Dreaded shading and micro-inverters

[PvNewbie]

Hi Guys

Just looking at using SolarEdge for the first time in a difficult shading position but note that MCS states a shading analysis has to be done for each panel. Is this right?

I was wondering how you guys do this? Do you actually (say) produce 16 shading charts for a 4kW system? Isn't it a bit of overkill for the customer/quote?

Thanks

/newb

 

[The Solar King]

This is an issue. How we approach this is to do a normal shading analysis to meet MCS, and then provide an alternative, better and more accurate estimate. The PV guide quotes the following:

"For systems connected to multiple inverters, or a single inverter connected to more than one MPP, it is acceptable to do a separate calculation for each sub array (each array connected to a dedicated MPP tracker)."

and later,

3.7.9 Additional Estimates
Additional estimates may be provided using alternative methodology,including the use of proprietary software packages, but any such estimates must clearly describe and justify the approach taken and the factors used and must not be given greater prominence than the standard MCS estimate. In addition it must be accompanied by a warning stating that it should be treated with caution if it is significantly greater than the result given by the standard method."

It doesn't say each method has to be done on the same basis or have the same equipment configuration. I don't feel due consideration was given to the use of optimisers or microinverters whose use will increase. AC panels will be launched in 2015 which will further muddy the water. Solar edge has been used in 50kW systems. It would be totally unrealistic and laughable to suggest providing 100 or 200 sunpath diagrams.

In our report and quotation to the customer, the following statement is made.

"We are obliged under the terms of the Microgeneration Certification Scheme to provide the following estimate of performance based on a specified manual method. This estimate will be that shown on the registration document for your system. This method may produce inaccurate results with a wider margin of error than that stated for the following reasons:

• It takes insufficient account of geographic location.
• There may be significant variance in irradiance at your location in comparison to that given in the data provided.
• The estimate takes no account of the performance of equipment specified or its ability to mitigate the effects of shading.
• It is prone to further inaccuracy in estimating the effects of shading due to human error.

Professional Performance Estimation by Solar Kingdom
Following the mandatory performance estimate is one prepared using expert system software specific to your location, based on the survey undertaken, and taking full account of the equipment specified. The financial analysis provided is based on this figure. In our opinion this better reflects expected performance and is subject to the standard disclaimer given."

We use PVSol, but you could equally use some other method, such as that from the solaredge software. The important thing is to ensure you have confidence in whatever method you use and can fully justify it to your MCS assessor. (Mine thinks the MCS method is pants as well, but can't say so).

The MCS system is better than that it replaced, is cumbersome and often inaccurate. In parts of NE Scotland it is over 20% out. Gavin A and myself even challenged it at the MCS PV Working Group (without success).

I see numerous quotes from competitors, with no output estimation whatsoever. As long as you can justify what you do, I wouldn't worry too much.

 

[Gavin A]

with a string inverter (without optitrac global peak style shading function) you should take the shading estimate from the worst shaded portion, with solar edge it'd be justified to take it from a rough average shaded position eg the middle of the array and use that to give you a reasonable ball park estimate for MCS.

Then produce a more accurate estimate using specialist software if you're that way inclined.

But explain to the customer that the MCS methodology isn't really designed to cope with individual panel tracking, and is only accurate to within +/- 10% for the shading anyway so should be seen as being a rough guide.

 

[PvNewbie]

Thanks for the responses. I was starting to 'panic' but your comments are reassuring!

SolarKing....AC panels. Can you elaborate? They sound interesting.

/newb

 

[The Solar King]

LG are due to launch AC panels in the U.S. in Quarter 4. They should come to Europe (if U.S. demand can be met) in 2015. These are basically panels with a micro-inverter integrated into them. Have been told on numerous occasions that various panel manufacturers were going to integrate optimisers (solaredge most mentioned) but they have still not materialised

 

[Gavin A]

I'd prefer them none integrated, as integrated ones would need the entire panel replacing if / when the micro inverter fails, which at a 0.2% failure rate is far far higher than the tiny failure rates for the panels.

 

[PvNewbie]

I'm inclined to think this is a technology for technology's sake! Call me old-fashioned but I'd rather a single-point-of-failure inverter, tucked away in the warm and dry!

I'm worried about the direction the panel manufacturers are going. It's great....so long as the technology doesn't fail! These aren't washing machines that can simply be replaced! :frown2:

 

[Worcester]

I'm inclined to think this is a technology for technology's sake! Call me old-fashioned but I'd rather a single-point-of-failure inverter, tucked away in the warm and dry!

I'm worried about the direction the panel manufacturers are going. It's great....so long as the technology doesn't fail! These aren't washing machines that can simply be replaced! :frown2:

Remeber the UK is different from the Rest of the World,
- Spain hop up a ladder, what H&S....
- US and Oz, Step ladder up to the shingles ....
Did you see all the panels on the roof of the new Maracana Stadium?

We are a minor part of the world market.

 

[PvNewbie]

Remeber the UK is different from the Rest of the World,
- Spain hop up a ladder, what H&S....
- US and Oz, Step ladder up to the shingles ....
Did you see all the panels on the roof of the new Maracana Stadium?

We are a minor part of the world market.

...then my preference is still to install single-point-of-failure inverters, so long as manufacturers supply them, unless in a real difficult position, shading-wise.

 

[SolarEdge1]

Re the MCS regulation, we know it is unrealistic to produce 16 shading charts, so our recommendation is to do as The Solar King advised - 1 MCS shading chart + data from additional sources such as SolarEdge Site Designer, PV Sol or PV syst. Additionally, for your convenience, we took the trouble of preparing 16 shading charts in 4 typical scenarios, showing the added energy with SolarEdge in each scenario. Feel free to share this document with your customers: http://www.solaredge.com/files/pdfs/mcs-shade-procedure-se-advantage.pdf

 

[Worcester]

Thanks SolarEdge! note -we do install and specify solaredge as well as string inverters with optimised mpp tracking

We even have two back to back systems at our own premises one with SE and one with an optimised string inverter so this is based upon our experience - I would challenged your claim of 50% reduced output across all the panels in the string inverter layout.

I can see no reason that should be the case - how do you come up with that? The shaded panel, yes will have reduced output - what science / theory are you using to determine that the two unshaded panels will also have a 50% reduced output?

 

[SolarEdge1]

The way it works with traditional string inverter is - they track the Maximum Power Point per string. If one panel is shaded, and is producing, for example, 50% of the energy, the inverter will either bypass this panel completely (0 power from this panel) or it will drop the current of the string so the weakest panel is still producing, but the rest of the panels in the string are also affected, and produce less than their maximum capacity.
With SolarEdge, each power optimiser is tracking the MPP per panel, making sure it is working on its optimised current to produce maximum power, independently of the rest of hte panels in the string.

If you want to learn more, you are invited to attend one of our upcoming training events. We are going on a roadshow June 16-27: Training

 

[Worcester]

I've already attended the 'training' sessions last year, I would still challenge your statement though, as what you have described is MINIUMUM power point tracking, not Maximum Power Point Tracking with optimisation.

Firstly Inverters can;t bypass panels, only bypass diodes will do that in fact in your shaded example deoending upon the design of the byapss diodes in the panel it may be completely bypassed and neither a string or an optimiser can do anything about that.

Lets for the sake of discussion though assume that the bypass diodes aren't triggered, so that one panel (assuming stc) will be outputting 50% of the voltage of normal, the other two at 100% (assume 100v per panel the exact figure doesn't matter) so the inverter will see 250V as Voc and it will then adjust string Voltage parameters up and down to optimise the power (V*A) that it can get out of the string. - that won't be either 150V (all at 50%) 250V (simple sum) or 300V (all at 100%) see here http://files.sma.de/dl/7418/GlobalPeak-TI-en-12.pdf


The SE will do exactly the same on a panel by panel basis.

Why do I challenge this - because I have two setups here which I are constantly monitored, and time and time again we can see no significant benefit of SE over string.

Now, give me a complex roof, multiple orientations, multiple elevations, and it's a different matter.

SE has it's place, I have yet to see any convincing evidence that simple shading on a string gives it an advantage.

Please, prove me wrong
monio

 

[Gavin A]

The diagram is technically correct, but also badly misleading as neither solar edge, nor almost any other manufacturer make string inverters that only handle 3 panels.

Put in a more realistic say 12 panels, and even if the entire panel is bypassed that would mean 33 / 36 strings of cells within the 12 panels would be functioning, so would be operating at 91% of a fully unshaded array.

The solar edge would be getting at most maybe 20% of the power from the shaded panel vs the unshaded, so would have 33 strings of cells at 100% power, 3 strings at 20% power, so would be operating at approximately 93% of an unshaded array.

The solar edge would therefore give a 2% advantage in this scenario.

This assumes that the string inverter has a wide enough MPPT tracking range, or more likely periodic full scan function such as the optitrac global peak, to ensure that the system doesn't get stuck on a false peak, and there may be some additional losses in the first few minutes of shading until the full scan function kicks in to bypass the shading.

Without a full scan function the shaded system on a string inverter could perform very badly, depending on the nature of the shading, if it ends up on a false peak that the MPPT tracking can't cope with then it could ride that false peak all the way down to a 90% reduction in output from the string without bypassing the shading at all. In this situation the solaredge would give a huge advantage, but then a string inverter with full scan function would also give almost as much advantage for a much lower cost.

That's the real comparison, and it's the comparison that neither solaredge, or enphase will produce data for, presumably because they know it would destroy their marketing position.

 

[The Solar King]

I love shading. It sorts the men from the boys and brings out those of the highest technical ability (not me).

We really need a sticky where definitive technical data can be accessed.

As Worcester says SE has its place as do micro-inverters - it is about knowing when to use them. Technology moves so fast compared with other sectors that things may change. There are more micros hitting the market. The Renasolar offering looks interesting and has a 25 year warranty. Cheaper than Enphase but still not cheap.

Shading is a fact of life. Full scan functions are becoming more common. All installers (and customers) need to be more savvy as to what mitigation measures are available. It is time CPD was introduced as part of the MCS standard.

 

[SolarEdge1]

I agree with Gavin A’s analysis. For further reading about the effect of the bypass diodes in shaded conditions, and the benefits of module-level optimisation, read here: http://www.solaredge.com/files/pdfs/se_technical_bypass_diode_effect_in_shading.pdf
Specifically note the effect of bypass diodes in installations with multiple strings connected to a single MPP tracker at the end of the application note.

For a real life example, you can see the performance comparison of SMA vs. SolarEdge in shaded conditions, as tested by a third-party, NREL laboratory in California: http://www.solaredge.com/files/pdfs/performance_of_pv_topologies_under_shaded_conditions.pdf

 

[Earthstore]

Debate is a healthy thing, however I looked at the link (thanks) but it is like trying to compare apples with pears, who in their right mind would use a single string inverter with multiple strings when there is a shading issue?
I do have to agree with Gavin too, his post is very detailed and explains the options very accurately, which also backs up Worcester's post too.

System design is paramount in solar pv and having different options available to obtain the highest performance/financial returns for the customer should always be the main objective, as pointed out Solaredge really does have a place in this, but it is not always the answer if shading is a problem.