my solar output ... decreases over time ?

[sunmercury]

I have 4KW Canadian Solar System installed backed in March 2014. Lately, I have observed the output going way down. I understand it's common to get less output as we have less suntime compared to summer. However, my concern is peak output. For example, my report shows that output is 871 watt-hours on June 04, 2014 – 2:00 - 2:15 pm. Lately, it shows 613 watt-hours on November 03, 2014 – 12:45 - 1:00 pm. In any case, the peak output for any given day lately is much less compared to a few month backs (which happens to be summer). I contacted my installer, and they said it's probably "weather related"

Has anyone observed the same thing ? I am in Southern California, btw. From what I read, the panel should be more efficient during colder weather. Any input is appreciated.

[sensij]

Sounds pretty normal. Using PVWatts, which is not matched to the weather for any exact day, but is instead "typical", on June 4th the system modeled 804 kWh as peak 15 min energy output. On November 3, it modeled 622 kWh. Not only do you have less suntime total, but the sun sits lower in the sky during winter, affecting your ability to harvest it. With a difference of < 10% between model and actual and absolutely no attempt to match the model to your exact installation, it looks fine to me.

[J.P.M.]

I have 4KW Canadian Solar System installed backed in March 2014. Lately, I have observed the output going way down. I understand it's common to get less output as we have less suntime compared to summer. However, my concern is peak output. For example, my report shows that output is 871 watt-hours on June 04, 2014 – 2:00 - 2:15 pm. Lately, it shows 613 watt-hours on November 03, 2014 – 12:45 - 1:00 pm. In any case, the peak output for any given day lately is much less compared to a few month backs (which happens to be summer). I contacted my installer, and they said it's probably "weather related"

Has anyone observed the same thing ? I am in Southern California, btw. From what I read, the panel should be more efficient during colder weather. Any input is appreciated.

You will have less output at the same time of day during winter months as summer months because the largest component of the of the solar radiation - the direct beam component - the stuff that casts the shadow- will fall on the panels at a more oblique angle. Therefore, less solar "fuel" will be available.

The decreased output you are noticing is completely normal.

Assuming clear day conditions for a minute, not only are there fewer hours of sunlight for each winter's day, but the solar flux (intensity) for each hour is less intense than at the corresponding hour in the summertime ( BTW: don't forget about Daylight saving time, or the other smaller adjustments necessary if you want more accuracy on the solar angle of incidence calcs).

The panels do operate more efficiently when they are cooler. However, while they will probably operate more efficiently in winter due to the generally lower temps and lower solar flux, they will have less fuel (sunlight) to convert to electricity on an instantaneous basis.

An example of how the irradiance can change for those dates/times without the gory details: At my location in 92026, the measured global solar irradiance radiation on a horizontal surface on 06/04/2014 ( a "sunny" day) at 2:07 P.D.T was 958 W/m^2.
On 11/03/2014 (another sunny day) at 12:52 P.S.T. the reading was 606 W/m^2, or about 63% as much as for the June date/time. That ratio and the irradiance levels will be different for non horizontal panels, but I'd guess not too much different for most residential panels of typical orientation in S. CA. That's mostly why your output is less in Nov. than June.

A "clear sky" algorithm (Bird & Hulstrom) gives theoretical values of 979 W/m^2 and 612 W/m^2 respectfully. Most days are not that clear - even sunny ones.

[Sunking]

Lower angles means more atmosphere between you and the Sun which attenuates the Sun's intensity.

[J.P.M.]

Lower angles means more atmosphere between you and the Sun which attenuates the Sun's intensity.

True, but that's not the biggest reason for the difference under consideration. The biggest reason for the difference (assuming clear conditions and a stationary array) is due to the difference in the angle of incidence for beam irradiance on the system for the dates and times given.

One way to show this is to look at the Beam Direct Normal Irradiance.

Approx. values using the same Bird & Hulstrom method for my location including elevation:

06/04 2 P.M. P.D.T Beam Direct Normal=~ 885 W/m^2.

11/03/ 1 P.M. P.S.T Beam Direct Normal =~ 825 w/m^2.

Get me a lat./long., coll. az. and tilt, and I'll estimate clear sky irradiance in the plane of the collector including the beam, diffuse components and the reflected irradiance if given a reflection %.

[sunmercury]

Thanks to ALL input, especially very informative explanation from J.P.M.

[inetdog]

Thanks to ALL input, especially very informative explanation from J.P.M.

Just to point out one additional consequence of the panel insolation factor: If you were able to change the tilt of your panels to match the lower sun angle, the peak production would again match the summertime peak except for additional smaller factors such as the longer path through the atmosphere and weather effects. The total output would still be lower though, because of the fewer hours of strong light.
The problem with trying to change the tilt of an array of panels on your roof is that if you tilt the individual panels they will end up shading each other and the production will be even lower. If you try to tile the whole array, you are into major mechanical issues.

[J.P.M.]

Just to point out one additional consequence of the panel insolation factor: If you were able to change the tilt of your panels to match the lower sun angle, the peak production would again match the summertime peak except for additional smaller factors such as the longer path through the atmosphere and weather effects. The total output would still be lower though, because of the fewer hours of strong light.
The problem with trying to change the tilt of an array of panels on your roof is that if you tilt the individual panels they will end up shading each other and the production will be even lower. If you try to tile the whole array, you are into major mechanical issues.

To those points:

One common use of estimating programs such as PVWatts is to change array orientation - tilt and azimuth to estimate the orientation that results in optimum (maximum ?) yearly output or min. elec. bill. Panel arrays set parallel to the roof surface they sit on are usually not at that optimum orientation. It's common to do it that way because it probably looks less offensive to many folks, it's easier to install and thus has become the "accepted" and therefore the "correct" way to do it. Most folks are clueless about orientation. Fortunately, off optimum orientations, sometimes by a fair amount, are not usually disasters output wise.

Changing an array so that panels are arranged in rows, sort of sawtooth pattern or similar is sometimes done, individual panels less often. In either or any case, such arrangements usually cost more up front and, as Inetdog writes, care is needed to avoid induced self shading of the array. Also, and most often but not always, such adjustments are not worth the extra up front cost in terms of money, appearance and aggravation when compared to the potential extra production attainable.

[J.P.M.]

Thanks to ALL input, especially very informative explanation from J.P.M.

For my part, you're welcome. Hope it helped.

[paulcheung]

True, but that's not the biggest reason for the difference under consideration. The biggest reason for the difference (assuming clear conditions and a stationary array) is due to the difference in the angle of incidence for beam irradiance on the system for the dates and times given.

One way to show this is to look at the Beam Direct Normal Irradiance.

Approx. values using the same Bird & Hulstrom method for my location including elevation:

06/04 2 P.M. P.D.T Beam Direct Normal=~ 885 W/m^2.

11/03/ 1 P.M. P.S.T Beam Direct Normal =~ 825 w/m^2.

Get me a lat./long., coll. az. and tilt, and I'll estimate clear sky irradiance in the plane of the collector including the beam, diffuse components and the reflected irradiance if given a reflection %.

In Jamaica, because we don't have NETmetering, most of us have to work without the grid or with grid assist (mean we use the grid to charge the batteries or turn on the grid when batteries are low because we size the batteries small, 30% DOD nightly use). We try to aim the panel to the sun use winter months so we get the best angle on winter month when the sun hours are shorter, Summer we have longer hours so the lost on the aim will be compensated.

[J.P.M.]

In Jamaica, because we don't have NETmetering, most of us have to work without the grid or with grid assist (mean we use the grid to charge the batteries or turn on the grid when batteries are low because we size the batteries small, 30% DOD nightly use). We try to aim the panel to the sun use winter months so we get the best angle on winter month when the sun hours are shorter, Summer we have longer hours so the lost on the aim will be compensated.

Seasonal or more often adjustment is an option. If considered as part of the design from the beginning, it can be a viable way to increase annual or seasonal production in a perhaps cost effective way as evidenced by many megaWatt size arrays that use single axis tracking.

Some design considerations (among many) with most residential size arrays might include the safest support design that allows the greatest ease of adjustment for the cost involved, and deciding how often or the criteria for adjustment, along with what that adjustment would be in terms of tilt or azimuth or both.

[paulcheung]

Seasonal or more often adjustment is an option. If considered as part of the design from the beginning, it can be a viable way to increase annual or seasonal production in a perhaps cost effective way as evidenced by many megaWatt size arrays that use single axis tracking.

Some design considerations (among many) with most residential size arrays might include the safest support design that allows the greatest ease of adjustment for the cost involved, and deciding how often or the criteria for adjustment, along with what that adjustment would be in terms of tilt or azimuth or both.

A few years back, the panels are more expensive, lot of people in Jamaica use a single axis tracker with 12 to 16 200 or 250watts panels, they couple with 80 amps MPPT CC and a bank of 400 amps 48volts batteries, that will supply a house hold use about 250 to 350kwh a month.

[sensij]

Cost-effectiveness on a residential scale sounds challenging. The potential benefit of tilt adjustment is explored here, and many other places if you care to search. Although the results make some assumptions about the installation that may not be applicable in all cases, their data generally agrees with what PVWatts reports for sensitivity to fixed tilt angle.

The tl;dr summary is that relative to an array that has been set at the "optimum" fixed angle, adjusting it twice a year would give a 5.8% bump in energy generation, while adjusting it 4x per year gives a 6.5% bump. There is clearly a diminishing return on adjustment frequency, until you get into automated trackers that adjust tilt to follow the sun throughout the day, which can bump up output by over 20%.

Putting that in perspective, let's say you're considering a system to producing 10,000 kWh annually, and there is enough space to lay the panels out in a way that prevents self-shading. With an energy value of $0.20 / kWh, 6% more output is worth about $120 per year initially. Compared to the additional installation and maintenance costs, and the time required to adjust all those panels, that doesn't seem like much.

Even the value of just getting the panels at a fixed optimum tilt instead of the convenient flush roof mount tilt may not be compelling. For my installation the improvement that a 31° tilt would offer over an 18.5° tile (4:12 roof pitch) is just under 2%, or something less than $20 a year with the smaller size of my array. A non-flush tilt may offer some additional improvement over PVWatts estimation due to improved airflow and cooling, but that should be small relative to the improvement from the higher normal irradiance.

The math will be very different when comparing a net metering policy that rewards time-insensitive generation with something like an off-grid system that needs to keep the minimum generation as high as possible. My examples above assume unrestricted net metering.