Optimisers for solar panel shading

[rthorntn]

Hi,

My setup is 20 normal DC panels in two equal strings going to normal inverters, no microinverters or anything like that.

Coming into March in Sydney I've noticed bad shading coming from the council's eucalyptus in the street up till 11am. It's badly obscuring 3 panels on the string of 10 nearest the street.

I can see my 10 panel string furthest from the street putting out 1800w at 10:20am and the shaded string of 10 only doing 270w.

Now I assume as the sun gets lower in the coming months that this shading is going to get worse until like September..

I'd like to mitigate this somewhat, I don't believe council would take metres of healthy branches off the top of the tree.

So I'm thinking optimisers, the installer uses Tigo but says "what optimisers can do is to bypass those shaded panels and let the rest in one string produce normal current. So the improvement with optimisers is very limited"

Is it really as simple as just bypassing panels, for some reason I thought it was more complex?

Any thoughts, experience?

Thanks
Richard

[thm_w]

https://diysolarforum.com/threads/do-bypass-diodes-in-solar-panels-eliminate-need-for-micro-inverters-or-dc-optimizers.48241/
https://help.aurorasolar.com/hc/en-us/articles/115001339907-Global-MPPT-and-Bypass-Diodes

[bdunham7]

Optimizers should do a bit better than simple panel bypass, so if 3 out of 10 panels are shaded and the others have full sun, then you will see better than 70% output.  How much better depends on how much of the panel is shaded and how many bypass diodes it has. 

Properly configured modern string inverters should also do a lot better than the results you are seeing.  Find out if your inverters have a global MPPT function and if it is turned on.  The panels should have bypass diodes (typically 3) so that partial shading results in less power reduction.

[Siwastaja]

This is not normal; if 70% of your panels are not shaded at all, you should see just 30% drop, not 85%. Something's wrong with your inverter. Maybe you have a seriously misdimensioned inverter with too high minimum MPPT voltage, or an inverter with broken MPPT algorithm.

I have kinda similar situation myself because someone thought it's a good idea to use a 5kW inverter designed for ~16 panels and only buy 10 panels. Normal string inverters can deal with ~half shaded panels just fine (and if more than half are shaded, chainsaw is better solution anyway). But in my case, 4 shaded panels means I have only 6 panels producing, which is below the minimum MPPT voltage of the designed-for-16-panels inverter.

But you need to check the inverter specs to be sure (minimum MPPT voltage).

[rthorntn]

I have it from a reliable source that the MPPT algorithm in these Voltronic inverters (with the high PV array voltage) is rubbish, and often gets stuck in a local maximum that is not the global maximum, under shading...

Georg594 in Germany has been patching his MPPT code to address shading in this thread https://powerforum.co.za/topic/15621-axpert-max-8-kw-shadow-management-firmware-upgrade-beta-still-based-on-stable-9006/. I think he ported it to at least one other machine.

So I'm a bit stuck.  I've attached a picture (taken today at 10:36am) of the panel shading as visible from my neighbours wall, at least three panels look shaded to me, one is off to the left behind the tree.

I can see that this string performing at about 25% of the string on the right, the panels and inverters are all the same.  The good string is putting out 160w per panel

As the sun gets round the tree and more directly to the panels I get pretty amazing performance in summer, like over 9KW on 7.4KW of panels.

A thought...could the panel orientation affect the bypass diodes with the shade coming in this way, my panels have 3 bypass diodes?

Cheers
Richard

[fourtytwo42]

Could do with a clear picture of your whole installation with the dividing line between the strings marked as it's difficult to see much from the above.
Short answer is yes orientation does matter as the diodes normally divide the panel into 3 vertical strips when using portrait.
I guess this was "professionally" installed so you have little detailed knowledge of the system specifications, you need to ID the panels so you can get the specs for them and ID the inverter so you can check the spec for it. What you are looking for is the MPPV of the panels and the MPP range of the inverter to see if when several panels are shorted out the remaining MPP panel voltage is still withing the MPP range of the inverter.
Assuming all that pans out then the problem is likely as someone described the inverter is not suitable for shading and it's algorithm is not finding the true MPP but is locked into some lower power point. IMOP optimizers are a very expensive way of fixing what is essentially just a software problem.

[Geoff-AU]

There is nothing to be gained by microinverters or optimisers in raw performance terms.

If you have an inverter with 2 x MPPT inputs and you need to spread panels across 3 x roof aspects then a microinverter system might be cheaper than a second string inverter.  Aside from that, strings work perfectly well regardless of the amount of shading.

And then there's the whole "high voltage DC will kick your dog and burn your house down" myth.  Yes, the DC isolators were a poorly conceived idea.  But the new wiring rules (sans isolator) are excellent.

[bdunham7]

Aside from that, strings work perfectly well regardless of the amount of shading.

Except for when they don't, as demonstrated in the OPs first post.

And then there's the whole "high voltage DC will kick your dog and burn your house down" myth. 

Again, except for when they do burn your house down.  It has happened, it's not a myth.

Yes, if you have a properly set up Global MPPT string system with properly installed good quality connections, neither issue is likely to cause you any grief.  But those are two 'ifs' and no matter how dismissive you'd like to be, large power reductions due to minor shading and electrical arcing fires only happen in string systems, at least AFAIK. 

[nctnico]

What are the specs for the inverters? model number?
It could make sense to replace the inverters. The ones from Growatt are surprisingly cheap.

[Siwastaja]

Except for when they don't, as demonstrated in the OPs first post.

With the exact same logic, optimizers and microinverters do not work at all if you find a single example where they do not work (and you'll find).

Cars also don't work if one car breaks down, so we should use horses instead.

Demonstration of an implementation failing does not prove the technology failing. And this is no nitpick; most of the world's solar installations use string inverters and they have excellent overall track record.

(Oh I hate intellectual dishonesty and discussions where agenda overrules technical understanding. As end result, problems do not get fixed, but random technical choices are done until the solution works by luck, or someone gives up. Since the false FUD campaigns of SolarEdge and others causing such large damage to collective intellectual understanding, the least we can do is to not spread the same old myths.)

[bdunham7]

(Oh I hate intellectual dishonesty and discussions where agenda overrules technical understanding. As end result, problems do not get fixed, but random technical choices are done until the solution works by luck, or someone gives up. Since the false FUD campaigns of SolarEdge and others causing such large damage to collective intellectual understanding, the least we can do is to not spread the same old myths.)

Then you won't mind me pointing out the fallacy or misstatement in your critique:

With the exact same logic, optimizers and microinverters do not work at all if you find a single example where they do not work (and you'll find).

I've never said that they don't work at all, nor that one issue with one product condemns them all.   What I am pointing out is that simply because a theoretical solution to a problem exists, even if widely implemented, it doesn't follow that the solution has always been implemented.  I believe you have actually posted that your own string system does not track properly because the inverter is oversized.  The OP clearly has a string system that isn't working ideally.  Shading and DC-arc fires do happen and are almost entirely associated with string systems.  Neither your issue nor the OP's nor most solar system fires would have happened with micros.  Do you deny any of that?

As far as spreading FUD and 'old myths', I pointed out in my very first reply that a properly designed string system and an inverter with global MPPT would likely not have this issue.  Would your 'intellectual honesty' require me to suppress actual facts and examples of real systems that have these problems, however rare you think they are, in service of some mandated truth or forced consensus? 

I think the most you could possibly say and maintain any claim to intellectual honesty, even in response to what you consider false and misleading advertising, is something like this:

First, acknowledge that optimizers and micronverters do fully solve shading issues and further acknowledge that microinverters mitigate DC-arc fire hazards completely or nearly so.  Or, if you disagree, provide evidence to the contrary.  Then you may claim that properly implemented DC string systems with properly sized and operable global MPPT can acheive shading peformance as good or nearlys good as optimizers/micros, provided you don't have extremely heavy shading.  And you may also claim that properly installed DC strings that use properly matching connectors of a type that have not been shown to be defective are fairly unlikely to start a fire. 

But if you are going to be honest, you have to also acknowledge that there are a certain number of DC-string systems that don't meet the stated criteria.  Like yours, the OP's and a whole bunch of commercial installations on WalMart roofs.  Your claim that DC string systems are common and have a good overall record has nothing to do with discussing a particular failure mode or example.  Understanding these issues helps us all avoid them, pretending they don't exist to support a 'correct' narrative does not. 

b/t/w, just as an aside, I've seen a fair number of solar installations that are not working or not working properly.  It's not rare.  To the extent that newer ones might be better, it is because lessons have been learned, some of them along the lines of this discussion. 

 

[nctnico]

Part of the discussion is also price. If I look at the Enphase micro inverters (MQ7 / MQ8) then I notice that 1) they cost almost the same as an entire solarpanel 2) they can't deal with a modern day panel running at full power. Even with the reasonably low prices of the solar panels, the time the system starts to pay itself back is very long  (assuming you have the system installed). I agree DC strings are not the most ideal solution from a technical POV and need attention when being installed but they are about the only choice if you want to have a ROI at all.

[Geoff-AU]

All modern string inverters are within 2-3% of microinverters under any shading condition.