-powered solar panel that's 95% optimal
#16I'll just leave this here... https://www.youtube.com/watch?v=wL9PcGu_xrA&t=174s
Great, simple solution, especially if the axes of motion are un-restricted.
Re: I built an open-source, AI-powered solar panel that's 95% optimal
#17I'll just leave this here... https://www.youtube.com/watch?v=wL9PcGu_xrA&t=174s
Simple
Solution is often the best. Although I wonder what this adds to overall build costs?
Re: I built an open-source, AI-powered solar panel that's 95% optimal
#18I'd like to see a comparison with a dumb as a box of hammers controller like PI or extremum seeking on the local gradient. I can't imagine that it is learning much more than a simple strategy like this, but maybe that's my lack of imagination...
Correct. Basically all the gimbal is doing is minimizing the cosine loss. This is, literally, the simplest loss function possible. Even if you don't know where the sun is, damped gradient descent will solve this problem every time, with no learning required (albeit with some jitter because you're having to estimate the gradient by moving the panel around). If you treat it as an iterative learning control (ILC) problem, though, you can learn the correct trajectory for the panel in just a couple of iterations and take the jitter out.
Re: I built an open-source, AI-powered solar panel that's 95% optimal
#19Given any GPS coordinates, couldn't one simply calculate the location of the sun in the sky to point directly toward it?
My general approach is if we're not doing X and it sure seems like we could simply do X, we probably cannot simply do X.
Re: I built an open-source, AI-powered solar panel that's 95% optimal
#20Earlier quoted context omitted.
So large solar farms are usually single-axis tracking, which provides a huge energy production benefit over static panels. Consider this panel's application more for a standalone installment, where it could make sense to use a dual-axis tracking panel over a single-axis or static panel. Re: solar farms and cost, I actually learned that around ~70% of the cost of solar installation is soft costs, not the actual panels…
> around ~70% of the cost of solar installation is soft costs This is incorrect, especially for solar farms Panels are indeed only about 1/3 of the cost. Additional components, labour, inverter, mounting - another 1/3. The last third is indeed soft(ish) cost, but this includes profit (duh..), certification, survey, tax, fees, etc. This can be reduced, but it is not going to magically disappear ... Another fun fact :…
>low altitude
Low latitude?