Here is my hypothesis:
I love my trackers even though they have been trouble to keep running. So much trouble, I tend to recommend "NO TRACKERS" for most of my customers. So how are more permanent mount panels added to the best advantage? Panel output can be graphed as a parabolic curve. Years ago the belief about array direction to Solar South was the best way to maximize KWH output. Now we know that various degrees of change from South only move or shift the parabolic curve and unless to extreme, do little to influence KWH output. Over loading extra nameplate power on an inverter unit is good to a point. Adding another inverter and more panels to gain KWH may have some problems, site specific. Off grid more noon power is often unusable. My system usually fully charges the batteries by 10:30 am, rendering all but useless the extra panels except for a few weather events a year. The maximum current allowed into a grid also may be a factor in maximum PV nameplate power all facing one direction. So here is where my experiment starts: I believe that the output of two or three arrays can be peaked to system maximum usage by aiming one array towards Winter rising Sun and one array towards Winter setting Sun, and possibly one more array pointing towards Solar South. If we look at the output by plotting the parabolic graphs, you can see we flatten the power curve away from a large peak power at the noon window, and gain power at earlier and later times of the day. Basically with this arrangement of arrays we have full Sun on some part of our arrays at all times of the day. A thing that, through constant observation over several years, is a really good thing. Especially for off grid systems.
So this, (at one time frowned upon), idea, may have a great deal of value to expanding a system power to compensate for trouble with trackers. This arrangement of differing array positions has zero affect on power available during cloudy conditions. In fact may improve output when the Sun peaks out at different times of the day, early and late.
We have moved our trackers and turned them off. One facing the rising Sun on Dec.21, and one facing the setting Sun on Dec. 21. And a zenith suitable for the current yearly Sun angles. We plan on recording the output for a while and move one tracker to Solar South again and compare it's output to the array pointing East or West of Solar South. We will create a new page in our Sunny Portal to compare outputs of our two tracking arrays set to different positions. And if our data suggests a sharper output at morning and evening, and a flatter more usable power peak during noon, we will build our next off grid system, currently under contract, to these findings.
No one can argue the fact that if you are off grid, reducing the night cycle by increasing the duration of generation is a good thing. And a fringe benefit is creating a more usable power curve at the same time. All this by positioning more than one array for maximum usage, not maximum power at the noon window. This may have already been done. I don't claim to be first. I just want to know scientifically, for a fact, it works. AND...I am leaving out particular power harvest data because 1.)It is so variable day by day and short term testing will not give us a viable model and 2.)It will eventually predictably behave to revalidate our test towards off grid usage.
Below is a baseline graph of our two trackers fully operational during a cloud free day.
Below is a graph of the trackers set to where the sun rises and sets on Winter solstice. Clouds fowled the first two days of tests.
Below are three graphs respectively showing clouds coming in heavy around 11:00, 14:00, and the third partly clearing at 13:00.
Below is the first good day of sun. Perfect shift demonstrated. Settings are 145* and 220*. I want more separation than this so I moved the arrays this evening to 130* and 240*. I am looking for significant shift to graph a lower crossover at 13:30. I am very surprised how far askew we need to shift to lower noon generation.
Below is the first day at more extreme skew. Clouds affected morning gen just a little. Slight lower gen peak due to atmosphere differences day to day. The steep drop off on our afternoon gen is due to our being in the bottom of a deep N-S valley and loosing evening gen early.
Below is another good day of sun to plot good graphs for skewed arrays at 130* and 240*
Below represents our arrays pointing solar South with Zenith set perp to solar noon sun. Atmospheric conditions alone influenced greater magnitude in graph 5/24.
And just for fun, here again is graph from dual axis tracking.
This dramatically shows the wider power harvest available to off grid systems by skewing two arrays in opposing directions away from solar South. Site specific shading factors will play a significant role in determining final alignment. This can be seen by understanding the relationship of early evening shade to our pair of trackers and the greater generation diversity as we skew towards East as opposed to towards West.
Here is what I am trying to prove acceptable for off grid systems...Wednesday, May 15, 2013 7:00am. I have 55W power on an array of 3200W nameplate facing solar South. I have 380 watts power on our stationary 1594206201 pointing 130*, the red inverter on graph, with 1560W nameplate. All newest information made available to me by SMA training library indicates no significant power losses by array alignments in this range, and the two arrays could be separate strings going into a 1 channel inverter, not really needing two inverters or a 2 channel inverter. I believe this early & late available power can shorten the night cycle and there by the depth of discharge per autonomous night cycle, gaining battery bank lifespan and many other positive power storage factors. Is this a good practice on grid tie applications? I wouldn't do it intentionally, but it won't interfere with total power harvest more than a couple percent. The factor is, off grid you only can use what power you need not what you can harvest. Very different from grid tied. Off grid, earliest morning and latest evening has more value than a great deal of generation at solar noon. Our graphs suggest we don't really loose that much at solar noon anyway. COOL!
FYI, If you reached this paragraph, you may be a solar geek! So in conclusion, very little drop in solar noon generation was observed. A great deal of early generation was gained. Little change to evening generation due to shade, with some loss in overall array power harvest due to shading of exactly the area of sky where late generation could be gained.
This indicates application of skewed arrays may best be used in off grid systems. Intentions to widen and lower energy harvest curves to increase array size per inverter on grid tied systems may not benefit due to less of a drop in solar noon amperage than expected.
And here is a photo of two arrays skewed to mimic tracking for an off grid system.
For more photos of this system click here
Copyright © 2007