underperformance

[burge124]

Hi guys, my first post on here.

'm a little disappointed by my electric generation I have 14 x250w = 3.5kw system separated across two roofs (see photo, the hip of the roof is due south).
Location is Liverpool 53.41N 2.99W

In December the system only produces 1kw/day
April has been this years best month producing 11.2kw/day
Over a year the average is 6.6kw/day

I think because the system is across two roofs the panels are only working at the weakest link. I'm considering buying another converter and splitting the system but wondered if anyone has other ideas.

Thanks!

[J.P.M.]

In what way is it underperforming ? What instantaneous output do you expect from a 3.5 kW array ? What daily output do you expect as f(time of year) ?

[solarix]

How is your array wired? Two independent series strings or one string for all 14?

[burge124]

For every sunlight hour in April its generating 0.9kw for a 3.5 sytem 25% thought that is low.

Do you think the fact my system is only generating as much as the weakest panel is playing a negative factor?

I believe its one string for all 14, I only have a single converter. Would it be possible to create a second string and connect to my converter?

[J.P.M.]

For every sunlight hour in April its generating 0.9kw for a 3.5 sytem 25% thought that is low.

Do you think the fact my system is only generating as much as the weakest panel is playing a negative factor?

I believe its one string for all 14, I only have a single converter. Would it be possible to create a second string and connect to my converter?

Since the system is probably under some type of warranty, who is and what does the installer say ?

Also, see the net for something called PVWatts. It's a solar modeling tool. Fast learning curve. Read the hel/info screens and use a 10 % system derate rather than the 14 % default value the model uses. Use Liverpool,UK for a location. Run each orientation as your attachment shows and sum the results for total output.

Remember, it's only a model and best useful for long term average output.

Tip/trick: Get the hourly output and look for output on sunny days for dates you know/remember as sunny and have data records for from any monitoring. If you're input is a reasonable representation of reality, the results between actual and modeled will likely be in the same ballpark, but no promises.

[sensij]

If I take some guesses (30 deg tilt, half of the array 225 deg azimuth and half at 135 deg azimuth), using weather from Aughton, PVwatts models annual production of 875 kWh / kW for the SW array and 836 kWh / kW for the SE array. That would work out to something like 8.2 kWh / day on average for your 3.5 kW system

Monthly average production ranges from (0.58 / 0.59) kWh / kW / day in December to (4.83 / 4.7) kWh / kW / day in May. (SW / SE)

Again, that would work out to something like 2 kWh / day average production in December, and 16.7 kWh / day average production in May.

It sounds like your actual performance is underperforming that estimate by ~20%, but my assumptions could be off, or maybe there is shade or other factors I haven't considered.

In any case, the *range* of monthly production you are seeing looks relatively healthy. I think if all 14 panels were in a single string the effect would be worse, but haven't looked close enough at that scenario to be sure.

[kb58]

Assuming the OP is in Liverpool, England, shouldn't the tilt in PVwatts be more like 55 degrees?

To the OP, you said:

...April has been this years best month producing 11.2kw/day...

That's very strange, since the sun gets much higher in the sky toward summer. Assuming you don't have constant clouds, the system's natural peak should have been around mid-June. Makes me wonder about their tilt angle, or as other suggested, maybe shading is going on.

As mentioned, the two sub-arrays should have separate wires running to the inverter.

[sensij]

Assuming the OP is in Liverpool, England, shouldn't the tilt in PVwatts be more like 55 degrees?

To the OP, how much are your panels tilted?

Also, you said:

That's very strange, since the sun gets much higher in the sky toward summer. Assuming you don't have constant clouds, the system's natural peak should have been around mid-June. Makes me wonder more about their tilt angle, or as other said, maybe shading is going on.

Yeah, higher tilt tend to push the best output earlier in the year, but penalize the overall annual production. For the SW array, for example, going from 30 deg tilt to 55 deg (which looks steeper than what the image suggests) would (under the same assumptions I used above) reduce modeled annual production from 875 kWh / kW to 824 kWh / kW. The irradiance data on this wunderground site can be used to show how much the sun was available in April vs May vs June this year, which is worth looking at relative to the data in the weather file used by PVWatts.

[NEOH]

Liverpool England
===============
Scroll down to the CLOUDS Section ...

https://weatherspark.com/y/39878/Average-Weather-in-Liverpool-United-Kingdom-Year-Round

In July, at the peak of Summer there are only 20% CLEAR Days - and that is the BEST that ever gets all year.
Only one clear day every two week during the winter months - avg


Columbus, Ohio
==============
Peak is on Sept 21st, there are 47% Clear Days.
One clear day every week during the winter months - avg.

https://weatherspark.com/y/17263/Ave...tes-Year-Round


Would two MPPT Controllers be able to optimize each String independently to yield more Total Power vs one MPPT Controller?

[NEOH]

Assuming the OP is in Liverpool, England, shouldn't the tilt in PVwatts be more like 55 degrees?

To the OP, you said:

That's very strange, since the sun gets much higher in the sky toward summer. Assuming you don't have constant clouds, the system's natural peak should have been around mid-June. Makes me wonder about their tilt angle, or as other suggested, maybe shading is going on.

As mentioned, the two sub-arrays should have separate wires running to the inverter.

But it is hazy, hot & humid mid-summer and heat reduces voltage.
My higher output days occur in early summer and late summer, not mid-summer ( cool & clear )

[AzRoute66]

The irradiance data on this wunderground site can be used to show how much the sun was available in April vs May vs June this year, which is worth looking at relative to the data in the weather file used by PVWatts.

A hopefully brief detour - I have not seen that irradiance data displayed on wunderground before. I made a quick attempt to just 'change location' from the page that sensij provided to no avail. I presume that it is either because I am not navigating to the correct section of wunderground or that none of the stations that I have been interested in publish irradiance data. I can probably start plowing into it, but am pretty sure just asking the question might be both quicker and more accurate: What do I need to do to see irradiance data relevant to any given location of interest?

[sensij]

What do I need to do to see irradiance data relevant to any given location of interest?

Quick answer, it only seems to show on the daily or weekly view, not monthly or custom views. You have to find a weather station that includes an irradiance sensor. I haven't found a systematic way to do that, just trial and error (not unlike hunting for irradiance in Mesowest collected stations, either)

[J.P.M.]

Assuming the OP is in Liverpool, England, shouldn't the tilt in PVwatts be more like 55 degrees?

To the OP, you said:

That's very strange, since the sun gets much higher in the sky toward summer. Assuming you don't have constant clouds, the system's natural peak should have been around mid-June. Makes me wonder about their tilt angle, or as other suggested, maybe shading is going on.

As mentioned, the two sub-arrays should have separate wires running to the inverter.

While 55 deg. tilt may be closer to a good tilt angle,if staying parallel to a flat roof surface, the roof tilt being fixed means the array tilt (and azimuth) is what it is, regardless of what the optimum tilt might be. Also, and while all other things being equal, max. production will, to a first approx. be from an array tilted at latitude, off south orientations will increase that tilt. Even so, assuming hourly clearness indices are symmetric about solar noon, off south orientations will usually produce less energy/m^2 than south orientations.

The system's peak production season will be a function of many things, and because of the wider range of azimuth angles at high(er) latitudes, and the fixed nature of the arrays, may not be around the summer solstice. Daily GHI may be higher, but daily integrated sum of hourly ave. [P.O.A.* cos (angle of incidence)] may well be lower around the solstice. That sum may well be higher at a slightly lower tilt. The higher angle of incidence (A.O.I.) will, in particular, lower output as the glazing reflectance will be much higher for more of the summer time, in spite of ARC coatings which are mostly tuned to be most effective at A.O.I.'s less than 60 deg. That will lower output. Also, another fly in the high latitude ointment: If higher latitudes do indicate higher tilt angles (and the usually do) those higher tilt angles will mean more array or panel self shading as the solar A.O.I becomes greater than 90 degrees for more of the summer portions of the year. Again, slightly lower tilt angles can help that.

Still, as the latitude increases, the ratio of summer production to winter production becomes greater or about equal reasons of lore summer and less winter irradiance. In the limits, the poles get no winter sun and no summer darkness, while at the other end, the equator has 365 identical (or nearly identical) daylight/darkness lengths of 12 unvarying hours for 365 days. That range of summer to winter production can be tempered (lowered) a bit at higher latitudes with higher tilt angles, but it'll still generally be a higher ratio than if the array was at a lower latitude.

Solar geometry is a fascinating subject, but highly non linear and full of counterintuitive anomalies.

[J.P.M.]

A hopefully brief detour - I have not seen that irradiance data displayed on wunderground before. I made a quick attempt to just 'change location' from the page that sensij provided to no avail. I presume that it is either because I am not navigating to the correct section of wunderground or that none of the stations that I have been interested in publish irradiance data. I can probably start plowing into it, but am pretty sure just asking the question might be both quicker and more accurate: What do I need to do to see irradiance data relevant to any given location of interest?

FWIW, least in the U.S. most WU sites do not have pyranometers. Also, the irradiance/insolation data as published is Global Horizontal Irradiance (G.H.I.)data and must be converted to Plane of Array (P.O.A.) irradiance data to be useful in system analysis or system efficiency calcs.

[AzRoute66]

sensij , thanks. I've a double whammy. I am seldom interested in less than a monthly view, and the few stations I am most frequently interested in don't publish the irradiance data. I guess I will either expand my horizons or just consider finding it a weather Easter Egg.