Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Sun Tracking...need advice

Status
Not open for further replies.

PurpleButcher

New Member
hi guyz.... i am a hoobyist and designing a single axis solar tracker....

i am kind a stuck.....

Facts:
1)I am just monitoring the panel power
2)No sensor
3)Algo that i wanna use is P&O perturb and observe
4)the algo literrally involves a perturb and observe and continues to do so till the power gathered in the n step step is lesst than the power gathered in
n-1 step.... then it "moves back"

Problem:
i wana decrease this moving back.....so after the first two days data i wanna predict the angle of the sun..... for this prediction i am taking the average of the first two days data.... and then P&O will be apllied to just fine tune the panel....... thus i am hoping that the "MOVING BACK" will be reduced.... cause that is pure wastage of energy......

YOUR HELP:
i am kind ov stuck whether i am on the right track or not..... have i missed sumthing or is my approach kind a hoping against hope that angles can be predicted so easily???

Any advice will be welcomed :)

Thanx in advance
 
With a real time clock calendar and some amount of work you could build a setup to follow the sun with no feedback.

You are talking about a learning system.

For a tracker to know where the most power output it has to move past the desired position.

You could have it act as a tracker for a few days to get a base line. Then have it act as a positioner without looking at the power output. Each day it will be less correct.

You want to come up with a mix of the two. It is possible.

If it were me I would add the sensors.

3v0
 
If you have an XY tilt base it'd be easier to come up with the code to calculate the suns position in the sky at your solar panels location, and simply directly aim the base at that location. No MPP tracking required as it's directly pointed at the sun at all times.
 
The problem with using the power output for position feedback is that power output is affected by the weather. If you are moving the panel when a cloud passes over the sun then your algorithm will think it has gone too far and move back the other way. I would recomend the metod of using a clock to determine the position and a motor with position feedback. I have built two solar panel trackers using this method and it works acurately and effficiently. I have also heard of people using light sensors to acurately determine the position, but don't have any experience with this.

The trick to using a single axis controller to acurately track the sun's position is making your axis of rotation parallel to the axis of the earth. Then simply move 15 degrees per hour (360 / 24). Then change the angle the panels are at seasonally.
 
Another sourse of info would be to ask the local planetarium. They track the stars - tracking the sun should be "elementary".
 
Your could have a number of light sensors... A cluster of 5 being arranged relative to each other In the form north,south,east,west .. With the 5th sensor being in the centre of all of these... You will angle them to face slightly away from each other with the centre one facing straight.

Because of the angles some sensors will recurve higher light intensity thus producing greater voltage output... If the north and east are recoeving the most by a given threshhold RELATIVE TO THE OTHER SENSORS... It means your device must tilt/rotate to that direction... Untill the CENTER sensor has the most voltage output and the surrounding sensors outputs are approximately equal to each other by a given threshold with respect to each other.. HOpe this helps
 
I had to look at original post to understand the objective.

It is usually not efficient, from an engineering perspective, to do an empirical 'hunt and peck' system when the results can be totally predicted in advance.

The sun can be considered at optical infinity, so it is like setting up a telescope with a clock motor drive such the axis of rotation is parallel with earth's axis (north star pointing in astronomy). The sun's distance allows the movement of the true earth polar axis to your location, keeping it parallel to original true polar axis, with little tracking error at your location even across the earth's daily rotation.

The solar panel is then bolted down on this polar rotational axis 'pole' with an offset tilt angle that points directly at the sun. From a solar panel perspective this is good for about a month but because of the precession of the earth (23.45 degrees) it requires this offset tilt to polar axis be adjusted for time of year to realign the panel-to-polar axis offset to point to the sun. Assuming northern hemisphere it will range +23.45 degrees in summer going to -23.45 degrees in winter relative to the polar rotational axis.

You said you wanted a single rotational axis.

A simple way is just to do the polar axis rotation based on time of day (with appropriate stop positions to avoid running the panel into the ground). The panel tilt angle offset to polar axis can be reset manually every month or two to improve tracking across the seasons. Some systems just set the panel parallel to polar axis and take the error. There is about a 10% worse case loss for this but still much better then a roof top fixed mount panel. When you put the 10% worse case loss up against the losses for clouds it does not amount to much additional loss.

A more complicated mechanics is a year based cam motor offset to automatically adjust the panel tilt angle to polar axis for time of year.

You could use an acme screw drive used for geostationary satellite dish adjustment for the polar axis time-of-day drive and a small acme screw drive for the seasonal panel offset tilt.

Some sites that may help.

https://www.electro-tech-online.com/custompdfs/2009/08/solartracker.pdf

**broken link removed**

Astronomical Applications — Naval Oceanography Portal
 
...
The solar panel is then bolted down on this polar rotational axis 'pole' with an offset tilt angle that points directly at the sun. From a solar panel perspective this is good for about a month but because of the precession of the earth (23.45 degrees) it requires this offset tilt to polar axis be adjusted for time of year to realign the panel-to-polar axis offset to point to the sun. Assuming northern hemisphere it will range +23.45 degrees in summer going to -23.45 degrees in winter relative to the polar rotational axis.
...

Sorry to nit-pick, but isn't that actually +(23.45/2) to -(23.45/2) ??
 
What are the latitudes of Tropic of Cancer and Tropic of Capricorn?

Is it really that much? You're saying there is 46' difference in the angle of the sun between summer and winter?? :eek:
Not where I live.

(edit) Wow. I just looked it up the planet tilt is 23.44 degrees, total 47 degrees difference. I know here the sun is close to 90' on a summer noon, I never would have believed it was only about 40' above the horizon on a winter noon. For some reason I thought it was about 23' total difference between summer and winter.

I'm very close to the tropic of capricorn, maybe that's why I didnt notice the sun being so low in winter?

Now i'm really sorry I nitpicked because the nit was me haha.
 
Last edited:
Hi Roman,

I shouldn't worry about it.
Most people are only vaguely familiar with the path of the sun across the sky anyway.
Far less people have much idea of the 'sideways' movement of the arc over the year.

I'm only familiar with it myself cos i am looking into fitting up a 'wall-sundial'
while discussing this with various people, i was surprised at how few people have much awareness of planetary movement. Even compass points are very vague with some people.

You noticed, and you checked it out.
Good enough.

Regards, John :)
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top