I won't try to dispute the numbers in the table you show. But, it is an example of how things get confused by definitions and other things. Statistics don't lie, statisticians do - that type of thing.
Let's start with how long that table says the Redding claim is based on. Looks like 10 years. Not a long time considering 30 years is what is commonly considered necessary to make inferences.
Next, know that % of possible sunshine, while it's an easy number to get your brain around, is better used for planning outdoor activities than solar design, starting with the question of just what constitutes "sunshine". As in the question: Just what constitutes a "sunny" condition" ? How bright ? About 300 W/m^2 irradiance will cast a shadow, but that's a long way from a mid day direct reading of around 1,000 W/m^2. So, does that 300 W/m^2 condition count as part of the % of sunshine or just a "sunny" condition ? Also, even in VERY clear and mostly cloudless locations, about 20 % of the solar radiation received is not direct irradiance, but diffuse, that is, sunlight received after being scattered and filtered by the atmosphere. Any clouds will only serve to increase that percentage.
NOAA and NREL provide % sunshine estimates for info and it has some uses, but to use such records for solar design purposes to claim how sunny a place might be is a misuse of the data and misleading. It's also pretty useless except in some qualitative sense.
Next, one quantitative way of several the solar resource is estimated by knowledgeable folks is to use something called the "clearness index". It's the ratio of how much insolation is received at a location on a horizontal surface compared to how much radiation is received at that same lat. and long. but above the earth's atmosphere. Redding, while sunny, is not # 2 % in that category. The clearness index for the Redding airport using TMY data is about 0.60. That's a nice number. but, that number is exceeded by many places. Inland San Diego is 0.62. Phoenix is 0.67. Most every place in AZ is > 0.65. Albuquerque is 0.65. Most of NM is > 0.65. Pueblo is 0.62. Much of CO is >0.60.
Many locations south of San Francisco and west of, say, Denver have a clearness index greater than that of Redding, that is, > 0.60. Redding is sunny, but it's not the 2d sunniest place in the U.S. as estimated by the methods used by solar energy researchers. Use the TMY databases to get an estimate of annual terrestrial GHI to extraterrestrial irradiance in the same plane. You'll find many locations with clearness indices >0.60.
By way of confirmation of the clearness index data, I just ran PVWatts for a few locations, using a 1kW array, facing south and tilted at local latitude.
Redding: 1,378 kWh/yr.
San Diego: 1,784 kWh/yr.
Phoenix: 1,874 kWh/yr.
Chicago: 1,353 kWh/yr.
Boston, MA: 1,421 kWh/yr.
Even Buffalo, NY will produce 1,237 kWh/yr. - not as much as Redding, but maybe not too far off. Make the Buffalo array 10% bigger and arrays in the two locations will probably produce similar average annual output over time.
-
J.P.M. you got me thinking when you suggested Redding isn't the second sunniest city in the country. I realized it was just something I'd heard around town. I thought it was worth a look for my self. What I found was various sites with lists of the sunniest cities. Redding was second on some lists, but absent from other lists. What I saw was some of the lists are filtered for cities over 100,000 population, or 1,000,000 population. Redding is only 90,000 populatioin. Then I went to the source NOAA and found this. https://www1.ncdc.noaa.gov/pub/data/...pctposrank.txt which does list Redding CA in second place after Yuma AZ.
So it turns out to be true.
The upgrade of my system is moving along pending lead times, design changes, and city permits. I've decided to upgrade to the 7600 model of SolarEdge so I should have margin and won't have clipping. I'm adding two more panels as well. The system was sized for the house but we've added the car with it's 17kW battery.
Leave a comment:
-
FWIW, that 12,000 kWh may seem like a lot for a 6.6 kW system (12/6.6 = 1.82 kWh/STC kW), but possible. SAM has my system (zip 92026, az. 195.75 deg., tilt 18.75 deg.) at about 1.81 kWh/STC kW unshaded, and production seems to confirm that's possible (with the 4+yr.,365 day actual output after ~ 5 % late afternoon shade loss off the total at 9,010 kWh/yr. or ~ 1.72 kWh/yr. per STC kW on a 5.232 kW system). Given the short time frame of 4+ years and the weather variability, reality is a lot closer to models' estimates than I'd have thought.
You and I won't do that well every year, and that annual total will also decrease with time as the system deteriorates, but unless you have a way off south system orientation, or a boatload of shade, or the system size is different (smaller) than 6.6 kW, whoever told you ~ 9,900 kWh/yr. is dancing with your leg.
Not knowing anything else about your application, I'd probably also go w/ a 6 kW inveter, but I'd look at other stuff besides Panasonic - good stuff but maybe overpriced. Above and beyond some basic quality level, panels are pretty much a commodity. They're an appliance - not a lifestyle.
BTW, all production guarantees are useful as marketing/sales tools but otherwise useless B.S. Read the fine print and tell me how claim verification is even possible much less likely.
Now, let's you & me get out of so-cal=burbs thread.Leave a comment:
-
I ran some analysis in PVWatts using a 10% system loss parameter and the results were helpful. For one, it suggests that one of the companies bidding for our business is severely underestimating the potential production of a 6.6 kW system at our location (he calculated 9900 kWh vs PVWatts estimate of just over 12,000 kWh AC system output). Since this company is offering a production guarantee, I can understand why they want to underestimate and overproduce.
I looked at the hourly results as well and it looks like over the course of a year the system would clip an estimated 757 total hours if I used a 5kW inverter, for a total of 271 kWh lost due to clipping. I basically summed up the data points where the DC Array Output exceeded 5000 W. If that's not the right way to look at the data, please correct me!
But it sure seems like a clear sign to go with the 6kW inverter.Leave a comment:
-
To your question about heat generation: Unless conservation of energy and the consequences of entropy have been violated, the electricity that does not get to or is not used by the inverter(s) and is not dissipated within the inverter(s) is converted to (additional) heat, mostly within the array, raising it's temp. above what it would be if all the generated electricity were used by the inverter(s). Any energy of any form that does not make it from the array to the inverter for whatever reason will be dissipated as heat, either by the wiring on the way to the inverter(s) or, most likely and mostly by the array. That electricity will increase the array temp. above what it would otherwise be if all the electrical output of the array were to be used by the inverter(s). That increased temp. will decrease array efficiency.
As for PVWatts system loss parameter: That is a number that combines several points of system loss. See the PVWatts help/info screens. Shading and system availability alone each make up 3 % of that 14 %. If there is no shade and the system is on, the system loss parameter will be (or can be treated as) 14 - 3 - 3 = 8 %. Keep in mind that PVWatts is for system design and the results are best or most accurate for long term (many years) annual output.
10 % is, IMO a more realistic number, but if you're going to take the time to do as you write, I'd suggest you consider using the more detailed input for system losses by clicking on the info button and then the calculator button next to the system loss parameter on the PVWatts input screen.
The 14% is also what a lot of peddlers use as one of many stealth sales tool to B.S. customers into oversizing a system. Think like a peddler. They make money putting systems on property, not saving users money. No one ever got fired for oversizing. A little oversize is prudent. Stacking up fudge factors is unethical. Caveat Emptor.
Many users who have taken the time have found that using 10 % as a system loss parameter seems to produce modeled output that is closer to what their system actually produces. But, due to the nature of the resource, the lack of monitoring equipment and knowledge about what's being measured, reliable quantitative data that's more than anecdotal is pretty scarce.
With that said, many users incorrectly think PVWatts and other solar generation models are predictors of output on a daily basis. They are not.To the extent that the weather (including irradiance) for any particular short time period matches the weather PVWatts or other models use, they can produce output that pretty much matches what a correctly modeled system produces. But if, for example, PVWatts uses a cold, cloudy day's weather and the actual weather that day is bright and cloudless, the modeled and actual output will not be the same. Using longer periods (say, a month) tends to smooth things out a bit, but because short term weather is more chaotic than long term climate PVWatts, or any solar system model is more accurate as the time period under consideration gets longer. See the PVWatts help screens. There, it's stated that actual system monthly output to model monthly output le can be off as much as +/- 30 %, or maybe +/- 10 % when the time period is annual.
An example, FWIW: Using running 31 day actual output for my system against a running 31 day PVWatts modeled output with the system loss parameter adjusted to 8.2 % so that PVWatts annual output matches that of SAM, after ~ 1,560 running 31 periods, the actual to modelled output ratio for those 31 day periods is 0.994 with a std. dev. is 0.099. that would mean that if my data is valid, at the 99% confidence level, the variation in my output for any prior 31 day period was something like +/- 23 to 25% or so of what PVWatts modeled for my system output using a system loss parameter of 8.2 %.
- However, that does not mean that the next 31 days' output, or any 31 day period's output will be within, say +/- 25% of the average. Chances are it will, but climate's what you expect and weather's what you get.
PVWatts can be bastardized to produce a daily or even hourly output, but caution is advised, mostly due the actual weather being different than what the model uses.
For my situation, SAM from NREL, which might be best described as PVWatts on steroids, produces modeled output that's about 5-6 % greater than what PVWatts produces. Another model, TRNSYS output is about the same as SAM. My money's on most of that difference being due to the system availability and shading as mentioned above.Leave a comment:
-
Everyone's red is different. From what and how you write about solar energy and PV, it's somewhat obvious to me that you indeed gained very little from the book. I agree the book may need some updating, and I don't agree with everything in it, but most of the info in there is pretty basic and correct and therefore would seem (to me only) relatively time independent with respect to relevance.
A background in high tech electronics is nice, and worthy of respect, but it's of limited help when it come to the basics of things like resource assessment, solar geometry, use reduction and other things, not to mention, for whatever it's worth, economics. Solar is more about understanding the energy inputs and outputs, and where the come from and where they go than it is about simply counting electrons and potential differences.
When referencing quickly changing technology no printed source stays current for long. The book is a good source for those basic concepts and other matters. FWIW (but probably not much), learning about resource availability can help you discover that Redding, while quite sunny, is not the 2d or 3d sunniest place in the U.S. by any reasonable measure. Far from it.
Good luck.Leave a comment:
-
There is no (zero) reason to go with the smaller inverter for you. SolarEdge does not have the same issues for undersizing as string inverters.
If it is undersized there is no need to dissipate the energy at the inverter it is just not generated so stays at the solar module not inverter.Leave a comment:
-
New member to the forum, and I'm very interested in the outcome for OP as we're looking at very similar potential solar systems, main difference being that ours is still in the planning stages, and my home is in Southern California. We even drive the same car (Honda Clarity PHEV--I knew that I recognized your username from the Inside EVs forum, Jdonalds!). We're looking at a 6.6kW system with 20 Panasonic 330W panels combined with a Solaredge inverter. I'm trying to determine if our system would be best served with the SE5000H or the SE6000H (the installers that I've received bids from indicate that the price is a wash between both models). From what I've been seeing in this thread, it sure seems like I should select the 6kW inverter, which would give me a 1.1 ratio.
My question is, when a system is oversized with a higher DC/AC ratio, does this generate more excess heat in the inverter? I believe I read somewhere that any input energy beyond the rated power output capability of the inverter needs to be dissipated somehow, and it seems that a lot of it would be dissipated as heat. The HD-Wave line of inverters does not have active cooling fans, only natural convection, so the less heat that is generated within the inverter, the better, it seems to me.
I'm going to try running that hourly PVWatts analysis that J.P.M. suggested. Is 10% system loses a more realistic estimate nowadays? The default in PVWatts is 14% and that's what I've been using to run the annual production estimates.Leave a comment:
-
Thanks I'll take a look at PVWatts.
I do expect to gain information once a proper sized inverter is in place. Without clipping I'll be able to see the top of the bell curve of our daily production. While this February has more sun than most years we do have lots of clear sky days here throughout the year. Temperatures are moderate, seldom very cold mid-day. So my thinking is for 3/4 of the year a new inverter will more than pay for itself in increased production. I expect between 1 and 3 KWH per day increase or about 10% on average. There's nothing I can do about the summer months and it will be interesting to see how much power is lost due to hot panels.
I did buy Solar for Dummies. That book is now out of date and needs a rewrite. It is full of duplicate information too. I gained very little from the book. I may be underselling myself calling myself a novice. I do have a 40+ year background in hi tech electronics. It's just that I haven't studied all the in's and out's of solar systems.
I still don't see any of my data on PVOutput. isn't it supposed to automatically pull data from my system?Leave a comment:
-
Jdonalds: Have you run PVWatts to estimate your array's output ? What is your array's azimuth, tilt and zip code "
While the inverter may be undersized, and that sounds like it's being addressed, and while the panels over producing, and that will perhaps adjust downward some as the ambient air temps work to increase panels temps., it looks like you're missing quantitative estimates of possible output.
PVWatts is one way to get a ballpark # for that output. Read the PVWatts help/info screens a few times, use 10 % system losses and choose the hourly output option. Then, make a few runs. Look for days with daylong high output for, say, Feb., and compare the modeled output for those clear and sunny days with what your array actually produced for clear and sunny days on/around that date. The comparison will not be perfectly accurate, mostly due to the wind and air temp. and - even on clear days - some variation in irradiance. But it it may give you a SWAG of what's going on and also perhaps what to expect in warmer weather. If the model consistently or often predicts an hourly output far in excess of inverter rated max. output, while that is not a smoking gun or a "gotcha" kind of thing, it's probably a good piece of information to have. At the least, if the input is reasonably reflective of actual site conditions with respect to array orientation and equipment, the model's output has shown to be a reasonable reflection of likely long term output.
I'll be the first to write that using the hourly output option is a bastardization of the idea behind the model, but if the user is aware of the model's limitations, the model's hourly clear day output on a day can be a reasonable 1st approx. where the actual weather for that day is clear and other conditions are either very similar to what the model is using or those parameters - wind and ambient temp. can be taken into account, provided the limitations of doing so are understood and kept in mind.
Also, a respectful suggestion meant in a friendly way: Get ahold of a copy of "Solar Power Your Home for Dummies". 20 bucks at bookstores/Amazon/etc. A good primer that will help you understand what myself and others are writing about.Leave a comment:
-
Yes we have access to the solar monitoring. That is how I discovered the system was clipping at 5kw. I use the PC application and Android app. I just registered at pvoutput.org with the name "House and Car power". I don't see any uploaded data there yet. I guess it takes some time.
Notes on peak output in February. 1) We've had a particularly rainless February. Normal rainfall is 5.1". So far (Feb 22) we have not had any measurable rain. So we've had a lot of sunshine. 2) Our weather has been running in the 40s to 70s F this month; warmer than average. So the panels must be running cool. 3) Redding CA is the 2nd sunniest city in the United States.
The LG panels are rated at 365W at NOCT of 44C. So we are likely well below that so I'm not surprised the panels are behaving above "spec". I'm pretty much a novice at this so I may not understand what I'm talking about.
Our house/solar panels are very close to due south. Off by just a few degrees.
I can't find the original quote with the cost per watt. I now realize I didn't scan it into the computer so it's filed away in the file cabinet. I'll try to dig it up. But the final cost of the entire system was $25,950 for 6.57KW so I guess that works out to about $4/watt. Right? I don't have the breakdown of costs. That figure of course includes all materials and labor.
My motivation for installing the system (I doubt if I really have to explain it to this forum) was three fold.
1) We have such good sunlight it made a lot of sense. Federal, State, City, and Electric Utility rebates and tax intensives really sweetened the deal.
2) We bought a plug-in hybrid vehicle and wanted to charge it with solar. That is working out quite well. The Honda Clarity (fantastic car) never takes more than 2 hours and 5 minutes to charge which is part solar and part grid. Once the car is charged we are feeding power back into the grid. On balance the car is being charged on solar every day. Sweet.
3) Rather than focusing on the pay back period I focused on other financial benefits.
a. I'm retired and am somewhat dependent on our investments
b. Stocks fluctuate and there is risk in a market portfolio
c. Gains in the stock market are taxable
d. Identity theft could wipe out our portfolio
e. Solar systems save money, they don't generate income. They do not generate taxable income
f. Solar systems are not subject to Identity theft
g. Solar system benefits are likely to improve over time as electric grid rates are bound to rise.
So by installing solar we didn't spend money, we simply invested in something other than the stock market. The returns are lower but they are steady. The money invested in solar is much safer and stable than other investments.
I'm quite happy with the outcome. I just hope our super hot summers don't result in too much efficiency loss.
Last edited by jdonalds; 02-23-2018, 03:03 AM.Leave a comment:
-
Hi jdonalds
You're the first poster that I have seen on this board with the new LG R-series panels. Are you able to access Solar Edge monitoring? Can you confirm that the panels are actually producing above the STC (365W) at peak sun -- in February???
If you don't mind me asking, what was your cost per watt DC installed? Any idea how much the panels themselves were? More than $1/Watt?
Your low light performance sounds impressive. My (37) Panasonic panels (330W) do not seem to perform that well without direct sun. Though solar insolation here in NJ is quite a bit less and my tilt is suboptimal. I also have P400 optimizers with a 11.4kW inverter.
Be sure to link your system with pvoutput.org to share and track your production.
Thanks,
JonathanLeave a comment:
-
-
-
My installer just called. They had communicated with SolarEdge who agrees my system is way overproducing what they estimated. The panels are generating quite a bit more power than their rated specs. So we are going down the path of installing a replacement SolarEdge inverter that can handle the extra power.
When my installer quoted a system back in October 2017 they wanted to use some lower cost lower performing (285w) panels. They wanted 22 panels. I did my own research and selected these rated at 365w which would only require 18 panels. I thought that was smarter because it would mean less chance of shade from trees on some of the outlying panels, and would leave room should I want to install more panels. Also I spent my entire working career in the semiconductor industry and understand quite well the behavior of semiconductors. I knew that if I got better panels they may very well perform quite well. The installer today said the system is producing about 20-25% more than they estimated. This morning it was generating 1.6KW with 100% cloud cover.
We haven't yet worked out who is paying for what. They are offering free labor and paperwork. We're working out a cost for the new SolarEdge inverter.
It looks like I'll have a SolarEdge inverter to sell. Is there a market for these things?Leave a comment:
Leave a comment: