Slowing Down the Summer Sun

[J.P.M.]

JPM,

"software capabilities"? "thermal design parts"? Yes, I shall certainly attempt to stay away from that in the future. Thanks for the heads up. Could you suggest which "resources availability" I should concentrate on? Any clarification appreciated, so far I think I've only looked at sunlight..

Respectfully, I believe you are out of your knowledge depth. I don't have it in for you, but you are lacking knowledge in the basics. Because of that, a lot of what you are doing is at least inconsequential to a meaningful analysis, but given where you seem to be knowledge wise at this time, I'm not sure you can see that. If you backed up and understood the basics, IMO, you'd see what I'm writing about. Do as you wish.

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

[ButchDeal]

Your 3-5 day worst case observation. Too big of a can of worms for me. Just as chained inefficiencies in a system add up, 0.98 x 0.96 x 0.99 x 0.94 always comes out lower than you think, over use or inappropriate placement of 'worst case' and 'fudge factors' can make something over-engineered to the point it loses feasibility.

Your entire analysis is based on backwards use of fudge factors....

PVWatts has always enjoyed a good record on the macro scale, I'm going to ride it, at least until it starts asking me what global warming model I wish to use. I think this addresses any interest I might have in the TMY data as well. Let's say I find those bad days, how will they impact insolation? Can I expect to know better than PVWatts does?

Can you know better than PVWatts? some of us do. It is just a model, no better than the setup and data put into it. SAM is a better model. Aurora has an even better model.

BTW, PVWatts uses TMY so as Sensij mentioned, picking the worst days and modeling them is important because those worst days ACTUALLY HAPPENED.
Thats how TMY is made, they pick typical days out of the past set of years (number of years varies from location to location) and stick them together to make a Typical year. so Monday could be from 2015, Tuesday from 2000, Wednesday from 1999, Thursday from 2010 etc. Often the bad days (atypical) are grouped together each year.

[AzRoute66]

sensij,

I'll try to address your points in order. As far as SAM goes, when I read the NREL blurb on SAM it very much looked like it was more of a cost analysis platform for different rate structures, etc. that was built on a PVWatts engine. I was so eager to get to my analysis I went with the tool I had actually used once upon a time and for what I needed it worked great. Adjust the azimuth, copy a column, rinse and repeat. Took longer to create the graph template than pulling and populating the data. With your input on it's parametric capabilities I plan to look at SAM probably this weekend.

I have seen you advocate the SW - SE configuration. From what little I just did I think that from a cumulative harvest priority aspect that might indeed be very close to right for a wide range of latitudes. As you can see, I was just trying to maximize the duration of Bulk Charge capable sunlight for my location, and some cumulative harvest observations fell over the side. It would be easy to design the data pull for PVWatts. If I knew the API it would have been two minutes or less of code. I wouldn't try that until after I have evaluated SAM as it sounds like SAM would be a better tool for that. Here is the $64 question. What 'large' off grid application would be the poster child for that?

Your 3-5 day worst case observation. Too big of a can of worms for me. Just as chained inefficiencies in a system add up, 0.98 x 0.96 x 0.99 x 0.94 always comes out lower than you think, over use or inappropriate placement of 'worst case' and 'fudge factors' can make something over-engineered to the point it loses feasibility. PVWatts has always enjoyed a good record on the macro scale, I'm going to ride it, at least until it starts asking me what global warming model I wish to use. I think this addresses any interest I might have in the TMY data as well. Let's say I find those bad days, how will they impact insolation? Can I expect to know better than PVWatts does?

"sufficient autonomy". All off grid conversations swing around that like a pendulum. I must admit, I learned enough about batteries from this forum to start to get an appreciation for how much I have to unlearn to get comfortable with them again. Electrically I have always known them as well as I will ever need. Chemically I am just getting started I think. Practical knowledge is somewhere in-between. My little 4 kW 'system' I was outlining above will get a 'supply side' battery treatment soon. I need to learn more and I am still contemplating my hypothetical load case as well Probably after my impending rant on charge controllers gets shot down. Sunking (and all of you with interest) will hopefully grade my battery efforts well. Once I have a better handle on batteries, I'll start pondering harder on autonomy.

Finally, the Chris Olsen philosophy. I read his sticky in off-grid, and a few of the others. He sounds like a guy who knows how to 'let go' when something doesn't work for him - a rare quality. Generator, solar, and wind. I can't believe he doesn't have (or at least hasn't mentioned yet) how he created an 80' head off of the creek running through his hardwood acreages to power a hydro mini turbine package if you know what I mean. He is using the his generation in a balanced approach it seems, and his utilization of two different sized generators I thought was a real step outside the box. If I was designing an off-grid lodge I would try to mimic what he has done, but I am just designing for fun and the learning aspect, and having the magic generator bullet would take too much of the fun out of it right now.

Thanks for the input on what I did. I suspect I threw too much out there at once. When I saw that Bruce had added 2.5 hours in front of his south array output to reach a 50% power point I was amazed. Now I am amazed that so few seem interested. Even though it is limited in winter, it is a powerful off grid concept.

[AzRoute66]

JPM,

"software capabilities"? "thermal design parts"? Yes, I shall certainly attempt to stay away from that in the future. Thanks for the heads up. Could you suggest which "resources availability" I should concentrate on? Any clarification appreciated, so far I think I've only looked at sunlight..

[J.P.M.]

AzRoute66:

I don't know where you got your training or information, but from the post I just read, I'm of the opinion that your knowledge of the basics of solar energy, and what available software is and is not capable of doing, is way below what you need to continue. A lot of what you just posted just doesn't seem to have much, if any basis that folks knowledgeable in solar energy would accept as useful.

I certainly wish you luck, but from what I've seen of your stuff, FWIW, I think you need to start at the basics. You're spinning your wheels. Read Duffie and Beckman. Skip the thermal design parts and concentrate on resource availability for a start.

[sensij]

Since the total power harvested is now in play, we'll go back and cipher it the same way for December and it comes to 142 kWh / 31 = 4.58 kWh/day. Not much, but it would cover my freezer with a about 800 Wh to spare. Nothing to sneeze at during short winter days. And, if it turns out to make sense to the East, it should also to the West, but I know PVWatts knows that clouds in this area form in the afternoon and in the West so will have to do those numbers independently but in the exact same manner.

One suggestion... monthly averages are nice, but you might want to key in on the worst 3-5 day stretch within each month that the TMY files present. If you are really into it, you can even look at Wunderground history and try to assess how well the TMY file is truly representing that worst case. After all, if your battery is sized with sufficient autonomy, the worst-case day alone doesn't kill you, it is only when they string together that you need to have a backup plan if you aren't sized to handle it.

There is another whole philosophy to off-grid design, finding a happy balance between energy from a generator and from PV, if you can find the posts by ChrisOlson

[sensij]

Of course, this whole novel was presented so that the assembled notables can shoot gaping holes in it. I exist for target practice. I hope everyone presents their observations and objections.

I don't have any objections. Consider using the parametric analysis in SAM to save yourself from the drudgery of running PVWatts every time you want to check out a new orientation. In that software, you can vary both tilt and azimuth in a single run and find the true maximum for both. It may be hard to pick out from the history of the forum, but I've consistently suggested that SE+SW may be a better choice for large off-grid (or grid-tie backfeed limited) array design, but it is rare for that to actually be a viable choice in typical rooftop installation, and isn't very space efficient in a ground mount (compared to the E/W assembly that bcroe demonstrates).

[AzRoute66]

[...]but a single orientation array with the same area under the curve will have fewer panels for the same production for the same period that's longer than a few days or a week or so.

Amen brother, but will it help you keep from undercharging your off grid batteries? Let's throw some dead standing pine on that fire.

Below we see the first of many pictures that may help us answer this question. The values within are from PVWatts. All lines represent a 1 kW array, in Tucson, tilted at 65 degrees (arbitrarily selected as the midpoint between 80 degrees where the sun is just over the horizon groundclutter and 50 degrees which is the highest the sun will ever be when it is due East). I am going to rerun the analysis for best tilt angle at a later time. PVWatts does not designate 'times', only hours, so when you see the time 0600 on the graph it actually represents 'Hour 6' in the PVWatts data --- close enough, I'm not going to get into a Daylight Savings Time or is 'noon' actually noon conversation.

So what I did in this picture is I took my 1kW array, on 21 December, and every hour instantaneously slew the azimuth from 70 degrees to 140 degrees in 10 degree increments and measured the wattage of the array. I also took the wattage reading from my due south facing 22 degree tilted fixed 1kW array for reference. The weather that day could only be described as 'typical for the area'. My lovely assistant and I were very diligent.

The first thing that jumped out at me was that the 70, 80, 90, and 100 degree azimuths are just 'non-starters' for winter sun, I don't care how well they perform in the summer when the sun is actually rising in the north.
PVW_Winter1.JPG


So, before we debate the detailed merits of the 110 - 140 azimuths, let's take note that none of the angles are going to give us anything for Hour 7. In hour 8, we can say that 625-ish W could be made available, and in hour 9 there are 750-ish W and that is as good as it gets.

We sidetrack here just for perspective. I already know that the 1 kw south facing 22 degree tilt array is ACTUALLY a 4.0 kW nominal array. How do I know this? Because I am a supply side designer. I measured the only south facing roof and determined that 77" panels are too long to efficiently cover the surface, I judged that the smaller 100W - 200W panels would require excess racking and connections, and that 16 of the 65" panels will fit just right with setbacks and racking considered. The most likely deal on a pallet of 65" panels will be in the 250W - 280W range, so I say a 4kW array is my design minimum understanding that it could go up to 4.5kW with a single entry on eBay. Several people just fainted, "There is no way you can specify an array without knowing the LOAD", they exclaimed as they collapsed. When they wake up, tell them I am a supply side engineer and they will learn to live on the amount of electricity I decide to give them, just like the POCOs. As an aside, this is also where I decided on a 48V battery system (without giving one little damn about no stinking loads). 18 more people just fainted - oh well. Also, when I did that roof measurement, I noted that there is plenty of available space for east and west facing panels, otherwise this whole analysis would never have been performed. A glance told me there were no shade issues or I would have done a shade analysis at that time.

Back to it. In Hour 7, my fixed south array is making nothing x 4 which is still nothing, in Hour 8 we have 286.364 W x 4 = 1145W. Great, but what do we need? Remember, the goal is to lengthen my charge day. What we need is Bulk Charging Current plus. We don't know the 'plus' at this point but we do know that optimal maximum bulk charging current is 20 Amps. 26 more people just fainted, leave them on the floor. I decided 8 six volt batteries will be a good starting point for a 48V system and 200 Ah is a reasonable starting point to look for 6V battery deals. C/10 = 200Ah/10h = 20A. Get over it. It will all work out later, I promise. So we need 20 Amps x 48V = 960W. In hour 8 we are making 1145W which is plenty just from the south array. So, from nothing in Hour 7 to plenty in Hour 8.

Am I going to add four 250W panels (and associated power management) just to get some 650ish load watts for ONE stinking hour. No way. Even if my bulk charging amps was 10A [not happening], I would still say no way. I'll tuck that extra 625-ish W in my back pocket for use if my clients get 'testy' about how late their 'loads' electricity is arriving. There, I said the word, should make somebody happy.

Now, I am not going to make eight PVWatts runs and collate the 8x24x365 data elements just for that tiny bit of somewhat intuitive knowledge. While the ink is still wet, I'm going to look at the summer numbers as well.
PVW_Summer1.JPG


Whoa. Looks much better, much more 'usable' [Of course it is, it is in the middle of SUMMER, you Solarholics.] As expected, our South-ish azimuths are now clearly the worst and we can eliminate many of them, specifically the 120 - 140 azimuths. Guess what is left if we combine that with what we did for the winter, we are left with the 110 degree azimuth as the best choice for a year round East array that favors winter just a skosh. THAT is good progress. When (if) I make the run varying tilts to find the optimum I will use a fixed 110 degree azimuth array. And now I can refine the 'ishes' in the winter analysis above if I ever want to..

What do we have here? I go back to the numbers for the 110 degree azimuth and add up the watts from Hour 5 to Hour 19 and I get 5,578 Wh/day. I don't know what that means. If I look at the 180 array for 1 kW, I get 7196 Wh/day. So in the summer it produces a little over 5578/7196 x or 77% of what a south array does. Just looking at the area under the curves I would say that looks about right. Also, I know that the sun is already 50 degrees high in the sky before it ever even gets to due East (at 0923 hours) so I know there is lots of morning sun that never hits a south facing array. No foul - it is at 22 degrees elevation and actually high enough to hit the panel from the rear at 0715 hrs, I recognize that. Sanity check time, PVWatts reports 201 kWh for the month of June for this array. 201 kWh/month / 31 days/month = 6.48 kWh / day. Sounds right, after all, our 7.2 was exactly on the solstice day.

Now, OUR south array is actually four times that big so 4 x 6.48 kWh/day or 25.92 kWh/day for June, 28.8 kWh for the solstice day.

WTF?!?! I did all that and I still don't know what it means. Back to June 21 only. East array of 5,578 / 28,800 main array = 19%. Over the course of a day, our 1 kW 110 degree azimuth array will generate 19% as much power on the solstice as our 4k array. Sounds like a lot. How many athletes attain 119%? - the average one only operates at 110%. It doesn't help us with the length of day charging situation, which is predicated on winter days, but it looks like something else we'll want to keep in our back pocket because we know in the summer we have to cool down this hut in Tucson and that will take some kind of cooling (evaporative cooling, probably 1/2 hp fan motor, as I noted when I measured the roof). We also know we can't squeeze another square inch of panel onto that south facing roof if we sit up and beg.

Since the total power harvested is now in play, we'll go back and cipher it the same way for December and it comes to 142 kWh / 31 = 4.58 kWh/day. Not much, but it would cover my freezer with a about 800 Wh to spare. Nothing to sneeze at during short winter days. And, if it turns out to make sense to the East, it should also to the West, but I know PVWatts knows that clouds in this area form in the afternoon and in the West so will have to do those numbers independently but in the exact same manner.

For length of charging day, we had to go to worst case (Dec 21), but now that we're interested in using it to cool the place in the summer we can't use the summer solstice very best case (June 21) to make our decision. We can assume that power available is a linear straight line from December to June for many rough guesstimations, but if our clients decide to proceed, we'll have to go back to the numbers and do some more quick calculations to cover the month to month power production. The numbers are already there waiting.

We know a boatload more than we did before we started, and most of it would apply to any install in this area. We know where to smartly point an East array and what would impacts of any given number of panels, we will shortly know about the West, and can legitimately talk about capability, options, and cost. Did I mention that while I was measuring the roof, a gander at the backyard revealed that there is room along the north fence for a six panel ground mount facing south - which would also be a nice landing spot for any leftovers on that pallet purchase for the 16 roof panels. Time perhaps for a shade analysis before we meet. Supply side rules !!!

Let's go get 12 months or more of electric bills from our client because there is much more information we can offer to them without knowing if they leave the fridge door open or not. We'll hand them a Kill-O-Watt meter and a checklist while we're there. (AND THE CROWD ROARS !!!!)

Of course, this whole novel was presented so that the assembled notables can shoot gaping holes in it. I exist for target practice. I hope everyone presents their observations and objections.

[J.P.M.]

No argument, but your situation isn't my situation. Meanwhile, the science experiment goes on. Bruce Roe

And make no mistake, I support your efforts.

[bcroe]

Bruce: I appreciate as much or more than many,if not most, that a lot of what you do is for other than economic reasons. I've done the same for years. FWIW, I honestly applaud you for it, an I've got a lot of respect for your experience and electrical engineering knowledge. But the idea that justifying less than ideal panel or array placement by lower panel costs doesn't work for me. Weather considerations aside, or for battery charging considerations, keeping arrays in close to or the same orientation, with that orientation being the one that maximizes the solar resource for the period under consideration usually means on orientation. For grid tied systems and probably for most applications, close to equator facing will result in the least cost for the panels because you will need fewer of them for the same energy yield. Unless you have limited ability to accept the generated power, the least cost system will usually have one orientation with lower costs for fewer south facing panels. You have a lot of curves with flat tops for multiple orientations, but a single orientation array with the same area under the curve will have fewer panels for the same production for the same period that's longer than a few days or a week or so.

No argument, but your situation isn't my situation. Meanwhile, the science experiment goes on. Bruce Roe

[J.P.M.]

I guess directed to me, glad it came through. If you use the QUOTE button, possibly edited down
to a sentence, I could be sure. Others will point out you can generate the most energy from a panel
by pointing it south, but with panels now so cheap, other consideration might dominate your app. its
still experimental, will be digging a foundation again soon. Killing phantom loads was a major project
over a couple of years, but with better results than seemed possible. Bruce Roe

Bruce: I appreciate as much or more than many,if not most, that a lot of what you do is for other than economic reasons. I've done the same for years. FWIW, I honestly applaud you for it, an I've got a lot of respect for your experience and electrical engineering knowledge. But the idea that justifying less than ideal panel or array placement by lower panel costs doesn't work for me. Weather considerations aside, or for battery charging considerations, keeping arrays in close to or the same orientation, with that orientation being the one that maximizes the solar resource for the period under consideration usually means on orientation. For grid tied systems and probably for most applications, close to equator facing will result in the least cost for the panels because you will need fewer of them for the same energy yield. Unless you have limited ability to accept the generated power, the least cost system will usually have one orientation with lower costs for fewer south facing panels. You have a lot of curves with flat tops for multiple orientations, but a single orientation array with the same area under the curve will have fewer panels for the same production for the same period that's longer than a few days or a week or so.

[bcroe]

I weathered the storm of photobucket pop-up ads and saw all of the great pictures of the EW Array. Aptly named. I would have thought of that configuration and dismissed it as people would have made fun of me, but your 'south' array in the background alleviates that I'm sure. Finding a spot that has a low elevation view in both East and West was nice from a main wiring run standpoint. Thanks, it inspired me.

I guess directed to me, glad it came through. If you use the QUOTE button, possibly edited down
to a sentence, I could be sure. Others will point out you can generate the most energy from a panel
by pointing it south, but with panels now so cheap, other consideration might dominate your app. its
still experimental, will be digging a foundation again soon. Killing phantom loads was a major project
over a couple of years, but with better results than seemed possible. Bruce Roe

[AzRoute66]

I weathered the storm of photobucket pop-up ads and saw all of the great pictures of the EW Array. Aptly named. I would have thought of that configuration and dismissed it as people would have made fun of me, but your 'south' array in the background alleviates that I'm sure. Finding a spot that has a low elevation view in both East and West was nice from a main wiring run standpoint. Thanks, it inspired me.

[inetdog]

My wife had me add an array with an eastern orientation, to get power faster in the morning for toast. Electric Toaster is the bees knees. Stove top, pan fried, all fail.

Have you tried an infrared type propane heater and a frame to hold the bread? You do have to turn the toast manually though.

[bcroe]

Well, examining the other threads on a 'virtual tracking' search yielded little besides mentions, and explanations, no numbers. I've only scratched the Google-sphere on this topic. So, another request for pointers to analysis, case studies, simulations is probably called for. That said, it might not be needed. The graph in Bruce's post #10 does just about all I need to see. Enough to motivate me to use PVWatts to check out a few specific scenarios anyway. Any other tools that would be quick and useful? What is 'SAM'?

Bruce - Thank you so much for that perfect graph in post #10. You know I'm interested in the location, configuration, etc. I'm pretty sure that at some point in my burrowing I saw you have a website resource where I can get more of your stuff. If so, would you mind re-mentioning it here?

Told you in my intro post that I love me some homemade trackers. The more Rube Goldburg the better. If I had the space for a single axis tracker array, I'll bet I could rig something practical for my own use, maybe even with some seasonal tilt.

The graph in 10 is close to being the best. I ran curves periodically with 3 matched test panels, over
the last half of 2016. Getting the center the flattest was somewhat critical, about 59.5 degrees tilt on
the E-W panels on the longest day. Tilt adjustments went on, but seemed to be less useful approaching
the shortest day. My conclusion was to go for best performance in the longer days; the shortest days
tend to be so dominated by clouds that tilt hardly matters. Instead snow will be the dominating
consideration; near vertical panels.

The graph in 4 ought to get broader, once I get all the strings moved to avoid premature shadowing.

The location here is 42 deg 08.403 min N, 89 deg 13.561 min W, Northern ILL. The E-S-W array
shows on Google Earth.

Experiments with trackers should be interesting, show pictures. See some stuff on conserving energy
and building arrays on my PHOTOBUCKET site album ENERGY CONSERVATION. Here, click
BCROE, then ABOUT, BASIC INFORMATION some half way down is INTERESTS which has
the link, since links won't go thru the posting process here. Bruce Roe