0

I need all info on heliostats (sun following collectors)
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

I need all info on heliostats (sun following collectors)

by virgilshewchuk on Sat Jul 11, 2009 10:06 pm

I am trying to prototype a Heliostat (a device that reflects the sun at an solar collector). I'm using an arduino, servos and a mirror to reflect the sun onto a solar collector. The East West orientation seems easy just move the servo 180 degrees in 12 hours and then back. The seasonal change is a little tougher. I have to tilt the mirror north and south depending on the day of the year. This probably would be buggy if the arduino had to count a year in milliseconds. I would also like to look at the possibility of controlling the mirror with light sensors this would be a more dependable system I think. Any hints would be appreciated.
virgilshewchuk
 
Posts: 2
Joined: Sat Jul 11, 2009 8:35 pm

Re: I need all info on heliostats (sun following collectors)

by John_NY on Sun Jul 12, 2009 3:08 am

The year is too long for milliseconds but you could increment a variable with the day of the year (DOTY). One complication would occur if the arduino reset and forgot the day. Perhaps writing the EEPROM would keep the day of the year in its memory, else, you should be able to input the starting day of the year if you need to restart the device after power cycle. I'd consider an external timekeeping device, which would be cheaper than having your technician reprogramming each heliostat every time you change the battery or lose power.
I did a quick search on "I2C time clock" and found PCA8565. You could probably run it on a separate lithium CR2032 battery or something similar, and the only connection with the arduino would be via the I2C.

I'd assume that you've already gathered the equations for pointing the instrument, and just need Arduino implementation advice? I also have unsolicited hardware implementation advice. If you've positioned your reflector in the very common Altitude-Azimuth configuration (Alt/Az for astronomy-types), the algorithm for combining time-of-day with day-of-year may be a cumbersome since you need to move 2 axes continuously throughout the day (the continual 2-axis re-positioning of your reflector may also be power-inefficient). Try making an equatorial mount stage for your reflector, and changing the mirror's declination* angle to match the solar altitude at noon for the given day of the year. (*declination is the up-down direction when you are on an equatorial mount). You'll only need to change it once a day (or longer), and you'll have one axis to move for the time-of-day. Also, you'll be able to move in evenly spaced servo steps throughout the day. Look up "Alt/Az versus RA/Dec" mounting if this subject is unfamiliar.

You mentioned rotating the servo 180 degrees in 12 hours, so perhaps haven't built your algorithm yet, or perhaps you're using a heliostat more advanced than just a simple reflector that I'm envisioning. If you point a mirror at a light source or laser pointer then tilt the mirror (servo) 45 degrees, you have reflected the light source 90 degrees, not 45 degrees. However, if you were thinking about a mirror mounted on a rod, where the altitude angle of the rod is 90 minus your latitude on Earth, and the rod points north if you are on the northern hemisphere, and where the rod rotates on a servo, and the mirror is tilted to point toward the solar panel, then 180 degrees is appropriate, as it controls the rod, not the mirror.

your friendly neighborhood astronomer,
-John
Last edited by John_NY on Sun Jul 12, 2009 4:51 am, edited 2 times in total.
John_NY
 
Posts: 55
Joined: Sun Jul 05, 2009 2:54 am
Location: SoCal

Re: I need all info on heliostats (sun following collectors)

by John_NY on Sun Jul 12, 2009 3:56 am

by the way -- changing the declination angle of your mirror is not the same as tilting your equatorial mounting stage -- you'll want to have a second axis either manual (for your prototype) or servo-controlled. If you are demonstrating the technology or if you have your solar panels in an easy-to-access place (not on your roof) you could use a manual stage. If it's automatic, you should run the declination-direction servo as a sine function, with a peak at summer and a minimum at winter.

Solar angle at noon on the equinox is declination 90 degrees exactly. You've already used an equatorial stage to mount your heliostat, so you've corrected for your latitude on the earth. With the elevation stage you interested in sun's position as declination angle instead of altitude angle (alt-az algorithms cost extra -- I charge by the hour). Since the change in solar angle is caused by earth tilt, the change in solar angle is a sine function with a period of one year, and the sine function starts on March 21st (day 79, since jan31+feb28+mar21 = 80 and January first is day 0).
There are ~365.25 days of the year, but this heliostat isn't rocket science, so call it 365. Your external clock will tell you what day of the year it is (or you can calculate it).

DOTY is day of the year. Change in declination from winter equinox to summer equinox is 47 degrees, so the amplitude of the sine is 23.5.

Solar Declination = 90deg + 23.5deg * sin((DOTY-79days)/365days * 2 *pi)

You'll need to consider where you are pointing your mirror, but the change in mirror angle should be half the change in solar declination angle.

-John
Last edited by John_NY on Sun Jul 12, 2009 9:30 pm, edited 2 times in total.
John_NY
 
Posts: 55
Joined: Sun Jul 05, 2009 2:54 am
Location: SoCal

Re: I need all info on heliostats (sun following collectors)

by John_NY on Sun Jul 12, 2009 4:10 am

Regarding controlling your servos through photocells, you'd use more power with this approach, but depending on the amount of power you get from your solar cells, it may be worthwhile. The simplest standard method is called a quad-cell. The goal of the quad-cell is to get an even amount of light on all 4 cells. I'd use this for tracking, not for initial positioning, since you could get some interesting problems in your heliostat (it might point your heliostat at the nearest light fixture or a reflection instead of the sun. The quad-cell should be used with a lens and baffle and a suitably dark neutral density (ND) filter placed before the lens (so that the sun hits the filter before it hits the lens). The ND filter should be dark enough to make sure the sun does not fry your photocell over time. If you know where to point the scope at dawn, you can point the quad-cell/baffle assembly at the sun's position, and track it.

I see one complication though -- the quad-cell will point toward the sun, but your mirror only should move half as fast as the sun. The quad-cell will move 180 degrees (at equinox) from sunrise to sunset, but your mirror should only move 90 degrees (unless of course you have the cell mounted on an axis that rotates 180 degrees). The solution then is to position the quad-cell not ON the heliostat, but on your solar cell. Use the input from the solar cell-mounted quad-cell to position the heliostat.
(If you have a rod-mounted mirror and mount the quad-cell there also, then you have one axis of motion and only need a bi-cell, which would be a quad cell where you average the two sides that are in the declination axis of the rod. You can use the quad-cell to rotate the rod with the sun, but cannot use the quad-cell to correct for seasonal motions even if you mount it on the mirror because the mirror has to point halfway between the sun and the solar cell).

Working out the bugs with the heliostat is your job naturally. For example, if you reposition too frequently, all the power from your solar cell will be used to move the heliostat, which means the solar cell isn't very useful.

-John

ps -- it sounds like you have a fun job. Is this for a commercial demonstration or school project?
John_NY
 
Posts: 55
Joined: Sun Jul 05, 2009 2:54 am
Location: SoCal

Re: I need all info on heliostats (sun following collectors)

by redrok on Sun Jul 12, 2009 12:27 pm

Hi All;

Heliostats are very cool.
I've been working with solar heliostats for many years.

Sure, the big guys are often using Vertical axis, (ALT/AZ),
mounts such as Solar II in Barstow CA and many others. See:
http://www.redrok.com/main.htm#practicalsolar
They work well but generally require a processor that can
do the complex, though not hard, calculations to track
them. See:
http://www.redrok.com/analemma.htm#burnett

There are other mounts that work better in small scale:

1. Yes, the Polar axis, (Equatorial), mounts are
generally easy to do but only when the light is
reflected along the polar axis. See:
http://www.solare-bruecke.org/English/s ... fler_e.htm
A second mirror can then redirect the light to the
desired location. These are properly called coelostats.
Pronounced "see low stat". See:
http://www.redrok.com/main.htm#coelostattype1

2. The Receiver Axis or Target Axis mount. There is no
equivalent astronomical mount like this. The main axis
is aimed in the direction you want the light to go and
rotated to align with the sun. The secondary mirror
axis is tilted from the main axis to the bisector angle.
See:
http://www.redrok.com/heliolighting.htm
http://www.redrok.com/images/mrea2005_3.jpg
http://www.redrok.com/images/mrea2005_4.jpg
http://www.redrok.com/images/mrea2005_5.jpg
The tracker for this tracking mount is quite easy to
do and doesn't need a micro. Just a 2 axis sensor
mounted to the 2 to 1 gear assembly. And another:
http://www.redrok.com/electron.htm

In general, permanent magnetic DC motors with very high
gear ratios, 100000/1 or more, are a much better choice
than stepper motors in solar applications. They don't
consume power when not moving. If you need positional
feedback DC servo motors can be used.

Duane

--
Home of the $35 Solar Tracker Receiver
http://www.redrok.com/led3xassm.htm [*]
Powered by \ \ \ //|
Thermonuclear Solar Energy from the Sun / |
Energy (the SUN) \ \ \ / / |
Red Rock Energy \ \ / / |
Duane C. Johnson Designer \ \ / \ / |
1825 Florence St Heliostat,Control,& Mounts |
White Bear Lake, Minnesota === \ / \ |
USA 55110-3364 === \ |
(651)426-4766 use Courier New Font \ |
redrok@redrok.com (my email: address) \ |
http://www.redrok.com (Web site) ===
redrok
 
Posts: 1
Joined: Sun Jul 12, 2009 10:07 am

Re: I need all info on heliostats (sun following collectors)

by virgilshewchuk on Mon Jul 13, 2009 6:10 pm

I want to thank you guys John and Duane. Not only did you give me all this stuff to ponder over you did it on a sunday. Thanks for getting me on the right track. I thought it was a simple task when I started but one look at an analemma blew my mind. I think the trick is to simplify the project. Perhaps a hybrid of sensor and clockwork use the best of both

Virgil
virgilshewchuk
 
Posts: 2
Joined: Sat Jul 11, 2009 8:35 pm

Re: I need all info on heliostats (sun following collectors)

by Brendan_CA on Wed Sep 02, 2009 12:02 am

Hi Virgil,

I'm interested in heliostats also and had decided to do the same project: arduino + servo. Looks like a few of us are trying this one (try some other google searches). I've been an IT guy a long time but have little hands on electronics experience so hoping the learning curve won't be too bad.

I'm interested in an array of them pointing at a single target (consumer version of a solar power tower), so I need a pretty flexible algorithm since each 'stat would have a different position relative to a single target. For this reason, my initial thought was to construct an algorithm that would run externally (i.e., an excel spreadsheet) and generate a long list of yes/no's that my counter would check once per minute (or whatever time increment is a good resolution) -- and then either move the servo a step or not. Call it a 'player piano' or 'music box' approach to tracking. It would allow the microprocessor to ultimately be cheap, but would probably require some external memory. Don't know if that will turn out well or not. However, I'm also wondering about clock accuracy and how code might go so errors tend to cancel themselves out rather than accumulate.

I'm also looking at servocity.com for some of their gear boxes (got em for servos) or pan mounts to save some time. I've got more ideas than practical experience at this point, but I'm hoping to alter that ratio in coming months. Anyway, I'll check back as progress is made.

Hi Duane! I've used redrok many times. Thank you for your site -- a great public service.

Brendan
Brendan_CA
 
Posts: 1
Joined: Tue Sep 01, 2009 11:32 pm

Please be positive and constructive with your questions and comments.