Optimal panel array? Shading issue

[Micho]

Hi, newbie DIYer here... please be gentle

What would be the optimal serial/parallel panel arrangement to address what I believe is a shading issue?

I recently installed an interconnected to the grid solar array and am noticing a significant drop in the harvested kWh.

So this is what I have:

• 10 (Ten) 250-watt solar panels (open circuit voltage rated at 35 volts; all of them wired as a single series)
• 1 ThinkPower 2200 watt central inverter with max input of 13 amps and 500 volts

So I had everything installed in September (I'm in the northern hemisphere); during the first days/weeks I was generating as much as 14-15 kWh per day. As year has passed I'm currently generating 5-8 kWh per day. At this rate, I'd be surprised if I'm generating more than 2-4 kWh per day at the peak of the winter.

Here's another piece of information that's important: two of the panels are shaded. When the panels were installed the shading was minimum if any, but as we get closer to the winter I'm thinking the shading could be as much as 40% of these two panels.

I was wondering if I by re-arranging the panels into several series in parallel I could isolate/optimize the array so as to minimize the degrading effect of the shading.

Input greatly appreciated.

Thanks.

[littleharbor]

Not sure how keen you may be with spending more money but if you cant eliminate the shading issue you might consider microinverters or Solar Edge optimizers/inverter options. In a single string configuration it only takes one shaded module to kill your production.

[bcroe]

Another solution might be reduce the unshaded panels to a string of 8, and use micro inverters on those shaded. But
production is always going to be considerably down in the short days of winter; my summer months do about 3 times
the worst winter month. Bruce Roe

[Micho]

Not sure how keen you may be with spending more money but if you cant eliminate the shading issue you might consider microinverters

I'd rather try to work things out with the central inverter I have (given how much I spent)... also, I can't eliminate the source of shading



It is my neighbor's wall that's blocking the sun of the top most panels (the other 6 panels are on the opposite end of the roof and are shade free).

Is there an serial/parallel array that may help to isolate if not diminish the effect of the shading?

Should I consider removing the two shaded panels from the serial string altogether?

Since the max inverter DC input is 13amps and each 250w panel is rated at 35v, it seems another option would be to set two 5-panel strings in parallel (one would be running with no shading, the other would have the 2 shaded panels).

/scratching my head

Please advise.

Thanks

[inetdog]

What is most important to consider is how much of the time and at what part of the day the shading is happening.
A few hours of shade first thing in the morning will not make much of a difference. The bypass diodes will keep those panels from limiting the string current so you just have to worry about whether your string inverter will work at the lower voltage. But you cannot parallel a full string and a shaded string into the same inverter input. The off-noon hour shading will be a small part of your production, so it would not be wise to simply remove the panels.

On the other hand, if the shading (even a small part of the panels) is near solar noon you have a bigger problem.

How many independent MPPT inputs do you have on your inverter(s)?

[littleharbor]

How about raising the racking up enough to eliminate the shading?

[Micho]

...if the shading (even a small part of the panels) is near solar noon you have a bigger problem.

During the summer months the shading is neglible (I'd say maybe 3 months of the year) but as we move to the winter, the shading worsens (I'll know for sure on Dec 24th but I think it'll be like 40% of the surface). The shading is at its worst in the middle of the day (again, not an issue during the summer as the sun falls in straight vertical 90 degrees).

How many independent MPPT inputs do you have on your inverter(s)?

The inverter specs state 1 MPP tracker and 2 strings per MPP tracker.

But you cannot parallel a full string and a shaded string into the same inverter input.

I'm new to this but may I ask why?

Thanks!

[bcroe]

There is no problem connecting a shaded string and an unshaded string of the same Vmp in parallel,
that happens all day long here.

In theory a good MPPT controller will run run a string of 10 as a string of 8 when 2 are shaded. I wouldn't
want to push the bypass diodes in the shaded panels that hard, until confirming they run reasonably cool.
Bruce Roe

[Micho]

How about raising the racking up enough to eliminate the shading?

I can't... I live in a gated community with super strict rules... neighbors complained that my original design was visible from the street and I had to reduce the height (thus the shading).

This was something that I did not take into account when I bought the panels, inverter, etc.

[inetdog]

I'm new to this but may I ask why?

Thanks!

The purpose of an MPP tracker is to set the DC input load characteristics so that the panel string is operating at the combination of current and voltage which produce the most power from the string of panels.
If you have two strings with different Vmp voltages (because they have unequal numbers of panels or one string is shaded and the other is not), then it is impossible to match the Vmp of both parallel strings at the same time.

Rule of thumb: if two strings or panels have Vmp equal within 5% you can parallel them without losing significant power.

In your case if you have, for example, two ten panel strings and you lose the output of two panels of one string, the difference would be 20% and you could lose most of the output of one string. Or if you are lucky only 20% of the output of the unshaded string as well as the loss from the shaded string.

You should study how MPPT works and come back to us with questions.

[solarix]

From the pic it looks like you might have room to rearrange the four panels as a row of four instead of 2 x 2 in order to get out of the shade. Second strategy is to rewire the panels as 2 parallel strings of 5 so that the 2 shaded panels only bottleneck one string. The string inverter might not start with just strings of 5 though.

[ButchDeal]

That is about the worst install possible for you. The best solution ( cheap ) solution would be to raise the lower side of the modules so that they are unshaded. Angel would not be good but better that than the shade.

Better solution would be to do that and add optimizers.

[PNjunction]

Butch - that's a brilliant strategy - mount them horizontal out of the shade - although output won't be perfect, they won't be down for the count.

[Micho]

That is about the worst install possible for you. The best solution ( cheap ) solution would be to raise the lower side of the modules so that they are unshaded. Angle would not be good but better that than the shade.

I agree; this is my best option.

I am going to rotate the panels 90 degrees on the vertical axis and lay them side-by-side and raised away from the shade as much as possible; the angle should remain the same. Hopefully I'll manage to reutilize most of the original aluminum base/framing.

Will report back in a few weeks.
Thank you all for your input.

[ButchDeal]

You are better off with then in landscape, most modules will perform more poorly in portrait with even a little shadow at the bottom edge. Kept them landscape as they are and get longer virticle for the lower edge.

[Micho]

From the pic it looks like you might have room to rearrange the four panels as a row of four instead of 2 x 2 in order to get out of the shade.

Hi, missed your post... yes, this is the first thing I'll try.

Second strategy is to rewire the panels as 2 parallel strings of 5 so that the 2 shaded panels only bottleneck one string. The string inverter might not start with just strings of 5 though.

I'll try this depending on the results. Two strings should yield a DC input to the inverter of 130 to 175 volts which is above the 75 volt power on requirement.

Thanks.

[inetdog]

... Second strategy is to rewire the panels as 2 parallel strings of 5 so that the 2 shaded panels only bottleneck one string.

I would not recommend this at all.
If you do this the voltage of the shaded string would drop low enough that it probably will not contribute at all when placed in parallel with the five unshaded panels.
Since you have only one MPPT input processor, I would, strange though it might seem, put one shaded panel in each string of five.
That way you would get the full production of 8 out of the 10 panels instead of only 5 out of the 10 panels.
(That would actually be the same as putting all ten in series, which would also give 8 out of 10.)

Some panels may not be completely happy running several hours a day through the bypass diodes, but a well designed panel should tolerate that just fine as long as you have good ventilation around the junction box where the diodes are located.

Do not be confused by the common misconception that one shaded panel will reduce the current output of the entire string. It will not do that.