Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.
Alastair Senior Member Joined: 03/04/2017 Location: AustraliaPosts: 161
Posted: 12:01am 25 Jan 2018
Copy link to clipboard
Print this post
A recent tread about multiple strings for solar PV systems prompted me to move my bucket list plan to build a solar monitor up the project list. This links to that thread.
Rather than let that thread get more sidetracked I will outline my ideas here and hope that other knowledgeable folk will keep me sensible. Madness and Warpspeed have already made comments.
My idea is to make a 'simple' monitor that will probably be based on a Micromite to map the solar intensity with time in a way that is useful/relevant to solar PV users and those perhaps make decisions re string directions. It may turn out to be no more useful than just making good visual observations about cloud cover and sun direction at different times of the year, but I like the idea of a combined electronics/mechanical project that is within my skill set.
Initially I was thinking of a small solar panel with its field of view constrained in some known way, mounted so it can be driven by steppers by the MM. I am now thinking that an array of 'focused' panels may be better. Until I do some numbers and test the practicality of defining a panel's field of view I am not sure how well this would work. Having a single panel makes measurement easier than having to do auto sensor calibration but would require all the mechanics. The array idea may become too large and cumbersome to get the required spatial information despite being in principle simpler.
I know there is a large amount of data on the web from good sources but what I am looking at is being able to map local effects. In my own coastal location there are very definite daily cloud patterns which I have observed but suspect that my interpretations re PV may be skewed. I think that it is these local variations that make important differences in panel/string orientations. The best solution is of course to put in as many panels as possible and cover many directions. Many others here have pointed to the cheap availability of s/h panel which makes having large strings quite viable.
In my case and I expect most people, the ability to put large numbers of panels up is constrained by cosmetic factors (neighbours, wives,regulators) so making the best choice is essential. I chose to have 2 strings of 9x285w ea oriented one to the N and the other to the W. This works well and I am happy but intrigued to see if my observations led to the best decision.
So my plan is: 1. troll the web re good local info rather than large area generalizations 2. look for a small panel to be used as a sensor that has similar response as large panels though this can be normalised by calibration against a standard panel which I have several 3. the panel chosen needs to be viable ($) if i decide to use an array rather than a single steerable sensor 4. experiment with focusing a panel and do the numbers re field of view (FOV) for spatial resolution 5. numbers to determine the size required to get useful spatial resolution 6. numbers to decide the required sampling to get useful temporal resolution 7. current guess is that a fov equal to 1 hour of sun transit would be a good compromise. With a map of that against time if one knows the effective collection window of a given panel then good choices could be made. 8. if I use a single sensor then it will have to be incremented continuously throughout daylight for at least a whole year to give the data required. longer to cover yearly variations. This leaves a lot of opportunity for mechanical problems and hence data gaps. An array is preferable from this pov. ? cost - must look.
more to come - have to goEdited by Alastair 2018-01-26Cheers, Alastair
Phil23 Guru Joined: 27/03/2016 Location: AustraliaPosts: 1664
Posted: 10:42am 26 Jan 2018
Copy link to clipboard
Print this post
I've got a small garden light panel on the roof to measure energy for one of my projects.
It's loaded down by a 10 Ohm resistor & straight into the Micromite.
It's calibration is not too bad, compared to the Pyranometer I have. Under 5% out in the 800-1200W range, gets worse as the level lowers & could be corrected in code, but it does the job.
If you are going to use a MM, how would a couple of Servo motors be? You can drive them directly from basic.
Phil.
Alastair Senior Member Joined: 03/04/2017 Location: AustraliaPosts: 161
Posted: 11:22pm 26 Jan 2018
Copy link to clipboard
Print this post
@Phil, My typical behaviour when I get an idea is to think about how I can do it and then charge off and order bits. By the time they arrive I have done more thinking and thought of other ways or possibilities. This has led me to having a nice stock of bits that usually get used for other projects. I am sure nobody else behaves like that.
This time I thought I would be smarter and do more thinking before buying.
I have no problem using servos or steppers with a single sensor but given I am trying to create a 'sky map' of the solar intensity with time I thought that perhaps it may be better to create an array that would sense a large fov at a time. This could then be stepped less often or perhaps only along one axis if it was linear.
This morning I was looking at one large panel I have and realised that perhaps I could access each cell by drilling through the plastic coating and then make a simple collimator to go on the front to given directional selectivity. Something like the interlocking dividers in carton of bottles. Need to do the numbers to see if this is sensible in terms of sampling.
Interesting what you say about response vs your meter. I assume, but don't know, that the small panels may generally be different to the larger ones but for my purposes a simple correction factor should be enough. This is actually in part why segmenting a larger panel is attractive. More samples at a time and better equivalence. Edited by Alastair 2018-01-28Cheers, Alastair
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 01:26am 27 Jan 2018
Copy link to clipboard
Print this post
Easiest way to build an accurate tracker is to do it the way the astronomers set up their telescopes to counteract the rotation of the earth.
All you need is "something" that rotates exactly one revolution in 24 hours, installed in exact parallel with the earths rotational axis. Its called an equatorial mount in astronomer language.
The rotating part can be a simple low cost 24 hour clock driven from a mains synchronous motor. It will always point to where the sun is even at night. It beats a sun seeking tracker because it still works in cloudy conditions.
That idea could be modified to track forwards for exactly twelve hours, and reverse for exactly twelve hours using a stepper motor. THat way your wiring does not twist up. I gathered a large gearbox, and all the mechanical parts to build a large tracker, but like many projects it was never finished. I was planning to use a second stepper and gearbox with a crank and long connecting rod that made one rotation every 365 days. That would cause the solar panels to nod up and down with the seasons.
Still have all the parts here, but solar panels are now so very cheap, and the fixed virtual tracker I have now is much simpler and is perfectly adequate.
Cheers, Tony.
Alastair Senior Member Joined: 03/04/2017 Location: AustraliaPosts: 161
Posted: 01:47am 27 Jan 2018
Copy link to clipboard
Print this post
Ahh but what I am planning to do is scan the sky continuously with an effective aperture of approx 1hr so that I build up a map for my location that shows at any time of the day and year what the solar intensities are. This then allows one to make a decision on the 'best' orientations for strings. Total overkill and probably a waste of time but ....
Agree with what you say re tracking but that is not what I am doing.
My playing with a panel at the moment suggests that I can possibly make a multi-collimated sensor from a 20W panel I have. The risk is that I stuff the panel but it did not cost a lot.
Got to do the numbers ... Cheers, Alastair
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 02:11am 27 Jan 2018
Copy link to clipboard
Print this post
I think you will find that orientation is not as critical as you might be thinking. It depends only on how much panel area the sun can actually "see" (in direct full sunlight).
As the sun angle varies, there is a change in perspective. Directly in front the sun sees the full physical area. Completely side on the sun sees nothing, just the non functioning edge of the panel.
The sun angle needs to shift quite a lot before output falls significantly.
Any shading is pretty lethal of course, and at sunrise and sunset you are not likely to see the real horizon unless you are looking out over the ocean.
In full cloud, sunlight defuses into a dull even grey, where its not really possible to see where the sun actually is. fairly typical in mid winter, in which case you can face the panels in just about any direction, including due south. Facing straight up probably being very slightly better in that situation.
I think it was George that asked why I had my east/west facing virtual tracker panels mounted at the very steep angle of sixty degrees. This is why. You can get an almost constant output all day long sunrise to sunset.
Edited by Warpspeed 2018-01-28Cheers, Tony.
johnmc Senior Member Joined: 21/01/2011 Location: AustraliaPosts: 282
Posted: 03:44am 27 Jan 2018
Copy link to clipboard
Print this post
Good day All
I have 3, 2kw arrays that move 15 degrees every hour to roughly track the sun between 9am and 6pm (due to position restraints) the only benefit is the first 1 hour in the morning and 1 hour in the evening this give me a net gain of about 4-5 kwh per day. During the summer we use 25 to 35 kwh per day for house water pumps and aircons (when the sun shines which is most days) As I add more solar panels they will all be fixed.
john
johnmc
George65 Guru Joined: 18/09/2017 Location: AustraliaPosts: 308
Posted: 04:26am 27 Jan 2018
Copy link to clipboard
Print this post
If this is an exercise you want to do for fun and because you enjoy playing with it, fair enough. It's going to be a long exercise though because any cloudy days will mean you have to wait a year to get that days data. If it's cloudy again that day....
If you are doing it for practical purposes, I think you will find as said, the angles and orientations are not as critical as what is often made out.
I have been crunching numbers ( a LOT of Numbers) for my own setup. Pretty much flew in the face of everything I had read and is prescribed by solar websites and " authourities".
I wrote a piece with facts and figures on another site as to the difference between different orientations and angles. I used a 5 KW array as an example. In so many instances such as putting panels west instead of north or at differing degrees according to lattitude, any shortfall was made up buy adding 1 or 2 panels to the array over the course of a year.
Roof angle not " ideal? Well in many cases as I found that could give you more total generation but any shortfall would be made up by adding 1 or 2 panels the same size. I called it add 2 panels to your 10 panel array and still calling it a 10 panel array and it will be the most efficent array going within a very wide range of paramaters.
Same with tracking. You cannot get a tracker for the price of 2 panels, especially if they are used panels as I buy. Slap the things on the roof orientated close to north as you can get, put a couple of extra panels in and you will outdo a tracker every time. Tilt frame , the same. Leave them fixed, couple of extra panels thrown in for good measure and you have every bit as good efficiency far cheaper and with less work , effort and potential problems.
Not saying a couple of extra panels in an array facing due south at 90o is going to make up for the problems with that setup but certainly for things like roof angle or less than perfect orientation, sure goes a heck of a long way.
And the one thing where more Cubic ( or Square) inches rules supreme again is when you really need it.... Cloudy weather. Nothing beats more panels when the sun is shining but glowing dimly. That's when the more panel option really comes into it's own and no efficency improving strategy can match that.
Anyway, if you are doing it for fun and learning, hope you keep us posted. If it's to optomise your system, I think you'll be chasing your tail when you can catch it very cheap with a couple of extra panels thrown in the array.
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.
Print this post
