Prince Albert solar planning

[Russel Delainey]

I'm planning to install solar on my property in Prince Albert, SK: https://goo.gl/maps/SqoF9GcaRtQMGfBp8
This project won't be started until at least summer 2020, as finances won't handle it right now. So I'm recording some information about the planning, with the hopes that it helps somebody else later on.
As an avid DIY'er, and reasonably handy, I'm hoping to do the install myself.

My property is not ideal for solar. The garage is small, with one side of the roof facing South. About 10'x25'. Shading is minimal, with only the house blocking morning sun midwinter until about 11 am. Standard 4/12 slope.

The house peak runs North/South, 4/12 slope. So I have an East side and West side. About 15'x45'. No shading expected, will know for sure once I do the sun survey. Pine trees to the East may be tall enough to block morning sun in winter.

So I think that I prefer to place my system on the house, on both sides of the peak. My output will be spread across more of the day. The math is below, but I'm thinking 22 panels on the East side (two runs of 11, in parallel), 11 panels on the West side (one run of 11).

The 7.2kW inverter has two string inputs with mppt. It's going to be lopsided, with 22 panels in one, 11 panels on the other. But still within the limits of the inverter as best I can tell.

I'd like some input on if I should splurge for the 10kW inverter instead. My thinking is that eventually I could purchase another 11 panels, making it even. I don't know how likely it is that I'd be able to get matched panels. I assume that it's a bad idea to match different panels in a string, the inverter will not be able to optimize well. It doesn't make sense to purchase the panels right away, because it's significantly more power than I use. The credits only stay for three years.

Energy consumption from April 2017 to April 2019: 16206 kWh (SaskPower provides a csv export that makes this easy!)
Yearly average: 8203kWh

PVWatts yearly output for 1kW of panel Prince Albert, 4/12 slope, roof mount (not rack)
East: 923 kWh/kW/year
South: 1105 kWh/kW/year
West: 900 kWh/kW/year

PVWatts max DC output per 1kW of panel Prince Albert, 4/12 slope, roof mount (not rack)
(The largest DC output value from the PVWatts hourly data spreadsheet)
East: 774w
South: 849w
West: 778w

Panel power to cover yearly usage
East: 8203/923 = 8.9kW
South: 8203/1105 = 7.4kW
West: 8203/900 = 9.1kW

Inverter sizing
East: 8.9 * .774 = 6.9kW
South: 7.4 * .849 = 6.3kW
West: 9.1 * .778 = 7.1kW

# of 295w panels
East: 8.9/.295 = 31
South: 7.4/.295 = 26
West: 9.1/.295 = 31

Max voltage, cold winter day, -35c, per panel (Thanks for pointing this one out Rob!)
Based on 295w hanwha panel spec sheets
Open circuit voltage@25c+((((temp diff)*temp coefficient%)/100)*Open circuit voltage@25c)
39.48v+((((-35c - 25c)*-.28%)/100)*39.48v)
39.48v+(((-60*-0.28%)/100)*39.48v)
39.48v+((16.8%/100)*39.48v)
39.48v+(0.168*39.48v)
39.48v+6.63v
46.1v

Max voltage, hot summer day, 65c, per panel
Based on 295w hanwha panel spec sheets
Open circuit voltage@25c+((((temp diff)*temp coefficient%)/100)*Open circuit voltage@25c)
39.48v+((((65c - 25c)*-.28%)/100)*39.48v)
39.48v+(((40*-.28%)/100)*39.48v)
39.48v+((-11.2/100)*39.48v)
39.48v+(-.112*39.48v)
39.48v+(-4.42)
35.06v

Max voltage per string (As per electrical code, thanks Rob): 600v
Max panels per string: 600/46.1 = 13.015 panels

MPPT range: 250-800v
Min panels: 250/35.06 = 7.1

Todo:
Get proper measurements of house and garage roofs
Do sun survey from house and garage

Links:
PVWatts - https://pvwatts.nrel.gov/pvwatts.php
Solacity sizing info - https://www.solacity.com/how-to-size-a-solar-system-that-really-works/
Saskatchewan incentives - https://www.solacity.com/solar-rebates-and-incentives-in-canada/#Saskatchewan
295w hanwha panels - https://www.solacity.com/product/hanwha-solar-q-peak-blk-g4-1-295/
Fronius Primo 7.6 - https://www.solacity.com/product/fronius-primo-7-6-1-advanced/
Fronius Primo 10 - https://www.solacity.com/product/fronius-primo-10-0-1-advanced/

[Joe Blake]

Hi Russell, and welcome to the forum.

A very detailed set of data for everybody to read. I have a somewhat similar layout for my place, with my solar hot water system on the (ideal?) northern roof face (I live in the southern hemisphere), so I had to install my panels on the western face. I discussed installing a second set of panels on the eastern face and running the output to my existing inverter. However, I was advised against this, and it proved to be good advice. I eventually installed the second array with its own inverter. At many times during the day each array generates vastly differing amounts of power, being shaded differently - eg in the morning one set generates 100 watts while the other might be generating 500. The technicalities are beyond my accurate description but it seems that the lower generating array "drags down" the output from the other.

Perhaps Rob Becker could give you a more detailed (and accurate) explanation.

Hope that helps.

[Rob Beckers]

Hi Russel,

I wish everyone would do their research like you are! :cool:

I see you read the "PV sizing" article I wrote. Be sure to correct PV-Watts output as indicated. With the shallow roof slope you'll not get much during winter (snow cover), and PV-Watts tends to be a bit optimistic to begin with.

Regarding the MPPT Voltage for panels during the warm days, it's a bit different from what you wrote. It starts with the MPPT Voltage, not the unloaded Voc Voltage. Most spec sheets don't give a percentage for the change in Voltage per degree C, but it's around 0.5%/C. So, starting at around 30V for Vmp at 25C, it goes down to about 24 Volt at 60C (yeah, they get that hot in summer in the sun!).

How inverters responds to too low an MPPT Voltage of the string varies a bit. For Fronius string inverters it's not a big problem, they keep on trucking to a much lower Voltage, the limit really is DC input current, and consequently they can't reach maximum output Wattage any more below a certain Voltage.

Anyway, this is a level of detail most people will quite happily do without!

For net-metering you don't care about 'spreading production over the day'. All you want/need is maximum annual production. For that the south garage roof is better, though east-west can work very well too on a shallow roof like yours, and with the limited space on the garage roof it may make sense to still opt for the main house roof.

Over here (southern Ontario) the rule of thumb I use is that you can oversize the PV array be 115% of the inverter size and really not see much clipping or losses. Beyond that and there's progressively less bang for the solar buck (the inverter doesn't really care, you can oversize a great deal and it will simply clip). You're quite a bit further north, making it less likely to see rated output out of panels, so likely 120% oversize will not really have much effect on annual production. For a 7.6kW inverter that puts solar at 9.1kW before you'll see real clipping (and possibly even more since east-west spreads output over the day).

You can actually combine dissimilar panels, though with some constraints: For a string of panels it is the panel with the lowest output current that will determine the current for the whole string. As it happens for solar panels the Voltage of a 300W panel is just about the same as that for a 250W panel (both being 60 cells), the real difference is in the current those panels produce. So, putting a 250W panel in a string of 300W modules will limit all those 300W modules to just 250W maximum. On the other hand, combining a string of 250W panels with a string of 300W ones (with all the panels in the strings of the same type, so only 250W ones in string 1, and only 300W ones in string 2) will work just fine. Their MPPT Voltages are close enough that there's no real problem in combining them, they each add their respective amount of current to the overall output. For the same reason there's no issue either combining east-west strings into a single set (yeah, the MPPT algorithm will do ever so slightly better if they each have their own input on the inverter, but the difference is very small).

Where things go wrong though is if shading gets involved. Combining partially shaded with unshaded strings has effects beyond just those shaded panels.

That is about it!
And now everyone is thoroughly confused... :weird:

-RoB-