tracking solar array

mmoreland · 2011-05-17 15:25

I recently completed work on a single axis tracking solar array (24 Sunpower 230 watt panels). It's operated using a linear actuator powered with a CoolMuscle motor and a manually operated momentary contact toggle switch. I want to add automatic tracking, but I'm less clever with electronic devices than I am with mechanics. A short video of the array in operation can be viewed at: http://www.youtube.com/watch?v=gjWTwdLGGhs

I have a basic stamp 2, and I'd like to design and build a controller that would wake up the array in the morning, move it through about 4-6 changes of orientation each day, and put it to sleep at night. I need for it to be able to find real time, so my first questions are these:

Is there a sensor simpler than the Parallax GPS module that can fetch and keep track of real time? Is it possible that the Parallax GPS module can function also as an inclinometer?

I'm in way over my head, so any help will be greatly appreciated.

Michael Moreland
Mendocino, CA

erco · 2011-05-17 16:15

Why not skip GPS and differentially read 2 light sensors, one on either side of your panel? That would tell you which way to move the panel to balance the two sensors, which would make your panel perpendicular to the sun. Your stamp could regulate how often and how fast to move the panel.

mmoreland · 2011-05-17 16:49

Thank you, I could do that. It doesn't seem to be the method employed in the industry, but it's a good one. We have a number of overcast days when the sunlight is diffuse, and there is no particular point of strongest light, but I suppose that situation could be worked somehow into the logic that interpreted the two light sensors' input. I'm not sure, but I'll think about it. I'm somewhat committed to a plan that has a look up table with specific orientations for every day of the year so that it won't matter what the sun or the weather are doing. Nevertheless, I like your suggestion, and I'll take it seriously. mm

william chan · 2011-05-17 16:57

Would it be cheaper to add more panels than to add a tracking system?

prof_braino · 2011-05-17 17:41

mmoreland wrote: »

We have a number of overcast days when the sunlight is diffuse, and there is no particular point of strongest light,

If this is the case, then does it matter which direction the panels are pointing? For example, if this is the starting condition, and really big cloud moves in front of where the sun would be, then the differential sensor would find the better patch, I would think. Or do the panels require direct sunlight in order to work at all

Oh yeah, welcome to the forums!

mmoreland · 2011-05-17 17:41

Actually, the tracking aspect of the array is already built. It's just waiting for a controller. I did a cowboy cost analysis of tracking vs more panels before building it, and it was about twice as expensive to add more panels than to build a tracker with the additional problem that I have used up all the space I've got unless I put panels on our roof. I'm convinced that tracking, and my figures told me I'd gain 25% in energy collection by tracking, is a more cost effective way of increasing generation, and to top it off, it was fun to design and build. mm

mmoreland · 2011-05-17 17:44

Thanks for the welcome. I'm a real newbie at this. The panels don't require direct sunlight. They can operate at 25-75% of full generation depending on the thickness of the cloud cover. They respond well to diffuse radiation, and on such days, I don't bother moving them at all. Thanks for the suggestion. mm

prof_braino · 2011-05-18 06:21

Could you post links to your equipment? I've been interested in this but this tech is not so common in Chicago.

solar array (24 Sunpower 230 watt panels).
linear actuator
CoolMuscle motor

Thanks!

erco · 2011-05-18 08:54

Neat article on a simple tracker at http://www.motherearthnews.com/Do-It-Yourself/1977-11-01/Mothers-Super-Simple-Solar-Tracker.aspx

Quite ironic, Mother Earth News endorsing the use of Freon 12!

erco · 2011-05-18 08:56

This is more along the lines of what I mentioned earlier: http://www.reuk.co.uk/Simple-Solar-Tracker-Concept.htm

Franklin · 2011-05-18 12:07

prof_braino, all those items can be found with a simple Google search, try it.

mmoreland · 2011-05-18 18:57

prof-braino, Sorry about the delay. Sunpower has lots of stuff on their website www.sunpowercorp.com. Linear actuators are all over the place; I bought the one on the array on eBay and modified it. CoolMuscle motors are at www.musclecorp.com.

bee_man · 2011-05-18 21:17

How about a Real Time Clock chip. Many examples of interfacing a DS1302 or DS1307 can be found on this site along with sample code. The chip, a crystal, 2 resistors, and a watch battery are all you need to get it going, total cost 10 bucks.

Jim

Why only 4-6 changes of orientation each day, continually tracking will maximize your energy gain. You could easily do it with a photo sensor in a short tube, moving the array to keep the input of the photo cell maximized.

sjev · 2011-05-19 02:16

The goal of solar panel tracking is usually to produce max power output. You could do the following:
- once in a while execute a 'sweep' with the panel while measuring power output. Calculate max & position the array accordingly.
- However, looking at the video of your array, it seems too bulky. Another otpion is to use a photocel on a servo to find the optimal angle.
- If you don't want any moving parts, just build an array of multiple photocells at different angles, then determine which one produces most power (some calibration required).

Jev

NikosG · 2011-05-19 04:42

Hi all,
You can create a perfect solar tracking system with no sensors and only with a system that can keep the time and the date (a Real Time Clock chip like DS1302 or DS1307 as mentioned above by “bee_man”. Personally I have used my computer's clock). Of course a micro controller (Basic stamp or propeller) with the appropriate algorithm is necessary. You also need to know the latitude of the place (the Sun has different high in two places at the same time even at the same longitude. Think two different cities with the same time, one on South America and the other on Canada.) At this point you need a GPS but it is not necessary because you can set the latitude of your place as a constant in your program code. The whole system will have the inverse function of a sundial. The sundial can tell as the time in every specific day of the year according the Sun’s position. On the other hand the solar tracking system with algorithm can tell as the Sun’s position according the date and the time.
The second approach of this subject is of course the use of sensors. “Erco” described above the main idea of this method. Personally I have tried both methods and I think that the first is more cheap, more easy to build and more reliable. I have never tried the third method mentioned by “sjev” which based on output power’s measuring so I can't express an opinion about this.

prof_braino · 2011-05-19 16:39

Franklin wrote: »

prof_braino, all those items can be found with a simple Google search, try it.

Thanks Franklin. I wanted info on his particular components, rather than the entire offering from a manufacturer. In Chineese.

But thanks for the flash on the Google thing, it sounds like it might go someplace.

Franklin · 2011-05-19 18:21

Quite ironic, Mother Earth News endorsing the use of Freon 12!

Yes, but in 1977 who knew?

kwinn · 2011-05-19 22:34

NikosG has by far the most practical suggestion. It is simple, reliable, and energy efficient. No hunting for maximum power or balanced detector output to waste energy and wear out mechanics.

mmoreland · 2011-05-20 11:35

Hello NikosG and bee-man, As you can see from the video, my tracking array has only one axis of tilt. The axis runs east and west, and the tilt follows the sun's altitude rather than its azimuth. but it's a modified altitude that it tracks. If it were a double axis tracker, it would orient itself toward the sun's azimuth and then raise itself to the sun's true altitude, but for a single axis tracker, except at noon, the sun's altitude is a trigonometric step removed from the true altitude. In addition, my array doesn't face due south, but due to geographical constraints, it faces 8 degrees east of south. For all those reasons, I'm recording solar altitude figures as they relate to this specific array, and I will have a full year of data by next August. I will know when the sun is perpendicular to the array (in one axis) every half hour for every day of the year.

As it turns out from my observation, it's not necessary for the array to be oriented perfectly toward the sun in order for it to have very nearly 100% generation. Five degrees above or below perpendicular is as good as perfect. Additionally, the sun's altitude varies about 20 degrees on any given day although that interval processes from nearly horizontal in summer to nearly vertical in winter, still on a daily basis, the array travels through about 20 degrees to be maximally oriented toward the sun. Given the +/- 5 degree flexibility and the 20 degrees of travel, 4-6 movements in a day are quite adequate esp. since most of the change happens before and after a long stretch of time at mid-day when the sun's altitude changes very little.

Finally, in this rural community, we have frequent power outages, so when we come back online, I want the array to be able to find out for itself what is the date and time. Using that information, it will go to a look up table where the appropriate array angles will be stored for each day and time, and it will set it's orientation automatically.

Thanks for all the great suggestions. They were all very interesting to read and ponder. In the end, I got my original questions answered, that is, GPS modules cannot be used as inclinometers, but the information they provide can be parsed for that which is desired. Everything has been extremely helpful. Thanks to all. mm

NikosG · 2011-05-21 01:26

Hi mmoreland,
Your idea to store the appropriate array angles for each day and time is correct but it is a litle bit taskwork. A mathematical algorithm could give you
automatically these angles (from sunrise till sunset every half hour or even every minute or every second!!!) . Of course every one of these indivindual angles changes during the year as you noticed. For instance the sun's altitude at 10 a.m. in December is completely different from the sun's altitude at 10 a.m. in August. It is exactly for this reason that a solar thacking algorithm need the Date as an input. The day's duration also constantly changes day by day as well as the time of the rise and set. But it is not also a problem because the argorithm calculates all these parameters having as inputs the date and the latitude.
For one axis tracking system like yours you can isolate the output of the system that concern the altitude and ignore the others that concern the azimuth. The shift of the system's orientation 8 degrees east of south is logical because the magnetic North is different from the geographic North.However I don't understand why "sun's altitude is a trigonometric step removed from the true altitude".
The "frequent power outages" is actually a problem if the clock system doesn't have an autonomus power supply. But I don't think that it is a serious problem if you use a propeller board with a DS1302 chip because they have small consumption of electricity.(a smaller cheap 6V photovoltaic pannel or a small rechargeable battery could help to keep the clock system active for ever )
Finaly after of a power outage it is very easy for the algorithm to calculate the new angles of the Sun and target the new Sun's position.
I like your video and your solar Park. Feel free to ask my any question.

mmoreland · 2011-05-21 12:41

Hi NikosG, I have used a number of mathematical algorithms for calculating the correct angle for the array on a given date at a given time, and what I find is that they all have degrees of error. All are quite accurate in the hours around mid-day, but they fail to accurately represent the sun angle either in the morning or evening, so I have decided to be practical about it and measure the angles as they occur. In the end, I think I'll be able to have a look up table with fewer than 150 items for these reasons. For a month around the winter solstice and again at the summer solstice, the panel need not move at all. At the summer solstice the array will remain at its most horizontal position for four weeks and in winter at its most vertical position. The two sets of five months between the solstices are pretty much mirror images of each other, and it's not so important to have a position correction each day as it is each week, so I imagine I'll have five months (21 weeks) of six movements each day for a week and then another slightly different set of six movements for the next week.

It was a big disappointment to me to discover that the algorithms weren't accurate. Now, it's true that I didn't develop the algorithms myself but found a number of them on the www and tried as many as I could find. I'll try to say what I mean by "a trigonometric step removed from the true altutude," but I apologize if I can't do a very clear job of it. The sun's altutude, as I understand it, is measured as if one were facing the sun. So, just after sunrise, the sun's altitude would be low to the horizon and measured by pointing a device at the sun. If one had a single axis solar array with the axis on an exact east-west line, and if it were the equinox, the sun would rise and set roughly at the end of the axis of the array. As the sun rises, to take measurements of its altitude, one must continue to orient oneself to the direction of the sun's azimuth, so the altitude is related to an observer who can rotate and follow the sun's progression along its azimuth. The array can't rotate, so at sunrise, it's best angle will be horizontal, and as the sun continues to rise it processes toward the south somewhat, and the array will begin its tilt toward the south as well so that it continues to be perpendicular to the sun in one plane. At noon, or sometime around noon, the array, if it has continued to tilt, will be perpendicular to the sun in both the east west plane and in the north south plane, but only for that moment, and only at that moment is the sun's altitude an accurate indicator of, or the compliment of the correct angle for the array. In all other moments during the day, the sun's altitude is not the compliment of the array's best angle, and I have worked out a formula that corrects for that problem, but even my formula is not perfectly correct, so I am taking readings. Today is one of those days that I'm collecting a list of sun/array angles every half hour, on the 6th and 21st of each month. I know this doesn't explain it well. I can only suggest that you take a few measurements yourself. I have a drawing that shows the problem and is the basis of the formula I derived, but that doesn't help here. Thanks very much for your interest, and I read your entries with great interest. mm

NikosG · 2011-05-22 11:17

Ok mmoreland,
from my research I have found some very interesting information. One of them is from University of Nebraska. They have made an excellent simulation with Java applet where you can see the sun's altitude and sun's azimouth any day any time in any place.If you want take a look and tell me if the angles of that simulation are close to your measurments. My robotic version of this simulation is a solar pathfinder. If you want you can send me the latitude of your solar park and I'll send you a table with the sun's angles from my software any day you want.

mmoreland · 2011-05-23 14:29

39.31n; 123.81w. Try @ 11 AM and 4 PM tomorrow, May 24, Tuesday.

NikosG · 2011-05-24 03:09

Well...

according to my algorithm at 24 May in 39.31 latitude (does't matter the longtitude) we'll have the following solar path:

attachment.php?attachmentid=81453&d=1306231037

attachment.php?attachmentid=81453&d=1306231037

.

As you can see at 11 AM the altitude is 67.56 degrees and at 4 PM the altitude is 35.98 degrees
The position of the sun at these two moments are in the red circles (every star represents the sun's position every 1 hour )
Other information: Rise : 4.47 AM, Set: 19:12 PM, Day's duration: 845 min.
What are your measurments?

P.s.
The simulation of Nebraska University as you can see if you follow the link, at 24 May in 39.3 latitude, gives very similar results to my software (at 11 AM it gives altitude 67.9 degrees and at 4 PM it gives altitude 35.9 degrees!!!!)

mmoreland · 2011-05-24 18:42

Hi Nikos G...efharisto! Here's the rub, however. At 11 AM, the array was tilted at 19 degrees (from horizontal), and at 4 PM, the array was at 7 degrees (again from horizontal). Those angles are derived from my moving the array until it is normal to the sun in its east-west plane, and I have an inclinometer that provides the array angle which theoretically should be the compliment of the solar angle. The altitudes don't work because the array isn't facing the sun when the sun is at those altitudes that you gave me. It is, however, facing the sun at noon, and that's when the calculated solar altitude works as the correct complimentary angle for the array. It's a little like why a tapered, four sided box that has sides meeting at 90 degree angles doesn't have 90 degree angles cut on the faces of the mitered sides. There's some trigonometry there. I'll do some research to see if I can find a good, clear explanation. Thank you very much for sending the angles. Michael Moreland

erco · 2011-05-26 13:51

We have it pretty easy tracking the sun these days. I just got back from a trip to Chaco Canyon in New Mexico, ancient ruins from pre-Pueblo people circa 900-1250 AD. My interest in these people was piqued by their intimate knowledge of the movement of the sun and the moon. If you don't know about the "Sun Dagger" on Fajada Butte at Chaco, prepare to be amazed:

http://www.planetquest.org/learn/sundagger.html
http://www.angelfire.com/indie/anna_jones1/fajada_butte.html

This rock formation casts noontime shadows on rock spirals that not only show the date and annual soltice/equinoxes, it also tracks the moon's shadow which shows the moon through its 18.6 year lunar cycle:

http://en.wikipedia.org/wiki/Lunar_standstill

Friggin' amazing knowledge for these ancient people. There's a great TV show (The Mystery of Chaco Canyon) on PBS occasionally that is very interesting and highly recommended.

Same video on Youtube: http://www.youtube.com/watch?v=jwB2fniCWEM&feature=related

NikosG · 2011-05-27 23:43

Erco,
What amazing piece of information!
Ancient civilization had a lot of astronomical knowledge!
Although they hadn't computers they had a mathematical thought and they were able to solve problems like the time calculation and the sun trucking.
The Antikythera Mechanism and the Myriad year clock are two examples of amazing ancient robotic mechanism with this ability!

erco · 2011-05-28 15:53

Yes, Nikos, that Antikythera Mechanism should make you very proud of your Greek heritage! It is fascinating to learn how clever ancient people were and what mastery they achieved.

I just ordered several solar small solar panels & motors (from Ebay China) to make a small self-tracking solar battery charger. Will advise when I get it going. Solar power is going to be increasingly important and I don't want to miss the boat!

I guess that's the slow boat FROM China with my EBay goods aboard...

Tracy Allen · 2011-05-30 22:30

Erco, I was very impressed by Chaco Canyon too, a couple of years ago when I had a chance to drive that rough gravel road that leads in. I hope they have not improved it too much. The vastness of the earth and sky, the exacting fitted stone work in the structures, and the amazing colors in the rock climbing up to and along the plateau made for an inspiring day. I wish we had had time then to stay overnight in the campground to stargaze and to soak in the spirit of the ancient astronomical insight.

Tracy Allen · 2011-05-30 22:50

@mmoreland, To help myself appreciate what you are talking about. I took a book and pointed its flat surface toward a ceiling lamp (the sun). Then rotated the book ~60 degrees around a vertical axis away from the lamp while keeping the book's angle constant to the horizontal. Then, looking back at the lamp, I had to raise the book to a more horizontal position in order to get what would be the optimal solar panel angle.

I found the following formula for incidence angle on a panel...

[SIZE=3]cos(&#952;) =  sin(&#946;) * cos(&#945;[SIZE=1]s[/SIZE][/SIZE][SIZE=3])[/SIZE][SIZE=3]* cos(&#978; - &#978;[SIZE=1]s[/SIZE]) + cos(&#946;) * sin(&#945;[SIZE=1]s[/SIZE][/SIZE][SIZE=3])[/SIZE]

   [SIZE=3]&#952;[/SIZE]:  angle of incidence on panel, [SIZE=3]cos(&#952;)[/SIZE] is the factor that needs to be maximized.
   [SIZE=3]&#946; : [/SIZE]tilt of solar panel off horizontal, this is free to rotate in this installation, to maximize solar incidence.
[SIZE=3]  &#978;[/SIZE] :  azimuth of solar panel axis, this is fixed by this installation at 8° east of direct south.
   [SIZE=3]&#945;[SIZE=1]s [/SIZE][/SIZE]: altitude of the sun above horizon[SIZE=3]
  &#978;[SIZE=1]s[/SIZE][/SIZE]:  azimuth of the sun, measured from north to east

Renewable Energy 32 (2007) 1187–1205.png

equation 28 in:
Alistair B. Sproul, "Derivation of the solar geometric relationships using
vector analysis", Renewable Energy 32 (2007) 1187–1205

The azimuth of the panel is fixed, so the only free parameter is its rotation around the horizontal axis.

How does that compare with other math/trig you have evaluated with poor results?

erco · 2011-05-31 02:01

Tracy: Fear not: they haven't paved 13-mile route CR7950 into Chaco (except perhaps with good intentions) and it was quite a rough ride in our rental car. http://chacoalliance.com/

The only thing that marred our fine Chaco experience was when I noticed we lost a hubcap somewhere. Guess what? We found it driving back out of the park. It fell off on the roughest section of that road. It was in perfect shape; everyone drives so slow there, drivers had plenty of time to avoid hitting it.

tracking solar array

Comments