Question Regarding 120% Overloading on Inverters

[SunInTheFun]

Hello All, Long time creeper, first time poster :bye:

Little Bit of backstory... The customer in question has quite a few (15+) inverters and plenty of panels and was asking a few system questions to me. Most were pretty standard, however I wasn't quite sure how to answer one: "Is it better to have more inverters matched exactly with the AC rated output or fewer inverters with 120% overloaded" I know in alot of cases it is standard to overload by up to 120%, however I thought that this was more of a consideration to save money on inverters and make sure to squeeze as much power as possible out of each one. Since that is not a concern for this gentleman I was wondering what everyone here thinks about this, as I couldn't come up with an answer.
Also worth mentioning he plans on using thin film modules and is located in central Illinois, and there are no potential shading issues.

Thank you everyone here who makes this forum the go to place for answers!

 

[GoldDigger]

1. Do not use thin film modules if you want the system to last more than 5-10 years.
2. The overloading, as you suppose, does not have anything to do with number of inverters. You want to overload with panels if they are cheap and you want to maximize your production in the winter and during low sun angle hours (morning and afternoon) at the expense of some potential peak production. If the size of your system in kW is limited, overpaneling can increase the kWh you get from the same nominal size system.

PS: IMHO the only reasons for using thin film are near constant dappled shade or a need for flexible panels (as on a vehicle).

 

[SunInTheFun]

Thanks for the speedy reply, i believe the only use for the thin film is that he has the panels already, just the same with the inverters. Do you believe there would be a benefit to keeping the inverters with exactly 5kw of panels.... Let me see if i can come up with an easy numbers example.

Say he has has 12kw of modules, in your opinion do you think it would be better to have (2) 5kw inverters with 6kw of modules per inverter, or have 3 inverters: 2 with 5kw modules and one with 2kw of modules. I guess that is the question. Do you think one would produce more overtime or they would be roughly the same.

Thanks again!

 

[electrofelon]

Pretty much every manufacturers string sizing tool will call for more DC KW than inverter KW. You want more PV than inverter, the question is how much. I know on the larger .5-1.5 meg systems we do, the inverter capacity is usually around 75% of PV nameplate. Its actually a bit of an art to get that perfect PV to inverter size where you are not paying for extra inverter capacity, but not clipping too much power during peak times. I would look at the output of some string sizing calculators and compare the range of options to inverter cost for a start.

 

[GoldDigger]

Thanks for the speedy reply, i believe the only use for the thin film is that he has the panels already, just the same with the inverters. Do you believe there would be a benefit to keeping the inverters with exactly 5kw of panels.... Let me see if i can come up with an easy numbers example.

Say he has has 12kw of modules, in your opinion do you think it would be better to have (2) 5kw inverters with 6kw of modules per inverter, or have 3 inverters: 2 with 5kw modules and one with 2kw of modules. I guess that is the question. Do you think one would produce more overtime or they would be roughly the same.

Thanks again!

The three inverter setup will produce more, given that the amount of panel is fixed. Initial cost will, of course be more unless he already has three inverters.
He will have to spend marginally more on wire for the three inverter setup and may have problems with the 120% rule.

 

[SunInTheFun]

It is my understanding that the customer has purchased all the modules and inverters and is just looking to maximize power production. From what i have seen so far, (correct me if im wrong) it may just be best to match the modules to the inverters and have the remaining modules on its own inverter (after contacting SMA to make sure that there will be no issue throwing a lower KW array on the inverter than speced.

If anyone else has any thoughts let me know!


Thanks again everyone

 

[electrofelon]

It is my understanding that the customer has purchased all the modules and inverters and is just looking to maximize power production. From what i have seen so far, (correct me if im wrong) it may just be best to match the modules to the inverters and have the remaining modules on its own inverter (after contacting SMA to make sure that there will be no issue throwing a lower KW array on the inverter than speced.

If anyone else has any thoughts let me know!


Thanks again everyone

I would consult the manufacturer's string sizing tool. You panels will never/almost never put out their nameplate power.

 

[jaggedben]

It is my understanding that the customer has purchased all the modules and inverters and is just looking to maximize power production. From what i have seen so far, (correct me if im wrong) it may just be best to match the modules to the inverters and have the remaining modules on its own inverter (after contacting SMA to make sure that there will be no issue throwing a lower KW array on the inverter than speced.

My wager would be that buying the extra inverter is a waste of money. That's a foolish wager to make without actually knowing details, but nonetheless that'd still be my guess.

Here's the logic:

• Even if you overpower an inverter at more than 120%, you will only 'clip' energy output about 1-2% annually, because roughly 98% of the time your array will be at less than 85% nameplate output due to numerous factors. This means that...
• As you add module watts, the production you get from those watts is about 98% of what you were getting from the module watts you already had.
• As you add more inverter watts,the production you get from those watts is about 1-2% of what you were getting from the inverter watts you already had.
• It's clear which is the better bang for the buck. With module prices dropping to only 1-2 times inverters, you get 25-100 times the financial return by adding more module (or not buying more inverter).

Don't forget that inverter efficiency by itself takes care of the first 2-4% of DC overpowering, so in almost no case does it make sense for your AC to DC ratio to be above 95% (unless you got a deal on the higher power inverter).

With SMA in particular I have found that their voltage windows are somewhat strict and that not having fewer modules than the inverter is made for will cause problems.

 

[pv_n00b]

There are a couple of factors in play here:

When matching the DC to AC we usually use the STC power of the array and the AC rating of the inverter. The problem is that the STC power of the array usually is not approached in the real world, since the real world is not at STC. What this means that that you can usually go up to 1.15 to 1.2 and not have any significant clipping happening. Now to really dial it in you need to use software like SAM or PVSyst to predict how the system will perform hour by hour at the site. With this information you can design the DC/AC ratio for maximum efficiency so all the anticipated energy from the array makes it to the AC side of the inverter with no clipping. Now on to part II.

Once we have achieved maximum system efficiency we have to look at financial efficiency. Adding more modules past the point of maximum efficiency means that some energy will be lost to clipping. That means that every additional module added will provide less usable energy, i.e. the kWhr/kW ratio will go down and the marginal LCOE goes up. In the financial model we are going to have a target LCOE that we have to stay below for the project to work out but we want to generate as much energy as we can while staying under that LCOE. If our LCOE at maximum efficiency is lower than the target we can generate more energy by adding more modules, up to a point. That point is either we hit the LCOE limit, we hit the CAPEX limit, we run out of array area, or we hit the DC/AC ratio limit of the inverter. This will maximize the profitability of the project, even if it is not operated at maximum efficiency.

As a reference I’ve seen projects go in with DC/AC ratios as high as 1.55.

 

[SunInTheFun]

My wager would be that buying the extra inverter is a waste of money. That's a foolish wager to make without actually knowing details, but nonetheless that'd still be my guess.

Here's the logic:

• Even if you overpower an inverter at more than 120%, you will only 'clip' energy output about 1-2% annually, because roughly 98% of the time your array will be at less than 85% nameplate output due to numerous factors. This means that...
• As you add module watts, the production you get from those watts is about 98% of what you were getting from the module watts you already had.
• As you add more inverter watts,the production you get from those watts is about 1-2% of what you were getting from the inverter watts you already had.
• It's clear which is the better bang for the buck. With module prices dropping to only 1-2 times inverters, you get 25-100 times the financial return by adding more module (or not buying more inverter).

Don't forget that inverter efficiency by itself takes care of the first 2-4% of DC overpowering, so in almost no case does it make sense for your AC to DC ratio to be above 95% (unless you got a deal on the higher power inverter).

With SMA in particular I have found that their voltage windows are somewhat strict and that not having fewer modules than the inverter is made for will cause problems.

wow thanks for all the great information. I agree with you 100% but in this case the customer said that he didnt mind using more inverters, since he has them all just sitting around anyway, so i guess is there any downside to using an inverter or two that is at about half power (assuming that the manufacture says that its alright?) are there any pros or cons to doing it this way, or do you think that even if this is the case it would be better to overload the inverters? Again, thank you so much, and thanks everyone for the great responses and fast answers :thumbsup:

 

[pv_n00b]

wow thanks for all the great information. I agree with you 100% but in this case the customer said that he didnt mind using more inverters, since he has them all just sitting around anyway, so i guess is there any downside to using an inverter or two that is at about half power (assuming that the manufacture says that its alright?) are there any pros or cons to doing it this way, or do you think that even if this is the case it would be better to overload the inverters? Again, thank you so much, and thanks everyone for the great responses and fast answers :thumbsup:

Partially loaded inverter efficiency is going to be lower and it’s a curve, so if your array power is too low you are going to get really bad efficiency. But ½ load should not be too bad. As long as you have enough voltage to turn it on running the inverter partially loaded should not be a problem for the inverter.

 

[jaggedben]

wow thanks for all the great information. I agree with you 100% but in this case the customer said that he didnt mind using more inverters, since he has them all just sitting around anyway, so i guess is there any downside to using an inverter or two that is at about half power (assuming that the manufacture says that its alright?) are there any pros or cons to doing it this way, or do you think that even if this is the case it would be better to overload the inverters? Again, thank you so much, and thanks everyone for the great responses and fast answers :thumbsup:

You've really gotta look at the inverter specs and make sure that you can configure a string that will make it turn on and operate without dropouts. It may or may not be better to just spread the modules a little thinner over all inverters. I agree that if the inverters are a sunk cost then the arguments above are somewhat less relevant. However, it still might be more financially savvy to keep the extra inverters around as replacements for a few years from now, rather than underpower them.

The devil's in the details of specs and cost.

 

[ggunn]

There is no single optimal number for overloading an inverter. If, for example, I were designing a system in Canada with modules tilted at a few degrees off vertical I would overload the inverters more than I would if the same sized array were latitude tilted and located in south Texas. On a cool sunny day in both locations the modules in Texas would get closer to their STC power rating than would the modules in Canada. It's a judgement call. An expensive software program like PVSyst could give you a pretty accurate simulation of what the minute to minute performance of the systems would produce, and some clipping when the array is producing its maximum output is acceptable if the power lost to clipping is more than made up for by the increased production on the shoulders of the power generation curve. The bottom line is kWh per year produced by the system per kW of installed inverter AC power for a given array with a given orientation and location.

 

[SolarPro]

Right. And Illinois is the kind of place where a higher dc-to-ac ratio is generally beneficial.

 

[SunInTheFun]

You've really gotta look at the inverter specs and make sure that you can configure a string that will make it turn on and operate without dropouts. It may or may not be better to just spread the modules a little thinner over all inverters. I agree that if the inverters are a sunk cost then the arguments above are somewhat less relevant. However, it still might be more financially savvy to keep the extra inverters around as replacements for a few years from now, rather than underpower them.

The devil's in the details of specs and cost.

well after talking with the customer again this morning, it seems that he has more than 2x the inverterters he will need, thats why this is so interesting to me, i cant actually find hard numbers with oversizing arrays vs adding more inverters with the extra panels, i have been researching this for quite a few hours now and still cant really figure out which will make more power annually. I wouldn't think that this was the first time this question has been asked, but im starting to wonder. Since generally when doing a system you would look at the cost/performance first and make the decision on how many panels and what size inverters, in this case, everything was just purchased with no real plan which leaves me stumbling around looking for a solid answer. Thank you for your help everyone!

 

[ggunn]

well after talking with the customer again this morning, it seems that he has more than 2x the inverterters he will need, thats why this is so interesting to me, i cant actually find hard numbers with oversizing arrays vs adding more inverters with the extra panels, i have been researching this for quite a few hours now and still cant really figure out which will make more power annually. I wouldn't think that this was the first time this question has been asked, but im starting to wonder.

Again, there is no hard and fast magic number. A common rule of thumb is 120% - 125% overloading is where the maximum is if you plot kWh per year against DC:AC ratio, but that's a WAG and the actual maximum is highly dependent upon array location and orientation. PVSyst can calculate it pretty accurately if you are rigorous with your data input, but it's an expensive program (~US$1500) and not that intuitive to use. There may be a way to enter DC:AC into PVWatts, which is free, but NREL quotes a +/- 10% error margin for PVWatts, so it's not that much better than a guess.

This is not a question with a simple answer.

 

[wwhitney]

There may be a way to enter DC:AC into PVWatts

Yes, there is, under System Info/Advanced Parameters/DC to AC Size Ratio.

but NREL quotes a +/- 10% error margin for PVWatts, so it's not that much better than a guess.

Surely that's just a caveat about the weather data. That would be a systemic error, so for comparing two different options, it should be quite good on their relative performance. Of course, errors in the weather data will affect estimates of how often a system is clipping, so perhaps it is a bigger issue. Does PVWatts let you get numbers based on 365 sunny days a year? Then you could compare those to figure out the worst case clipping.

Cheers, Wayne

 

[ggunn]

Surely that's just a caveat about the weather data. That would be a systemic error, so for comparing two different options, it should be quite good on their relative performance. Of course, errors in the weather data will affect estimates of how often a system is clipping, so perhaps it is a bigger issue. Does PVWatts let you get numbers based on 365 sunny days a year? Then you could compare those to figure out the worst case clipping.

Cheers, Wayne

It's not just the weather data. When version 2 came out I ran a comparison between the two versions on the same PV system, and I chose a location right on top of a weather station to reduce/eliminate the effects of interpolation between the sites in version 2 compared to the using of the data from the nearest station that version 1 used. There was quite a difference between the outputs, so I contacted NREL to ask them which was the more accurate result. They just said that the difference in the results was within their +/- 10% margin of error.

Also, PVWatts doesn't allow for entering specific inverters and modules like PVSyst does, and you can get down and dirty with lots of variables that PVWatts just lumps into system loss. It can also show you precisely when and how much clipping you are getting. It uses the same weather data that PVWatts uses but it gives much more detailed and precise results. It's expensive, though, and using it effectively requires traversing a bit of a learning curve.

That said, however, for the OP's purposes, using PVWatts and running comparisons between different DC:AC ratios for what is otherwise the same system at the location and with the orientation of his array might be good enough.

 

[electrofelon]

well after talking with the customer again this morning, it seems that he has more than 2x the inverterters he will need, thats why this is so interesting to me, i cant actually find hard numbers with oversizing arrays vs adding more inverters with the extra panels, i have been researching this for quite a few hours now and still cant really figure out which will make more power annually. I wouldn't think that this was the first time this question has been asked, but im starting to wonder. Since generally when doing a system you would look at the cost/performance first and make the decision on how many panels and what size inverters, in this case, everything was just purchased with no real plan which leaves me stumbling around looking for a solid answer. Thank you for your help everyone!

I agree with the others that there is no single answer as there are too many variables, but if I found myself in your position of having extra free inverters lying around, here is what I would do: I would still want to "load up" the inverters to or above their power rating and use as many modules per string as possible. This would likely run the inverter where it is most efficient, and give me the least amount of voltage drop/loss on my DC conductors. Generally running the string voltage as high as possible is good too because under voltage during hot times has been more of a problem than high voltage during coldest times. Also this would reduce labor and material costs by having fewer strings, fewer inverters, fewer connections. I would consult the string tool and look centered around 110 % - 115% DC to AC ratio with the highest number of modules per string. Either that will get you there, or you will be on the line between having another inverter, at which point you will just have to make the call.

 

[pv_n00b]

Your situation is a little different than someone who is working from scratch. You have in hand more inverters that you need. Just match the DC of the array to the AC of the inverter, maybe go up to 1.1 or so but I would not sweat it, and put the other inverters in storage if there are any left over. The system will convert all the available energy from the array to AC and you are good to go. If you already have the inverters there is no reason to up the DC/AC ratio if all it does is to let you put inverters in storage. This would be a different conversation if you were trying to decide how many inverters to buy.

NREL SAM is free, better than PVWatts becasue it lets you specify equipment and unlike PVSyst is free. I use both depending on the situation.