7.6kW vs 8.2kW Inverter on a 9.9kW PV

[Victor Siena]

First of all, please bear with me if this has been a topic in the past. I'd be glad if someone would point me to the right threads if that is the case. I'm new to solar systems but I'm trying to learn quickly to fully understand the solutions presented to me before I decide how to move forward.

I'm evaluating two proposals for a 10kW system:

Solution 1 1. 30 330W Hanwha cells with 7.6kW SolarEdge inverter with optimizers, which claims to be able to produce 10,700 kWh per year in my location (burlington, on);

Solution 2. 32 300W Hanwha cells with 8.2kW Fronius inverter, claiming 9800kWh annual output (lower cost)

My question is, why would the second solution use an 8.2 kW when a 7.6kW inverter with optimized input can actually produce a higher output?

My view is that if a 7.6kW inverter with optimizers can produce 10700kWh, then I think the same output can be achieved without using optimizers by appropriately increasing the number of panels and using a 7.6kWh inverter. That would cost lower than the first option and yet closely maintain the cost of the original solution 2 (assuming I don't put value on the other features of optimizers such as panel-level monitoring, etc.).

There is obviously a big gap in my understanding. I really appreciate any thoughts. My personal benchmark is I should be able to explain the proposed design myself as if it's my own, only then will I decide to put my money on it. I trying to get there :)

[Victor Siena]

I have learned a little a bit more since I posted the above a few hours ago.

I understand now that the reason the provider cannot specify a certain inverter matching the exact requirements is because of logistics reasons; example, in the second solution above, the provider cannot use a 7.6kW inverter without optimizers because the manufacturer of the "non-optimizer" inverter brand they carry does not have a 7.6kw model, so they up to the closest which in the above case is 8.2kw.

I'm inclined towards the optimized solution so I'm trying to understand it a bit more.

I understand now that there will be two planes in my roof, on one side they propose to put up 15 330W panels (130 deg az), and another 15 on another side (220 deg. az.). I learned about PVwatts and I tried to model this installation. My parameters are not very accurate (example, I determined roof pitch by overly basic squaring with my thumb and index finger ;) but I noticed the variations in values do not result to a sea change in output so I will tolerate.

1. For plane 1, peak output would be 3.955kW
2. For plane 2, peak output would be 3.84kW

That gives me a total of 7.8kW, which exceeds the rating of the Solaredge (7.6kW). I assume that would result to marginal clipping.

Is this the norm in inverter dimensioning?

In the industry where I come from, we size standalone systems so peak utilization is 60% of rated capacity for various good reasons. One reason is from our studies, we were able to show that as utilization approaches rated capacity, the mean time between failure (mtbf) decreases. In redundant systems with failover, we would size that to 40%. If I apply this rule, this means that my inverter should be 13kW (7.8kW/0.6) rounded to the closest available model of 10kW for SolarEdge.

Does that make sense in the solar system dimensioning perspective? My line of thinking is that if I go for 7.6kW inverter, then I have to accept the marginal clipping and negotiate the 20yr extended warranty since I expect MTBF will be lower.

Thoughts?

[Rob Beckers]

Hi Victor,

The same panels, with the same orientation, will not produce very differently with or without optimizers. Given this the SolarEdge solution would produce about the same amount of energy vs. the Fronius string inverter solution, anything else is just marketing. I understand the Fronius solution has 300 Watt less and there will be a penalty for that, all I want to point out is that SolarEdge doesn't improve energy yield vs. string inverters.

There can be good reasons for going with optimizers or micro-inverters (by the way, despite the advertising shading is generally not one of them): If you have many roof surfaces that will have a few panels then optimizers make sense. They also void the need for rapid-shutdown and squirrel guard, saving some money there. Though SolarEdge is a very expensive solution. Panel-level monitoring is nice to have.

There are downsides to optimizers or micro-inverters too: You're putting complex electronics behind every panel in a very hostile environment, at least one will fail over the next 20 years (I can pretty well guarantee that). You're buying into a sealed universe; the broken optimizer can only be replaced by another SolarEdge and you hope that company will be around when you need it. Then there's the expense of fixing a broken optimizer or micro-inverter, even if you get a free replacement (pulling panels, and putting them back, lots of labour). Cost as mentioned is another downside.

PVWatts is pretty good at predicting annual energy yield, I use it quite a bit. Be aware that it is a little optimistic with its default efficiency settings, and it doesn't account for snow cover in winter. Still, it's a good place to start!

Regarding inverter size: PV modules very rarely produce rated output, you'll only see that twice a year (the rating is at a very high light intensity and very low panel temperature). It's perfectly acceptable to oversize the PV side 115% - 120% of rated output of the inverter, in practice that will rarely cause clipping and not affect energy yield. Beyond that you get into diminishing returns, as clipping will start to weigh in more and more.

It's not considered a problem to size the inverter the same as the panels (or even oversize the panels some). Solar is not 100% duty cycle, there's no production at night, and even during the day peak output is only around noon. Much of the time the inverter will be at partial power.

That said, we would have sized 9.9kW of panels simply with a 10kW inverter. The savings with a smaller inverter are marginal...

-RoB-

[Victor Siena]

Thanks for the comments Rob.

One of the reason I was leaning towards Solaredge was from my vendor's design, they have one string with some panels on different planes (130az which is smaller plane max 18 panels and 220az up to 24 panels). I thought optimizers were intended to address that scenario.. But I've been modeling in PVWatts various configurations putting all panels in same string on same plane and I find that desired output can still be achieved under the same total number of panels. So I guess by doing that the value of having optimizers is diminished.

I'm curious why there's more need of squirrel guard in the Fronius solution. Can you help elaborate?

Looking at Fronius data sheet, I see that there's two MPPT inputs. Pardon my elementary question, but is it correct to say that:

a) I can have different input voltages to each MPPT and that
b) the power to each MPPT is additive (ie, Total power=MPPT1+MPPT2)?

Been playing with the Fronius configurator, and I like the fact that stringing options is much more flexible compared to Solaredge (mainly due to large operating voltage).

At this point, I'm trying to understand more deeply the Fronius solution, and hopefully I can make a final decision by end of week which way to go.

[Rob Beckers]

Victor, panels mounted at 130 azimuth vs. 220 degrees will see very different amounts of light at the same time of day. PVWatts may show the same energy yield, it says nothing on what time of day that's made. Panels pointing in different directions should be connected as individual strings, into their own MPPT input on the inverter. That's why optimizers and micro-inverters are handy, they do MPPT for every individual panel, so you can connect panels on various roof surfaces together in a string.

Micro-inverters and optimizers have ground-fault protection (and possibly arc-fault) at the module level. Code doesn't require squirrel protection in that case.

Rapid-shutdown has to do with de-energizing the wiring coming off the roof in case of emergency (fire etc., this was introduced to improve safety for fire fighters). Micro-inverters and SolarEdge have this build-in; when the 240V AC is cut the optimizers cut their output to 1 Volt for SolarEdge. So no need for rapid-shutdown. For Fronius that requires a separate box that's normally put on a panel on the roof, so when power to the inverter is cut it in turns cuts the Voltage coming down the wires.

Regarding multiple MPPT inputs on string inverters (not just Fronius): Yes, each input does its own thing and can have vastly different numbers of panels. Total output is the sum of the inputs.

-RoB-

[Victor Siena]

Thanks a lot, Rob. You have clarified a lot for me.

I thought the squirrel guard is only to prevent squirrels from nesting. I now understand it's also to protect them. I'm learning :)

I think I now have an idea how to move forward.

I've built my own configurator spreadsheet for different conditions based on various scenarios and while not perfect in assumptions etc., I'm glad that I was able to approximately match the values of my vendors' designs. My goal was to be able to reverse engineer their designs so I can better evaluate it.

With your help and by reading the hundreds of threads here and the articles in solacity, I'm glad to say I'm at a comfortable place to make a judgment now :)

I'll update once things are in place. This might as well be my journal towards going green ;)

[Rob Beckers]

I thought the squirrel guard is only to prevent squirrels from nesting. I now understand it's also to protect them. I'm learning :)

I don't think anyone is worried about the squirrels, but they seem to have caused at least one house fire chewing through PV wires on the roof. After that Canadian electrical code mandated squirrel guard. Trouble is that it's a blunt rule: It's mandated everywhere, even places that have no squirrels (and it's painful to install, expensive, ugly...).

Don't get me wrong, there are certainly places that NEED squirrel protection. Just that there are many more where it is not doing anything useful.

-RoB-

[Kill Switch]

Hi,
I notice there was a comment regarding the common practice of over supplying power to the inverter.
My question is how far can I push this? My goal is to maximize my shoulder output.
I have an array of 4 x 15 panels mounted on posts(13.86 kWt) plus another 9 kWt on the garage roof. The 13.86 array has been running for 6 years with minimal problems. I just added the 9 kWt but have not hooked it up.
I have a pair of 5 kWt inverters ea with 2 MPPT inputs with an input power rating of 4.4 kWt. I can aim the pole array east and west to produce more power at shoulder time, and connect these to an input, and split the 9 kWt output between the remaining 2 MPPT inputs.
My reasoning is there will be attenuated output from the roof in the morning and afternoon, which will be compensated for by increased power from the reoriented east and west post arrays. The east array will produce very little after noon, while the west will produce little before noon.
Does this seem reasonable? or am I overlooking something?
Cheers,
Kill Switch

[Kill Switch]

BTW do I need squirrel protection on the 4 post mounted arrays?

[Rob Beckers]

BTW do I need squirrel protection on the 4 post mounted arrays?

No, ground mounted solar PV doesn't require squirrel protection. Depending where you are, there could be other requirements though. Ontario has lots of code about ground mounted solar, but other provinces don't (yet).

Regarding oversizing the DC input on inverters: It depends on the inverter. Most will handle fairly large oversizing, I've seen 180% of rated, but there is a limit. At the end of the day the inverter's processor (and other hardware) has to be fast enough to dial back the current by moving the power point up in Voltage, and keep it within safe limits. No idea where that limit is though, and it'll vary from brand to brand.

-RoB-

[Kill Switch]

Thanks for the reply, I will try angling the 4 arrays and monitor the input power. I can also put in some fuses to limit the current though that isn't a power limit. Do inverters give any clues as to their upcoming demise? Besides letting the smoke out?
Cheers,
Kill Switch

[Rob Beckers]

I don't think fuses or breakers are going to save you, things will happen much faster if they go wrong!

-RoB-

[Kill Switch]

Thanks for the comment, It makes me smile. I have a friends that put 14700 watts on his two 5 kWt inverters and another that put on16100, both without issue. Both have shoulder input that is low compared to a tracking array, and i thought this could bring me closer to a tracker with about a third of the cost and less complexity. Panels are so cheap now, I paid $0.65 per watt so the temptation to bump up production is high. If an inverter does fail, they are now about 1/2 the price I paid 6 years ago and 2 months production would more than pay for replacements.
I will let it run for a year and see what I get. hopefully the smoke doesn't leak out.
Kill Switch:

[Rob Beckers]

Yup! As mentioned, I've seen 18kW on a 10kW set of inverters (2x5kW, so 9kW on each). Worked fine at the time. It's been a few years and I've not heard from the installer. Presumably they're still working though. There's no big risk to the inverter as long as it's fast enough in controlling the input current.

Fully agree with you on adding more panels vs. a tracker. Even at the time when panels were still expensive it struck me as an anomaly to use trackers. Solar PV works very well and very reliably, while trackers add moving parts to the mix that are bound to break. History has proven me right: All the big Deger trackers (most installed type in 2010 - 2011 when the FIT program took off) have had or will have their elevation motor fail. An expensive part! Can't tell you how often I've seen a tracker in a field pointing in some odd direction, and still pointing that way a week later.

KISS (Keep It Simple Stupid) works!

-RoB-