New Guy..System Upgrade (actually system replacement) question

[SolTex]

PVWatts is fun to play with, and in our case it was very accurate. We used the calculator to check against the system "sizing" recommendations of our installer. Our goal was to cover 25,000 kWh of consumption per year. The calculator said our system would produce about 26,000 and after our first year of operation we are running at 101% of the prediction. So pretty close!

Note that most of the parameters in the calculator can be tweaked to your liking. For instance, you can run it with a DC/AC ratio of 1.2 (which is default). Then run it again with a DC/AC ratio of 1.0 and see how the total annual production number changes. Very interesting!

[Ampster]

Hmmm. Inputting my system in PVWatts, then downloading the excel spreadsheet for hourly production... looking at how many days I would be production over 11,400W

Is that the idea?

.....

No the most useful number to me is annual estimated production. I do not care as much about the daily peaks or even if I see clipping as long as I optimize production for the year. My goal is to cover my annual loads and have as small a bill as possible. To get a more acurate number in PV Watts, go to the advanced section and change the DC to AC ratio from the default 1.2 to your actual ratio of 1.29. You might see a slight difference but it will give you a sense for the assumptions you may want to make when configuring your system. If you are using micros again you will have a some latitude in choosing the best panel and micro combination.

[J.P.M.]

1. 95682
2. Arrays are placed due south. Dont want to alter this, holes and concrete and all...
3. No. And google isn't helping....
4. PVWatts looks to be the NREL tool for estimating solar production? Have not used it, just found it- and will start plugging away..
5. PGE
6. ?
7. My current sunpower 7.7kw system results in an annual true up of $3000 a year. Id like it to be much less.... Large home, two ACs, shop, pool, landscaping- on a well, so all water is pumped from 400ft down. And will probably by an EV in the next year. Presently it seems that my system is meeting 50% of my (current) power needs

Thank you.

#2: OK. Array azimuth is 180 degrees. What is the array's tilt relative to the horizontal ?

#3 and # 4: SAM is PVWatts on steroids. Check the NREL site. On PVWatts I suggest you read ALL the help screens and notes and get the array orientation(s) as close as possible before doing ay runs. It's a quick but necessary learning curve. Use a 10% system loss parameter.

#6: Question was of roof or ground mount. Omitted by my editing error on check before posting. Apologies.

More later.

[J.P.M.]

Hmmm. Inputting my system in PVWatts, then downloading the excel spreadsheet for hourly production... looking at how many days I would be production over 11,400W

Is that the idea?

Just took a quick spin. Getting late, will drill down tomorrow.

thx

How many days (or total annual hours) in clipping is a nice number but I'd wait until you get further into the design. Learn to walk before you run.

Also, a tip: Using a 10% system loss parameter rather than the 14% default seems to get the model's output closer to what a system is likely to produce for many users at least for those without way off the usual inputs.

To reiterate, reading the ALL the help screens can be an educational process as well as a necessary one for a model's output to have a pretty good probability of meeting what the actual system is likely to produce.

[Ampster]

How many days (or total annual hours) in clipping is a nice number but I'd wait until you get further into the design................

As I have often said people can get distracted by focusing on clipping. It is a "nice" number but is useless in deciding on a configuration unless the OP also calculates the excess produciton earlier in the day and later in the day that often happens with higher DC to AC ratios. Then thos two numbers would have to be netted against one another. I am not sure how one would even determine how much clipping there was from a PVWatts output. That calculation is much more complicated that jusr comparing estimated annual production numbers for various DC to AC scenerios. It is much easier to run different DC to AC ratios in PVWatts to see whether adding AC capacity or reducing DC solar capacity will have a cost benefit. Those scenerios can be used when the OP gets further into the design to decide the optimum or most cost effective design.

[ard]

Really useful inputs so far- thanks much

Tilt is 27 degrees. I had already dropped the system loss to 10.2...looking at the values, voltage drop and others, 14 seemed high)



I was dinking around with PVWatts, and outputted the hourly spreadsheets...figured out how to calc the clipping- but then was noodling on how to capture the delta-efficiency when the inverter is operating at lower powers..JUST like other posters mentioned.

Then had the idea of "dont worry about it day by day, just sum it for the year" and compare that way...again, just like some posted.

Then running the simulation for different 'inverter DC/AC percentages'

Getting there, did it quick and dirty, mostly to get a sense of how it runs and how I am treating the outputs.



Question: Default inverter efficiency is 96% on PVWatts.... should I use 99%? (For example, the solaredge 11400 has a CEC weighted eff. of 99%@240)

thx

[J.P.M.]

I am not sure how one would even determine how much clipping there was from a PVWatts output.

I'm not surprised you are confused.
I was able to do it for about the first 6 years of my array's operation given a good system monitor (that Sunpower gutted) and to model it with PVWatts seems like a no brainer. It's also easy to model using SAM or other software. Sometimes it takes a bit of creative thinking, but not much.

If we're talking about a model, such as PVWatts, get a modeled hourly output for 8,760 hours.
Then, for all hourly values of array output > (the inverter's rated capacity divided by inverter efficiency) get the difference between that hour's array output and the inverter capacity. That will be an estimate of the amount of inverter clipping for that time period.
Use an "if" statement to zero out cells in that column that have negative values.
Sum that difference column and voila' : The system's modeled clipping for a TMY (Typical Meteorological Year).
Flag the positive valued cells in another column with a "1" and sum that column to get a rough idea of the total # of clipping hours clipping hours over a year as I described in another thread where I got 273 clipping hours for my array with a 4kW inverter.
If you want to get creative you can do a lot of stuff with the clipping column such as frequency distributions, binning the data, cumulative frequency distributions and other stuff.

If we're talking about an actual system, if one has access to panel output and inverter output, the same method can be used, as ironic and perhaps counterintuitive as it may seem, actual short-term data won't be as applicable for design purposes as more generalized but it'll likely be of less use because it'll be based on a limited sample, but that's off topic.
Before Sunpower took all the nice stuff off the customer monitors, among that data were a lot of other array and inverter parameters in 5 minute increments, I got a boatload of input to and output paramet6ers from the inverter and the array and saved it daily.

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

[J.P.M.]

Really useful inputs so far- thanks much

Tilt is 27 degrees. I had already dropped the system loss to 10.2...looking at the values, voltage drop and others, 14 seemed high)



I was dinking around with PVWatts, and outputted the hourly spreadsheets...figured out how to calc the clipping- but then was noodling on how to capture the delta-efficiency when the inverter is operating at lower powers..JUST like other posters mentioned.

Then had the idea of "dont worry about it day by day, just sum it for the year" and compare that way...again, just like some posted.

Then running the simulation for different 'inverter DC/AC percentages'

Getting there, did it quick and dirty, mostly to get a sense of how it runs and how I am treating the outputs.



Question: Default inverter efficiency is 96% on PVWatts.... should I use 99%? (For example, the solaredge 11400 has a CEC weighted eff. of 99%@240)

thx

For initial design, use the inverter efficiency published with the inverter spec sheet. If more than one, or if charted against some other parameter, use an average. It'll not be that important until the last design iterations, and probably not much of a deal then, unless you're using micros.

SAM handles that rather nicely, but for most applications, most folks don't want to take the time. The learning curve can be long and arduous.

[bcroe]

To avoid extra error, the actual array output required to drive the
inverter to clipping should be considered. DC losses are near
1% here, and claimed inverter efficiency is 95%. If an inverter is
really 99% efficient (I have my doubts) that adjustment will be
different.

I have been taking the mfr rating, inverter efficiency is not easy to
measure with good accuracy. I did try to grab some input and output
numbers off the display, but got a likely flawed 93%. The monitors
that drive the display are very suspect. Sun and line voltage tend to
keep moving, so getting 5 accurate numbers at any instant is difficult.
Beside DC voltage and current in, needed is AC voltage, current, and
power factor out. Bruce Roe

[Ampster]

.........

If we're talking about a model, such as PVWatts, get a modeled hourly output for 8,760 hours.
Then, for all hourly values of array output > (the inverter's rated capacity divided by inverter efficiency) get the difference between that hour's array output and the inverter capacity. That will be an estimate of the amount of inverter clipping for that time period.
Use an "if" statement to zero out cells in that column that have negative values.
Sum that difference column and voila' : The system's modeled clipping for a TMY (Typical Meteorological Year).
Flag the positive valued cells in another column with a "1" and sum that column to get a rough idea of the total # of clipping hours clipping hours over a year as I described in another thread where I got 273 clipping hours for my array with a 4kW inverter.
If you want to get creative you can do a lot of stuff with the clipping column such as frequency distributions, binning the data, cumulative frequency distributions and other stuff.
..........
.

You can talk about clipping all you want but if you only do the calculation to determine clipping hours you are ignoring the other increases in production that sometimes come from higher DC to AC ratios.

I prefer a simpler approach that just compares the annual output of two different configurations and compare the costs of those two configurations to decide which one is optimal? In your example you obtained a number of hours of clipping but there was no kWs associated with that number so additional assumption would have to be made. That clipping figure in hours would not be useful to rank various configurations anyhow. How would you use just the clipping hours to rank two different configurations without looking at estimated produciton over time?

As I have said before, focusing on clipping is a distraction that keeps one from looking at the important numbers of system production for a given period. The fundemental economics are derived from the kWhs that a system might produce not the kWs that it might have lost due to clipping.

[J.P.M.]

You can talk about clipping all you want but if you only do the calculation to determine clipping hours you are ignoring the other increases in production that sometimes come from higher DC to AC ratios.

I prefer a simpler approach that just compares the annual output of two different configurations and compare the costs of those two configurations to decide which one is optimal? In your example you obtained a number of hours of clipping but there was no kWs associated with that number so additional assumption would have to be made. That clipping figure in hours would not be useful to rank various configurations anyhow. How would you use just the clipping hours to rank two different configurations without looking at estimated produciton over time?

As I have said before, focusing on clipping is a distraction that keeps one from looking at the important numbers of system production for a given period. The fundemental economics are derived from the kWhs that a system might produce not the kWs that it might have lost due to clipping.

FWIW, I prefer an approach that's as simple as possible (which is not to say it may not be necessary to be more complicated than the uninformed or ignorant may care to consider), while at the same time adding measures or calculation(s) or procedures that may provide what may be insightful information for little or no additional time, cost or effort.
In good engineering design, at least the way I learned it, among other attributes, such additional material or actions are usually considered useful and proactive provided they don't cause a distraction to meeting the project design goals one of which - for residential PV systems - is cost effectiveness.
Oversized inverters are usually counterproductive to cost effectiveness.

To your point of ignoring increases in production that can possibly result from higher DC/AC ratios:
Unless an inverter is initially undersized for the application - however the designer chooses to define undersized - (a situation that I'd suggest should not occur for a well designed system as that term is usually understood), I'm not aware of any way to increase annual system production in a well-designed system as defined here, including optimal cost effectiveness by only increasing an inverter's size alone.

If you (re?)read what I wrote, perhaps you'll be able to understand part of the way I manipulated the spreadsheet by adding a couple of columns to the model's hourly output spreadsheet.
If not, I'll humor you:
One added column was the inverter's size in W minus the modeled array output for a given hour using an "if" statement that made those cells == 0 if the array's output for any hour made those cells' value <0.
So, if the result in any cell in that column was a negative #, it was the modeled clipping for that hour.
The total clipping hours/yr. comes from flagging those hours with negative values with a "1" in an adjacent column and summing that column.

I got both numbers, total annual clipping and number of hours clipped as well as a lot of other stuff like a distribution of clipping by any number of parameters including clearness index should I want to investigate such things.

Note that I obtained the total clipping hours AFTER I obtained the total annual clipping output.
I believe I stated that as well as the reason(s) why I did not include the total annual modeled clipping in a prior post to this thread.

I also believe you, Jeff and I have all written in this thread that each of us is of the general opinion that clipping is only one of the parameters to be balanced in a good design and probably not the first consideration or even near the top of priorities.

So, what's your point ?

I believe I've added some substantive comment and some things to think about to this thread that are useful with respect to clipping. You don't like them ? Sue me.
I've said all to you I'm going to say about clipping in this post.

However, to perhaps clear up some confusion, I'd reiterate a suggestion I made a few posts ago that it may be productive or at least informative to have some discussion about how to qualitatively and quantitatively define just when an inverter becomes oversized or undersized.

If you and anyone else who considers this post to be too wordy, take what you want of it and scrap the rest.
Or, you might work on getting a longer attention span.
Otherwise, give it a rest. Your ignorance is showing.

[ard]

Still plugging away.... However a question/observation:

My current system is 7.7kw (36x215W)

My proposed new system is 144.4kw (36x400)

Basically TWICE the solar generation.... why does PVWatts *not* show me with twice the power... and or my new electric bill being zero? (ie my current system is about half my usage.... doubling should drop me to 0 net. Give or take of course. I think the issue is my usage numbers are not accurate.

Anyway, if someone has any ideas- all ears. Im still working on stuff. (and the paid work seems to get in the way of this.... )

thx

[Ampster]

FWIW, I prefer an approach that's as simple as possible. (which is not to say it may not be necessary to be more complicated than the uninformed or ignorant may care to consider), while at the same time adding measures or calculation(s) or procedures that may provide what may be insightful information for little or no additional time, cost or effort...

I also believe in a simple approach with out all the qualifications of the rest of your statement which actually makes that approach more complicated. What is the benefit of the additional measures or calculation(s) or procedures that is not found in comparing annual estimates? Is there any reason that just comparing annual production estimates would not be a simple approach to use to compare designs?

So, what's your point ?
..

It is very simple and I have stated it several time. I will repeat it and highlight it for you.
I believe that comparing annual production estimates from a model like PVWatts can be used to optimize system design.

It is that simple. I don't think any more words are needed to describe that process. As I have also stated, I also believe that clipping is a distraction that only looks as one impact of system design and ignores the other productivity gains that may occur when a system incurs clipping. However, I understand that, analysis is a discipline in which the analyst is free to chose which methodology works for them and in that context I am only offering this for other readers to decide for themselves. I have no expectation of changing your mind nor do I need to change your mind. I have taken your advice and chosen to scrape most of what you have said on this subject.

[azdave]

I guess our Admin felt the anti spam software was better then having to deal with spammers. Although it seems to catch just about anyone that the software feels is not right.

Every single one of my posts go to moderation no matter what. It's been that way for many years. No one ever figured out why.

[Ampster]

.....

My current system is 7.7kw (36x215W)

My proposed new system is 144.4kw (36x400)

Basically TWICE the solar generation.... why does PVWatts *not* show me with twice the power... and or my new electric bill being zero? (ie my current system is about half my usage.... doubling should drop me to 0 net. Give or take of course. I think the issue is my usage numbers are not accurate.

What assumptions did you make about inverter size in PV Watts? What were the two results of annual production for a 7.7 kW system compared to a 14.4 kW system? How much is your annual consumption?
I entered you parameters in PV Watts and it estimated your 7.7 kW system annual production number is 13,000 kWhs. How does that compare to actual production for the past year in kWhs.
As far as I know PV Watts does not compare dollars so I can't answer your question about about your bill.