Oversizing inverters

[J.P.M.]

I use the same simple analysis and agree that comparison of annual output of alternate designs is the best way to find a cost effective DC to AC ratio assuming all other design parameters are the same.
in summary, the important numbers using a 5.232 array are:
4 kW inverter estimated annual production 8,997 kWh
5 kW inverter estimated annual production 9,036 kWh
39 kWh extra production from upsizing the inverter.
If someone were to focus on clipping they might be fooled by the 273 hours of clipping and might spend ,$350 essentially to save $18 per year or as you point out, close to a twenty year payback..

Well, all the info I put in there is important or I wouldn't have put it in there.
IMO, you missed the purpose and sense of my post.
Some or most of it may not be important to you and perhaps to a lot of other readers but I wouldn't presume some or most of it is not important. I suppose opinions vary. And anyway, who made you the judge of what's important around here ?

My purpose in all that info was to flesh out some of the details behind a usable and easy method to maybe get some quantitative handle on inverter sizing (or undersizing).
I haven't seen much around here beyond a lot of talk and not much in the way of quantitative methods that might be useful.
I figured the easiest way for me and maybe an easier way for others to understand was by giving an example of how the method might work.

With that in mind, the 39 kWh/yr. penalty and each modeled system's output is actually unimportant.
If I had used a smaller inverter than 4 kW, the kWh penalty would have been different.
I tried to include enough info so that should someone have their curiosity sparked and take the initiative and do it themselves they have maybe a better go by from my post.
I put the 273 hrs./yr. of clipping in the post specifically to draw attention to the idea that sometimes things are not as they seem and perhaps provide a bit of insight on how things can be manipulated.

I try to provide information and answer questions as asked without spoon feeding people answers. The post describes a method and an example of how to estimate clipping in a PV system as f(inverter size, DC/AC ratio) and what might be expected. My purpose was to fill a need I felt existed and maybe add to the discussion, not simply puke out numbers as you seem to think, but provide something useful by example. I believe that's a better way to learn as folks may actually do a bit of critical thinking.

Take what you want of the above. Scrap the rest.

[davidcheok]

[QUOTE=J.P.M.;n439273]

What is "This" ?

This = Agree with this statement

[J.P.M.]

[QUOTE=davidcheok;n439280]

This = Agree with this statement

Thank you for clearing up my confusion.

J.P.M.

[davidcheok]

Here is something to talk about.

Specific yield for solar production has always been to use kWp of the panels (DC) as a measurement of system performance and that makes sense if the assumption that all energy will be converted by an infinitely sized inverter. However because of varying DC/AC ratios being used these days to system performance, these values IMHO no longer provide a feasible method of measuring 'yield'.

I did some calculations for my system below based on 6 months of data.

11076kwh / 181days = 61.2 kWh x 365= 22338 kWh/annum

Annual Estimated based on average past 6 months = 22338 kWh

kWp = 18 kW
kWac = 13.1 kW

Specific yield 1241 kWh/kWp or 1705.2 kWh/kWac

Capacity Factor = 22338/157680 = 0.142

Panel = 2.172 x 1.303 = 2.83 m2
Array = 84.9 m2
Efficiency = 21.2%

Nominal plant output = 1742.5 kWh/m2 x 84.9 m2 x 0.212 = 31362.9 kWh Expected

Performance ratio = 22338 kWh / 31362.9 kWh = 0.712

As you can see the specific yield will be much lower on oversized systems based on array size while high based on inversion capacity.

Discuss.

[Ampster]

Well, all the info I put in there is important or I wouldn't have put it in there.
IMO, you missed the purpose and sense of my post.
Some or most of it may not be important to you and perhaps to a lot of other readers but I wouldn't presume some or most of it is not important. I suppose opinions vary. And anyway, who made you the judge of what's important around here ?

Sorry, I agree with DanS26 that too much emphasis is on clipping and it is a distraction from what the actual numbers suggest. I understand what you write is very important to you and you are free to talk about clipping until the sun goes down.. I believe annual production is a useful tool to optimize a system. What I express as my own personal opinion does not mean to imply that I am the judge, any more than anyone else who expresses an opinion.

My comments are for the benefit of other readers who want to "step out of the box" and look at the entire picture not just the obvious flat top presented by a clipped output curve. There are a number of good reads on the subject and everyone is free to form their own opinion about the appropriate DC to AC ratio for their system. Here is one DanS26 posted earlier in this thread:
https://www.gses.com.au/wp-content/u...Oversizing.pdf
or this also posted earlier in the thread:
https://global.apsystems.com/wp-cont...generation.pdf

To me a DC to AC ratio of 1 to 1 is a waste of inverter capacity. I don't know that I have the time or the real estate to make a DC to AC ratio of 2 to 1 work for me like it has for bcroe but I am envious of what he has accomplished. What the optimum ratio is depends on a lot of factors, some of which can be input into PV Watts to get a rough ranking of alternatives.

[Ampster]

Here is something to talk about.

........... However because of varying DC/AC ratios being used these days to system performance, these values IMHO no longer provide a feasible method of measuring 'yield'.

I did some calculations .............

As you can see the specific yield will be much lower on oversized systems based on array size while high based on inversion capacity.

Discuss.

Yes that is another metric that one can use. There is definitely an optimum ratio if DC capacity to AC capacity. I believe it is greater than 1 to 1. The closer one gets to 2 to 1 the less effective the system becomes unless there are various orientations like in Bruce Roes setup. . I would not even hazard a gusss about the optimum configuration of inverter size would be for a 18kW system without knowing location, azimuth, tilt and other factors. I think we can agree that the optimal ratio depends on lots of factors.

I prefer PV Watts which allows one to enter a specific DC to AC ratio, and I use annual production as a factor to rank specific configurations. I am not suggesting it is perfect and I don't know how it compares to others like Aurora which one has to pay to use.
One of the things that can effect production is temperature. In 2018 I had a system installed on a home with significantly different weather than my previous home.That system had a DC to AC ratio of 1.5 to 1, and I was seeing clipping in March. Extrapolating to Summer with the longer days I assumed the clipping would be much worse. I was wrong and the reason surprised me. The answer was temperature. In March when I made my observations the average daytime temperatures were in the forties and the panels were producing above their STC rating. In summer the average temperatures were in the eighties and nineties and the panels were producing below their ratings and despite the longer days of sun there was less clipping. Tracking the annual production over the three years I lived there confirmed that for that specific location and configuration that higher ratio was optimal.
Depending on your latitude and tilt your system may have significant differences in yield from one season to the next. Is there a reason you did not use annual data? Did you run your system size and your hypothetical system into PV Watts to compare results of annual production for various configurations?

[davidcheok]

Is there a reason you did not use annual data? Did you run your system size and your hypothetical system into PV Watts to compare results of annual production for various configurations?

Yes because this system is only 6 months old. Solar is just starting to pick up here and we are on NEM 1.0. I had to calculate and extrapolate data from sources from the northern and southern hemisphere while we are on the equator and there is very little real-world data for me to work with in my research into solar.

I tried almost every calculator out there and got different results for all. The one I preferred was global solar atlas https://globalsolaratlas.info/ and this was about 5-10% off.

Clouds rather than temperature is the biggest factor affecting performance here all ceterus paribus. Below are two systems using the same DC/AC, one system with 10.8 kW and other with 18 kW with location about 10 km apart. The other factor was wind where the first chart has consistent wind to keep the micros cooler while the other is surrounded by hills and trees which makes the micros throttle down (later half of chart).

Thing is.. its difficult to compare apples to oranges using specific yields because different systems have different modifiers. Its very difficult find a good calculator to estimate production because modifiers like clouds and wind is impossible to estimate and play a huge role in estimating output.

Based on my observation, I would say the biggest variable DC/AC ratios affect would be the low light performance potential with larger array sizes being able to generate more for the inverters and produce the most versatility as opposed to arrays sized to match the peak solar irradiation cycles.






Screenshot 2023-05-01 at 5.27.40 AM.png


Screenshot 2023-05-01 at 5.28.23 AM.pngLuckily for me, the 18 kW system produces enough to keep my monthly draw to under 600 kWh which is the ideal zone due to our highly subsidised rates. Too much would have my system exporting excess to the DNO who does not pay back in terms of cash. Any excess at the end of the year is zeroed out. On average my monthly bill is about $3 for now until I decide its time to get an EV. Once that happens, I need to put in more inverters. For my area to get an additional 40 kWh, I would need to put up an additional 10.8 kW of panels worth. Oh well, thats maybe a few years down the line.

[Ampster]

........

I tried almost every calculator out there and got different results for all. The one I preferred was global solar atlas ,,,,,,,,,,and this was about 5-10% off.
.........

I also compared its output estimate to PV Watts and found the above tool to be 5% higher. As a tool to determine optimum DC to AC ratio, I did not see where there was any way to change that assumption in the calculations. nor did I see what that assumption was or what the assumption was for inverter size.
Therefore I do not thing it would be useful for the exercise of determining the optimum configuration in terms of panel size to inverter capacity.

[J.P.M.]

Sorry, I agree with DanS26 that too much emphasis is on clipping and it is a distraction from what the actual numbers suggest. I understand what you write is very important to you and you are free to talk about clipping until the sun goes down.. I believe annual production is a useful tool to optimize a system. What I express as my own personal opinion does not mean to imply that I am the judge, any more than anyone else who expresses an opinion.

My comments are for the benefit of other readers who want to "step out of the box" and look at the entire picture not just the obvious flat top presented by a clipped output curve. There are a number of good reads on the subject and everyone is free to form their own opinion about the appropriate DC to AC ratio for their system. Here is one DanS26 posted earlier in this thread:
https://www.gses.com.au/wp-content/u...Oversizing.pdf
or this also posted earlier in the thread:
https://global.apsystems.com/wp-cont...generation.pdf

To me a DC to AC ratio of 1 to 1 is a waste of inverter capacity. I don't know that I have the time or the real estate to make a DC to AC ratio of 2 to 1 work for me like it has for bcroe but I am envious of what he has accomplished. What the optimum ratio is depends on a lot of factors, some of which can be input into PV Watts to get a rough ranking of alternatives.

OK, back in town.

For those wishing only the bottom line of this post and no other information, a precis : Skip to the last paragraph of this post.

1.) I agree there is too much emphasis on clipping. It is one, but not the only aspect of residential PV design, and it's probably for most common residential applications a secondary consideration.
2.) Broadly writing, and to that end, any good design in any area of engineering I've worked in and as I learned and practiced it usually has top level goals of a safe, fit for purpose, including serviceability and cost effectiveness with lots of other goals and priorities as called for by the project design goals.
3.) Relative to residential PV, from what I've seen on many applications and also what I've read and discussed here, those goals will probably be similar for most residential applications with cost effectiveness being high on the list.
4.) If cost effectiveness considerations govern the design (after safety and code compliance considerations), then I'd submit that the idea of DC/AC ratio is a means to describe one of the outcomes of good design rather than a design goal in itself. That seems backwards to me.
5.) However, I'd suggest that for most residential designs being done today, the first design goal would be to meet the annual design load, whatever that may be, including any oversizing for any reasonably certain future system expansion or any reasonable contingencies.
6.) Only after that, and after any initial array sizing iterations or adjustments, the question of inverter sizing or methods (micro, string or optimizer ?) would be worked out. That is, only then would array to inverter size (DC/AC ratio) be addressed.
8.) In doing so, one possible way to start (for string inverter equipped arrays anyway), be with inverter sized to == the array STC wattage. Then, reduce the inverter size in steps until one or both of two things happen:
a.) The model's (assuming PVWatts or whatever other reasonably reliable model) output matches the design load is negatively impacted to the extent the system's modeled output is reduced.
b.) The cost differential savings of a smaller inverter to a larger inverter is not as great as the increased cost of the clipping.
When either or both of those things happen, the design's cost effectiveness will be negatively impacted.

9.) Whatever the final design DC to AC ratio winds up being, it ought to be arrived at as a result of design rather than an up-front design goal.

DC/AC ratio is best used as an end of design system descriptor rather than a design goal.
Putting it before the rest of the design process puts the horse before the cart.
Ampster: To your comment of and with respect to DC/AC ratios being 1.0 (or less I'd assume, with apologies if I'm being presumptuous) and so being a waste, without some logical method of inverter sizing, as long as the final cost of an STC W of installed array (however they are determined), including per W costs for wiring, structure and other B.O.S. equipment are greater than the per W costs (however they are determined) of inverters, using a DC to AC ratio of greater than one may wind up being a waste of panels. If that's the case (per STC W panel/B.O.S. costs > inverter/W costs), incorrect inverter sizing relative to design goals will have a higher probability of negatively impacting system cost effectiveness which is usually the highest system goal. <1.0> means nothing. It's a number. A cost-effective design can have any DC/AC ratio. It depends on the application.
Take what you want of the above. Scrap the rest.

[jflorey2]

6.) Only after that, and after any initial array sizing iterations or adjustments, the question of inverter sizing or methods (micro, string or optimizer ?) would be worked out. That is, only then would array to inverter size (DC/AC ratio) be addressed.

I think this is key.

Solar/inverter ratio is pretty far down on the list of things to consider when you start designing a system.

Are you limited to 40A backfeed due to the 20% rule, and are all your inverter options (up to 7600 watts) pretty similar? And will 6kW of solar get you to your goal? Then there might be no reason to ever overpanel.

Do you want to expand in the future? Then again maybe don't overpanel so expansion is just adding another string.

Do you have two different orientations? Then overpaneling becomes much less of a big deal for the obvious reasons.

Are you limited to 40A but need a 12kW system? Then you're going to be significantly overpaneling out of necessity - and it probably won't be that bad.

To put it another way, the overpaneling ratio should be an output of, not an input to, the equation.

[Ampster]

.........
DC/AC ratio is best used as an end of design system descriptor rather than a design goal.
Putting it before the rest of the design process puts the horse before the cart.

The DC to AC ratio is a metric that can be used to view the annual production of a design by entering that parameter in a tool like PV Watts. It has never been a design goal in my way of thinking. As I have mentioned, there is too much focus is on clipping and not enough on total production. Since starting this response jflorey summarized my thoughts very succinctly,
As far as your other question about ratios below 1 to 1 being a waste of inverter capacity it is a pretty simple concept that the most cost effective solution should be chosen. I don't think a longer explanation is needed. I have not seen any scenerios where a larger inverter costs less than a smaller inverter. At a ratio of less than 1 to 1 the limit on production would be the panel output so there would be no change in production of the system by oversizing an inverter. I do agree that if one wanted to expand the system later that would be a reason to oversize an inverter but that is a hedging strategy more than a strict numbers decision.
Are there any other reasons why oversizing an inverter would make financial sense?
It is ironic that the title is phrased, "Oversizing Inverter", but the OP started the discussion by talking about over paneling.

[davidcheok]

Lots of different opinions on what comes first.

In my opinion, my workflow would be:

1) Identify solar real estate first whether roof or ground mount and from there identify what the largest possible capacity should be.
There is no point dreaming of potential annual production if you do not have the space to put in those panels. Then identify the electrical infrastructure to see what it can carry or if upgrades are necessary to handle the potential maximum load.

2) Calculate the worst case scenario for annual load for the property and identify a target kWh to achieve first. Once you have that and the real estate data can you start trying out configurations to your heart's content.

Like i said before, if only it was as easy as dumping some numbers into a web page. So many factors affect output that these calculators cannot model especially with weather. You know the saying, lets talk about the weather. Things are complicated further when you live in a country where the sun is not constantly in the north or south but cycles between the two throughout the year or live in a house in a valley that has wind cut off at certain times of the day and shadows moving in from places you least expected.

Other things to note. Inverters have a lifespan of that necessitates at least replacing once within the lifespan of the array. There are pros and cons to both panel and inverter oversizing. e.g. 1. spending extra on a bigger inverter to eliminate clipping and provide future expansion may seem smart now when the array is new but as the panels age, output drops further increasing the gap between peak output and maximum real output. Adding more panels or increasing panel size instead of spending on the inverter may cause clipping but it also increases the usable energy at lower light levels across the mornings and evenings or in winter and allows you to maximise the inverter output across a longer part of the day. IMHO it just makes more sense to me to spend more on what im less likely to replace later and ensure I have ample potential production for my inverters to use rather than have bigger inverters that do less. Inverter technology is also changing and in ten years or more, i would be more likely to replace my inverter than to replace my panels. Cost of inverters would also likely to have dropped and they are far easier to replace than panels.

[Ampster]

......Ampster: To your comment of and with respect to DC/AC ratios being 1.0 ......being a waste,......

i did not have to look too far to find an exception to my statement that DC to AC ratios of less than 1 to 1 are a waste of inverter capacity. That exception would be my own hybrid system which consists of 3kW of DC coupled solar, 8kW of AC coupled solar with an inverter which has a capacity of 12 kW. However in the case of hybrid inverters which typically have batteries the AC capacity is more related to the need to power the loads of the home to shift loads or during power outages with battery power often without the use of solar. If you consider that the AC coupled solar already contains 6.6 kW of inverter capacity i could say that I have 11 kW of PV solar and 18.6 kW of inverter capacity. I believe my statement still applies to strictly GT systems.

[darkskies]

I have a 11.9 kw system and 7.6 kw inverter (1.56 DC-AC ratio). The PVWatts, which my installer used for estimating, had showed slightly over 14.3 MWh production over a year. Last year (our first full year of production), and the system generated about 11.3 MWh - about 20% less. I'm guessing that trees are affecting production vs estimate. Unfortunately, our consumption at this new home, is roughly the same, so were are not cost free, from that perspective. With the newer tariff that we have with the energy company for this house (not 1:1), I estimate about 13 MWh to break even.

I've been analyzing our consumption at this new place (about 1 MWh more than the old house we were at, so not too different), to see if there are any things I can do to reduce consumption (there are a few things). With the 16 kWh battery that was finally added last May, it has helped produce another 500 KWh for a partial year, so that will help as well.

I talked to my installer, mentioning the difference between projected and realized power. I also had mentioned that even though we have no electric cars at this time, there is a possibility of going that direction in a few years, as my wife's car is getting old and she mostly does short trips, so having more power generation would be desirable.

He suggested that we could upgrade the 7.6 kw inverter (all that was available at the time e bought the system), with a 11.4 kw unit that is available now - for the cost difference of the two units. He mentioned that we could also add more panels in the future (not something we want to do right now). We have quite a bit of clipping with the system right now, so I suspect he is suggesting to recover some of that.

Any thoughts on that approach?

[DanS26]

I have a 11.9 kw system and 7.6 kw inverter (1.56 DC-AC ratio). The PVWatts, which my installer used for estimating, had showed slightly over 14.3 MWh production over a year. Last year (our first full year of production), and the system generated about 11.3 MWh - about 20% less. I'm guessing that trees are affecting production vs estimate. Unfortunately, our consumption at this new home, is roughly the same, so were are not cost free, from that perspective. With the newer tariff that we have with the energy company for this house (not 1:1), I estimate about 13 MWh to break even.

I've been analyzing our consumption at this new place (about 1 MWh more than the old house we were at, so not too different), to see if there are any things I can do to reduce consumption (there are a few things). With the 16 kWh battery that was finally added last May, it has helped produce another 500 KWh for a partial year, so that will help as well.

I talked to my installer, mentioning the difference between projected and realized power. I also had mentioned that even though we have no electric cars at this time, there is a possibility of going that direction in a few years, as my wife's car is getting old and she mostly does short trips, so having more power generation would be desirable.

He suggested that we could upgrade the 7.6 kw inverter (all that was available at the time e bought the system), with a 11.4 kw unit that is available now - for the cost difference of the two units. He mentioned that we could also add more panels in the future (not something we want to do right now). We have quite a bit of clipping with the system right now, so I suspect he is suggesting to recover some of that.

Any thoughts on that approach?

Cut some or all of tree shading will be your biggest bang for the buck......