Sami, Welcome to the forum.

I'm not sure where the 1580 kWh/kW rate came from. Perhaps it is a specific model for your region and installation. If you want to calculate the correct ratio for your specific situation, many of us like a web tool called PVWatts. It does much better than a "rule of thumb".

Without knowing more, I would assume that the ratio is based on inverter AC output, but that's a guess.

Assuming that you want one string inverter, you can them it in various sizes. For example, SMA makes the Sunny Boy inverter in 3kW, 3.8kW, 5kW, 6kW, 7kW, and 7.7kW. Another brand, Fronius, makes the Primo in 3.8kW, 5kW, 6kW, 7.6kW, 8.2kW 10kW and even larger sizes.

An alternative is to put one inverter on each panel. These are called microinverters. That way, you can expand at any time, just by adding another panel and another microinverter. For the same power level, microinverters can be more expensive than the SMA inverter, but not by much. Some people like microinverters for their ability to work well when the panels get partial shade.

As for the size inverter to buy, we use a ratio of somewhere around 90%. If you are using 16 400-watt panels, you will get 6400 watts peak, so a 5700 watt inverter is close to optimum. A larger inverter will work well, but cost more money. A smaller inverter will put out significantly less, so you will be wasting panel capacity.

Although the inverter may seem undersized, it will only result in less-than-peak power for a short period of time around noon. The rest of the time, the panels will produce less than rated power and the inverter will handle it fine. That slight loss of power will be ~10% for a few minutes, so really just ~1% of total production. We consider this optimum because you are saving more than 1% of the installation cost.

One last option you may want to consider is QS1 quad microinverters from a company called APsystems. These have the cost advantages of string inverters and the flexibility and shading advantages of single microinverters. They save money by putting four independent inverters in one box. I have no vested interest in this product, but have used it successfully at my home.

Some people on this discussion group don't like microinverters because they require placing the electronics on the hot roof. String inverters are normally installed on the ground, where it is cooler, but may require optimizers or other electronic devices on the roof, so you don't avoid the heat problem either way.

I mentioned a few products here. If any interest you, use a search engine to find the manufacturer's website and learn more about them. Each has advantages and...

Thank you very much Bob. And for the 1580kWh/kW, it is the Peak Sun Hours multiplied by the losses factor of 0.8 and by 365 days (for annual calculation), this is how it is done here in my region.

Sami, If you have a few minutes, you may want to try running PVWatts and see how it compares with 1580kWh/kW. It is an on-line tool with weather data for the whole world. You can use it with all defaults or include as many details as you like. For example, if your panels are facing south-east instead of south, it accounts for that. Also, if your panels are facing upwards at 30 degrees or 25 degrees, it accounts for that.

I agree, dont waste your money until you have run PV Watts. Its easy to use and gives conservative numbers as long as you are accurate in taking into account local conditions like shade from landscaping and shadows from the structure.

I will, thank you very much

If your POCO requires you to be on a time-of-use rate plan, you may need to over produce in order to compensate. In our area, we have to design for 130% to 140% of actual usage in order to break even on the billing. Plus almost everyone uses more power as time goes on - all I ever hear is "more power, more power". Plus, the solar array does degrade slightly over time, although current stats are proving it to be less than 1/2% per year. So with solar PV panels being so inexpensive now, go as big on the array as you find affordable. Then put in an inverter that is as big as the largest array your roof can hold. The next person with a large family that lives in your home will love you. There is very little difference in price between the smallest and largest inverters. An oversize inverter will last longer than a smaller one operating at its max. Anymore, we almost exclusively install 7.7kW inverters (max size for a 200Amp service) for everyone.

Interesting calculation. Not sure what area of the world you are in but down here in sunny San Diego I'm getting better performance out of my solar array. I have a 4kW (dc) array on my roof and since installation in 2019 I'm averaging ~7,600 kWh (ac) of yearly production. That puts my yield at 1900 kWh(ac) per year per kW(dc). As others have said PVWatts should get you in the ball park.

Bob: I'm with you on PVWatts for Sami's modeling. Seems to me that Sami is more savvy about some solar stuff than the average bear, but if nothing else, getting up to speed on PVWatts will at least confirm or refute the 1,580 kWh/yr/STC kW. I use the specific yield number all the time and in various ways. FWIW, my system's logged actual 7 yr. specific annual yield is 1,726 kWh/(yr*STC kW). PVWatts comes up with 1,732 kWh/(yr.*STC kW). It can be used as a preliminary sizing tool. It has other uses as well.

Sami: Welcome to the neighborhood.

1.) Your methodology may work, but I'd get familiar w/PVWatts. Almost too easy not to use to get a bit more info. Read ALL the help screens a couple of times before you make any runs. The model is pretty easy to learn with that intro. Use 10 % system losses. The 14 % default value seems to underpredict output for many locations as confirmed by many folks around this forum and elsewhere.
2.) PV systems are generally rated by STC size. That is, (the panel's STC wattage) * (number of panels in the array). If so, for your numbers, that 5.7 STC kW number is the D.C. array rating of 5.7 STC kW before the inverter losses. If your number of 1,580 is the modeled # of kWh/yr. that 1 STC kW of panels will produce in a specific location and orientation, then the array size needed to produce 9,000 kWh/yr. as a long term average is probably something close to 9,000 kWh/yr./ (1,580 kWh/yr*STC kW) = 5.7 STC kW or so. The caution here is to make sure that your 1,580 number is the modeled specific annual output in units of kWh/yr. per installed STC kW.
Example from my array above : System size : 5.232 STC kW. Annual modeled output = 9,062 kWh/yr. Specific output = (9,062 kWh/yr.)/5.232 STC kW = 1,732 kWh/(yr.*STC kW).

2.) It's probably not necessary and it's probably overkill to have an inverter that has a larger rating than the array - or for that matter even as large as the array.
Reason: In actual operation, an array will only produce it's STC rating a few times a year, and then only for short periods of a few minutes or so. Unless you're in a cold climate and more than a couple km. above sea level, under good conditions of clear skies, low(er) temps. and a good breeze, about the best steady output you'll get out of a well designed and sited system is maybe 90-92% of an array's STC rating, perhaps a bit more, but if more, only for short periods of maybe a few minutes or less and only a few times/yr. The value of what's usually that small increment in annual output may not be worth the cost increase of a larger inverter. Example: I live near San Diego with close to a goldilock's climate. My array is 5.232 STC kW. My inverter is 5.0 kW. The highest continuous output of more than 10 minutes duration I've measured and logged has been 4.91 kW out of the array and into the inverter. array output is clipped by the inverter at 5.032 kW about 1X/month or so for maybe a minute or less.

So, if you want to oversize the array, and using your 1,580 specific output #, to meet a 20 % higher load, you may well need a (9,000kWh/yr*1.2)/1,580 kWh/(yr*STC kW) = 6.84 STC kW array. But you probably won't need a 6.84 kW inverter. For a lot of reasons too complicated to thoroughly explain here, mostly because I'm a hunt & peck typist, If your array will be split or a lot off south, the inverter can often be smaller yet.

If you haven't done so already, try this: More PVWatts. To reiterate, read all the help screens, particularly the parts dealing with DC/AC size ratios. Run up the array size until you get to whatever annual output you want. If you want, say,10,800 kwh/yr., and if the modeled array size comes out to be ~ 6.84 STC kW or so., the inverter will probably be less than 6.84 kW. Then, vary the DC/AC size ratio while holding the array size constant and see what happens to the annual output as the DC/AC ratio increases.

Note: An array's annual actual output (or maybe more accurately an array's running 365 day modeled output) will vary by maybe +/- 10 % or so from the modeled output mostly due to irradiance fluctuations from clouds, with very sunny and very cloudy climates often having less variation (example: After the first year's operation, my array's 365 day running variation seems to be ~ +/- 4.0 % or so, but I'm in a pretty sunny location).

While you're at it, download a PDF of a slightly dated copy of "Solar Power Your Home for Dummies", or get the revised hard copy at bookstores/Amazon for ~ $25 or so. A good primer on residential PV.

Thanks so much for all the help, I truly appreciate it. Very glad to be here.

1. your bill is based off AC usage, so any calculation based off that is also based off AC usage
2. the cost difference between a 5kw and 7kw inverter is pretty small. Always better to use a bigger inverter so that it does not heat up as much and lasts longer IMHO.

"the annual specific solar yield is 1580 kWh/kW."

your units are wrong. it is 1580kwh/year

if you use pv watts and put in a solar array size of 1kw DC then the website will give you your annual specific solar yield.

Actually some calculations from PV watts can be used to show a kWh/kW production.

When I ran a calculation a few years ago for a 6kW system PV watts showed about 9000kWh/year. That calculated to about 1500kWh/kW (9000kWh/6kW = 1500kWh/kW)

you dropped the unit of years in your calculation how did 9000kwh/year just become 9000kwh?

And as long as your separating fly frass from pepper and seeming to be attempting to start a pissing match, why didn't you capitalize the "W" in "kWh" as is customary for abbreviations of S.I. units named after people ?

Just like correct grammar, spelling and syntax in writing, if you want to be taken seriously, conventions and rules matter. Otherwise, fair or not, you'll probably look ignorant to those reading what you write and who do know the subject being dealt with. That can make it more difficult to be taken seriously.

The specific solar yield is in kWh/kW/year but I added "annual" to the sentence that's why I ommited the "year" in the units. Not a big deal, sorry for the complication.