More later, but the inverter temp. has little if any relation to the panel temperatures. FWIW, this time of year, and from a lot of measurements I've done w/ an IR thermometer from under the array, the ave. array temp. runs about 25 - 30 C. above roof ambient under clear skies at min. incidence angle time, a bit less as f(P.O.A.) off min. incid. angle.
The roof ambient runs about 1- 2 deg. C or so above ground level ambient for early/late day or cloudy skies and ~ 3 - 6 C. above ground level ambient location for clear skies, depending on wind/sun. The panel temps. are primarily F(P.O.A. irradiance, wind vector, ambient air temp. at the array) and a few other minor effects.
What's the orientation of your Davis solar sensor ? If horizontal, that 998 W/m^2 needs to be adjusted to P.O.A. irradiance for what you're measuring. For a bit more precision, the solar sensor reading needs to be corrected for temp. effects (~ -1/1500 per deg. F. above/below 77 F.).
I've got a Davis Pro II +, located about 4 ft. north off the N-S axis of my array. Been measuring/recording this type of stuff for a bit over 2 years or so.
-
Fair enough. Let's take it in pieces though ...
| proof that larger modules on small inverter is proof that clipping gets more production which is 100% not true
Not sure what that means. The paper demonstrates that money is better spent on larger panels than larger inverters. Given that micros tend on the small side this goes against Enphase if anything, I think.
| Larger modules produce more DC, and LARGER inverters produce more AC from the larger DC.
Yes ... but that's a bit of a tautology and sidesteps the point of the paper, which is a larger inverter relative to panel will just you marginally better actual output, from their real-world data. Basically they just clip just under ideal conditions which rarely occur.
| But ALSO F'n true that the same 350W module paired with an inverter large enough to have no clipping will produce even MORE power.
Oh seriously you're still not getting it, despite the bluster. Go read the paper and report back, better yet study it and ponder the different between an integral and a derivative. -
I point out that it is biased because many people bring up that paper as proof that larger modules on small inverter is proof that clipping gets more production which is 100% not true.
Larger modules produce more DC, and LARGER inverters produce more AC from the larger DC.
yes a 350W module paired with a 250w inverter will produce more than a 300w module with the same 250W inverter. True. But ALSO F'n true that the same 350W module paired with an inverter large enough to have no clipping will produce even MORE power. -
Re: bias; OK? The paper makes clear they tested with Enphase inverters, well because they work for Enphase, right? That's what most people do, hell I don't spend my time researching our competitors products, except for a competitive analysis.
Rest of your comment not making a lot of sense to me, sorry. If you read in detail the valid point is that integration counts more with panels than peak. This is the key takeaway, don't get your shorts in a bind that Enphase did some research. -
Enphase paper on sizing inverters is very very biassed towards THEIR inverters.
As an independent I would look at sizing inverters to the modules with an open pool of inverters.
Using this idea, bigger modules AND (AND!) bigger inverter perform far better. But like I stated enphase paper is completely based on the assumption of using their inverter, assumption is inverter size is fixed.
But why would anyone pay MORE for a bigger module and MORE for a smaller inverter, just to get less power than if they payed less for a bigger inverter from a different manufacturer?!?!
As for your clipping, if you are in an area that gets colder weather and some (any) wind then you would have more clipping with the LG315 S280 combination. As it is the clipping will be slight with more moderate weather. Though again the micros cost more than SolarEdge optimized solution.... -
I posted a new thread about panel clipping with data. Since it relates to this inverter and panel I'll do a repost here for general interest, hope it doesn't violate some rule or another
Had some perfect days here and noticed that at peak my system flatlined. Nearly straight lines don't often happen in nature so I suspected clipping, let's see ...
You can see my system below- Enphase S280 inverters
- LG 315W panels
Now to the data- Under STC (Standard Test Conditions) the panel voltage/current gives 315.4W, conditions are 1000W/m2 irradiance and 25C.
- My Davis science grade weather station measured 998 W/m2 at peak, convenient!
- Microinverter temperature at that time is 111F/44C, inverter temp isn’t panel temp but assume it’s close
- Temp derating then gives the panel output at about 90%, so now it’s a 283W panel
By the way Enphase has a great paper on sizing inverters for panels
Their conclusion - bigger panels perform better year round. And in fact, in this case I did a rough integration and figured I lost maybe 500w-1kW or say 25 cents worth of energy at peak, but at non peak am generating more power than I would with smaller panels. Which corroborates their study.
Leave a comment:
-
Fair and honest enough and this is not worth arguing. I respect what you are saying. But from my point in the end. 95 jobs in an isolated area vs a thousand or so, across the state using and comparing the CEC calc, we will see different results. Sorry if I misunderstood anything. -
1.) As I and others have noted, NREL revised PVWatts in response to a lot of comment that it was too conservative with respect to measured and observed reality. Apparently they (NREL) agreed with those comments. You, and/or your data, seem to not agree. So be it. This is a forum where information can be exchanged. Some of that information is opinion.
2.) It seems that we agree that I, you, or anyone can tweak PVWatts, or any model for that matter, to bolster an opinion. Thanks for the insight on a blinding flash of the obvious. FWIW, I'd not call an algorithm that estimates system output a calculator as much as a model. By design, (good) models can be tweaked - for good or for not so good purposes. If for good purposes - that's one way they're improved.
3.) I have a lot of tools. Some are models. As such, those models can be adjusted when warranted by observation and data to better reflect reality. That's what I'd call a good tweak, or at least a tweak with a well intended purpose. A bad tweak might be thought of as one that intentionally decreases how well a model's output compares to that of similar systems in similar environments, often capitalizing on customer ignorance for the purpose of financial gain.
4.) To clear up something: I don't believe I wrote, or meant to imply that the 10% # is set in stone, and I never said or implied "right off the bat". That 10% underestimate is indeed an estimate. FWIW, you seem to agree with that estimate in your statement: " But you are right about the difference at 10 or more %." (your post of 05/09/2016, 08:39 P.M. (sic on time stamp)). Also,a # like that 10% seems a rough consensus among those knowledgeable in such matters and possessing broad powers of observation. Is it 8% ? 12% ? More ? Less ? Maybe. Maybe not. It's a gut, a feel, a dart throw. Whatever it may be numerically, the majority (including NREL, or they probably wouldn't have acknowledged the reasons for the PVWatts revisions as they did) seem to think that PVWatts, and thus the EPBB tool are probably not the best in terms of estimate vs. observation. BTW, while on that part of your above post, I don't think the state of CA uses much of anything to calculate federal energy tax credits as you seem to think. For starters, I believe that is the responsibility of the taxpayer. Also, CA has nothing to do with the federal tax credit.
5.) As for any "proof " : Without seeming to duck out under the cloak of brevity, this is not the place or venue for long streams of data. However, as the guy in my HOA who reviews and recommends action on solar additions to the HOA Arch. Rev. Comm.,I've got about 95 or so files of neighbors in my HOA. Each one contains most all there is to know about their PV systems. Among other data/information, all contain vendor quotes, many contain several quotes from competing vendors. All have copies of the contracts, either purchase, lease or PPA. All have copies of docs required for county/SDG & E approval. All have system performance estimates from both the vendors, and by me - for my purposes only. My estimates include a PVWatts run (by me, or the bidder(s) or both), a SAM run (by me), and often, other estimates based on stuff I've written. All that's where some of my opinion about the ~ 10% number for PVWatts came from.
Some of the rest of my opinion came from comparing output from cooperative neighbors' systems with both PVWatts estimates and SAM estimates. Given the same inputs to the greatest extent possible, SAM estimates of long term system output have been about 8-10% higher than the PVWatts output using TMY3 Miramar weather, the 14 % default system loss parameter, all the while trying to be as true to actual conditions as the models' inputs will allow. As for neighbors' actual output vs. the models' estimates, I've used clear day estimates from the models for actual clear day output on a same date basis and compared that to actual output for that clear day. Sam is usually within a few %. I then "tweak" the PVWatts system loss parameter to match both actual output and SAM' s estimate. The PVWatts system loss parameter usually come in at about 8 - 10% for a reasonable match, vs the 14 % default.
I'm not out to prove a position. I'm looking for a way to accurately estimate current and future system performance. My experience is that the EPBB calculator underestimates likely system performance by something of the order of 10 %, and that a PVWatts system loss parameter of something like 8-10% has shown to be better # to estimate long term average performance than the 14 % #. Opinions vary, but until I find otherwise, that's my direction. I think I'm in good company with that opinion.
6.) Where's the "proof" of your claim: " For the last 14 years that calculator has been the most accurate overall in the predicted output here." ? That door swings both ways.
7.) We also seem to disagree on customer sophistication and knowledgeability with respect to solar energy, its applications, potential and limitations. I'd sum my opinion up by writing if more people had any clue, and knew some of what I think I know, there would probably be fewer PV systems on residential property, and those systems that did exist would be smaller, better designed and more fit for purpose. Or, to use your phraseology, maybe more systems would kick booty. I'll refer you to a pretty healthy portion of the postings to this forum for further reasons for my somewhat skeptical opinions with respect to the general public's knowledge of things solar. It's a target rich environment for shysters.
8.) Please do not continue to try to twist what I write into an inference that I think everyone is dishonest with such statements as: "Everyone who uses this calculator is a scumbag who rips the poor uneducated customer off". Not everyone rips customers off. But to think all vendors are scrupulous is naïve. My apologies to all the scrupulous vendors. Don't get me started on how many of the 30 or so vendors I've dealt with might know something of what they're doing. The brainless ones who hit/run force the good ones to drag the anchor of poor reputation and drag the industry down. At least we may agree on that point.
9.) I am not active in any industry. I have never worked in, what you call, the "PV industry". Since you seem curious, I'm retired with more time than money and more money than brains. FYI, I spent the first 10 years of my adult working life as a commissioned peddler and made a fair living. When my curiosity and interest in solar energy overtook me, I took a pay cut, returned to school, got a BSME, and was trained, worked, licensed and titled as mechanical engineer, designing industrial heat transfer equipment and pressure vessel and piping systems for power plants, refineries and other energy intensive industries. Solar and alternate energy was the driving reason I became an engineer, but no one was building solar equipment in my area when I started the engineering odyssey. Solar has always been more than a hobby and more like a second job. Being retired, and not having any skin in the game, I now get to bloviate all over the planet via the net, pretty much with impunity. Interpret my situation as you please.
Take what you want of the above. Scrap the rest.Last edited by J.P.M.; 05-10-2016, 05:11 PM.Leave a comment:
-
A good source for real world data in our area is Team San Diego, on PVOutput.org. It is very clear when looking at actual system data that the older version of PVWatts, on which the EPBB tool is based, was underestimating performance. NREL has said that much when they reduced the default loss percent in the latest version, and even 14% is too much loss for most systems going up today.
Oversizing a system, whether intentionally or not, costs your customers money. You can choose to perpetuate the ignorance and myths that are embedded in the industry, or you can choose to objectively look at the data and try to help people understand that they don't need to offset 100% of their usage to come out ahead, especially (in many cases) with TOU plans.
PVWatts can not be tweaked to "prove" any theory. It is a model that builds specific relationships between inputs and outputs, using equations and principles established through science. If a user intentionally (or through ignorance) runs the model in ways that don't reflect reality, that doesn't prove anything. The output of PVWatts includes irradiance and environmental data that can be used to assess how accurately the inputs into the model reflect reality.Leave a comment:
-
I'm in SCE area and most of my friends' system are in the same area. SCE only allows 100% coverage when you design you system. Yes, you can expand up to 10% and less than 1kW without re-apply net metering. oversizing/undersizing what what I have observed is pretty much: cash vs lease(monthly payment). Among of my 8 friends who bought their system size "properly" (good size) other 7 friends who is on lease payment undersizing their solar. I know it is not a good statistic with very small samples, but I'll stop right here.
POCOs like TOU simplly they can make more profit off 95%+ of their clients without solar. I like TOU plan, because I can use more kWh without paying more $$$. but I will not build a solar base on TOU plan.
My 3 yrs average is about 9,700kWh prior than solar installed. CSI EPBB calculator estimates 9973kWh and SAM estimates 10,599kWh on a Bosch 255W X 26 with SMA 6000TL-US-12, SW245 deg with 23 deg tilt. I have add 2 more panels within 6 months and gave me 11,500kWh real output in 2015. Am I oversizing? Not really, I still withdraw 3852kWh from SCE in 2015. Yes, I did endup with negative credit and pay almost nothing, but if I'm not in TOU plan I probably had to pay few hundred dollars. I have managed to charges the EVs in super off peak hours. If I cannot switch the loads to super off peak, how much more $$$ do I end up with?
Do your installer know all these infos? If you pay a fee and someone might come up with a more precise calculation base on your hourly usage, but people like to get info for free. If you like to save money, then learn what you don't know. You pay higher price on bottle of water than gasoline, have you complain about the B.S.?Leave a comment:
-
Well your results are a little different than what I see for systems 4-6 years old. Most are performing below what they should be and they were sized with the EPBB tool. For the last 14 years that calculator has been the most accurate overall in the predicted output here. I am sure that you can tweak PV Watts to prove any theory and I have already seen that done. Or you have some calculator that is more accurate than what the State of Ca used for federal rebate calculations. You say it is 10% difference by your calcs right off of the bat, awesome, where is your proof of this claim? And I do not see that in the real world. When I size a system my quote is usually smaller than what the EPBB calculator says the customer needs. It is closer to what PV Watts says. But in the end the customers are usually more educated than you are portraying and want a larger system, that is close to the 10% difference you are referring to. I am not adding another 10% on top either. It seems like you are putting a lot of effort into this to prove what? Everyone who uses this calculator is a scumbag who rips the poor uneducated customer off? Hilarious. Tell me, are you active in the PV industry? Or just spouting calcs from your desk from a few systems around you?Leave a comment:
-
FWIW, I'd rather try to get the model estimate closer to observed reality (whatever that means) than rely on underestimating performance to account for some deliberate and intentional oversizing.
It's a free country and I'm not the Don Quixote that's going to save it. We all get to do what we want. But freely made choices can and do get made by folks ignorant of what they're doing and the consequences that tag along. Oversize as needs/wants dictate, but do so from some position of some consensus as to what estimated long term system output actually is, not from some estimate that can be arguably shown to be oversized out of the gate, and then adding, say, another 10% or whatever on top of what could be argued is already, say, 10% excess. Even before that, know what the need (demand, goal, want... ?), actually is.
Oversizing is not a crime, and the boundaries of just what oversizing is, like performance estimates, are fuzzy. But, if it can be demonstrated to reasonably informed minds that a sizing/performance model consistently underestimates output, I cannot endorse the idea of adding to the size estimate based on such a model. Seems like double dipping - especially when the person or outfit paying for the oversizing may not know that the oversizing they want is already (and often/usually deceptively) in the estimate. That, IMO, may be a crime, or at least a scumbag, deceptive B.S. move. All that said, I'm certain it happens more often than not, and to the extent it happens at all, It still sucks.Leave a comment:
-
[QUOTE=sensij;n314621]There is another CEC calculator that I thought ncs55 might be referring to, that could conceivably be more accurate (as opposed to the version that is based on PVWatts V2). Link here:
http://www.gosolarcalifornia.org/too...ator/index.php
Looks like ncs55 confirmed the EBPP version.
Yea, I'm pretty sure the new CSI calculator is based on the TRNSYS model. I had a license for that until about 5 yrs. ago and used it for many years for solar thermal modeling. TRNSYS also does PV, and the economic modeling is pretty sophisticated, but was still sort of based on Chap. 11 of Duffie & Beckman (and that's not a knock) when I used it. I'm sure it's quite improved by now. Haven't downloaded the CSI application (yet). If you do before I get to it, please post comments. Thanx,Last edited by J.P.M.; 05-10-2016, 04:55 PM.Leave a comment:
-
I see the opposite, especially with new systems. Also, my very limited experience is that 4-6 yr. old systems produce about the same as when new, but that's more anecdotal than documented. Even so, given that the CSI calculator seems to oversize as much as it does - both for my system and others I've seen, usually by 10+ % or more - I believe it will take a long time for system degradation to catch up with such low estimates.
We size systems for what the customer needs, it is usually the customer that wants a bigger system and we have to requote. And 10% is pretty close to what they ask for over the original quote. some are adding a garage, some are having family move in or kids coming home for college etc. So, a 10% oversize is not uncommon nor uncommon practice to use that specific calculator for some companies. But you are right about the difference at 10 or more %. For system production and degradation. That is a little different and can vary widely with from many factors including product matching and quality, design and or installation techniques etc. We see that systems with modules installed a few inches from the roof, suffer faster degradation and loose performance quickly and in less than 6 years most of the time. Systems where the modules are properly installed at least 8 inches above the roof degrade slower and kick booty! We always jack the modules up as high as the customer allows, but never less than 5 inches off of the roof for proper airflow and cooling. These low mount racking systems look great but foster loss of power right from the start. Airflow and module height vs performance is being documented by a university professor and the findings will be out soon. Depending on the installation and or system design we see high degradation to failures usually in the first 4-5 years. For proper installations you would be correct it takes a longer time frame to notice it.Leave a comment:
-
Anyway, the idea that a system must be sized the achieve 100% offset to be successful is another fallacy perpetuated by solar installers. "Too small to cover their consumption" is in some cases the most cost-effective way to size a system.
I was referring to the EPBB Calculator although the NSHP calculator is handy sometimes.
In the beginning 16-17 years ago yes, in some cases you are correct. Peak shaving was good for awhile, until the utilities changed their tiers and rates. With the new structure coming with SDG&E, it will not be feasible to peak shave profitably. Sempra is out to stop grid tie solar because it has cut to much of their profits down. The battle has only just begun here with Solar and Utilities. I remember when the good folks in Germany had this problem and now they have solved it by going to municipalities.
Leave a comment: