Can I mix panels installed at different angles in parallel?

[willstay]

Can I mix two panels installed in two different angles in parallel to a single MPPT charge controller?

Let me explain why I want to do this. Let's say the battery can be charged at maximum current produced by two 100 watt panels connected in parallel. In Florida, sun is at 86 degrees in summer; 62 degrees in spring/fall; and 38 degrees in winter (taken from http://solarelectricityhandbook.com/...alculator.html). Ideally, best angle for panels would be 28 degrees (90 - 62). This setup would produce best output in spring/fall. Flaw with this setup is I have to resort to less output during summer and winter.

Now I want to install two 150 watt panels but panel A in 16 degrees and panel B in 40 degrees (these are in-between angles between winter-spring and summer-spring respectively for Florida sun). Panels A and B will be connected in parallel to one single MPPT charge controller. During summer panel A will produce more electricity; during winter panel B will produce more electricity; and in spring/fall both panels will produce slightly less electricity. The idea is to have more consistent output throughout the year. Neither panels will produce max output at the same time of the year hence higher wattage panels are to be selected.

[sensij]

In parallel, as long as the voltage is close, it will work well. If panels A and B have similar specs, the biggest difference in voltage between them will come from the difference in temperature between them. That shouldn't be much, although as shown, B might run a bit cooler with better airflow, except in winter when it is getting more direct sun and may run hotter.

Some people chose to do something similar to level out intra-day production by splitting the array between east and west.

You probably know this already, but if you really want to level out seasonal energy production, you would put both panels at their winter tilt. The penalty in peak power that you would take in summer is compensated by the longer days. What you've proposed is sort of a compromise between maximizing energy production and leveling it.

You can check out PVWatts for more detailed modeling of the different orientations you are considering... check out the hourly output option.

[willstay]

Thanks sensij for the PVWatts link, I am playing around and charting data in excel with various angles. Will check with 16 and 40 degrees and merge data.

[willstay]

My theory of installing panels A and B in two different angle does not work. I used PVWatt calculator and assumed panel of capacity 500 watts. Reduced all the default inefficiencies, increased efficiency of inverter to max it supported. Panel A tilted at 16 degrees in blue line in the graph below. Panel B tilted at 40 degrees in orange. They produce different outputs as expected meeting at two points during spring and fall. Output from A and B total is in gray line. What is shocking is total output (orange line) when both panels are installed in the same angle of 28 degrees.

[J.P.M.]

My theory of installing panels A and B in two different angle does not work. I used PVWatt calculator and assumed panel of capacity 500 watts. Reduced all the default inefficiencies, increased efficiency of inverter to max it supported. Panel A tilted at 16 degrees in blue line in the graph below. Panel B tilted at 40 degrees in orange. They produce different outputs as expected meeting at two points during spring and fall. Output from A and B total is in gray line. What is shocking is total output (orange line) when both panels are installed in the same angle of 28 degrees.

When I first read your post I thought about suggesting that, while your idea would work, it probably won't work as well as thought. You beat my response.

The common wisdom was, before computer modeling became ubiquitous, and pretty much still is, that the best, first cut, all round annual output is at an az. of due south and a tilt equal to the local latitude. As modeling improved, it became possible to tweak those orientations and optimize based on local conditions (early morning clouds, poor winter irradiance, temp. extremes, etc). For example, my optimum orientation w/ Sunpower panels is about 187 deg. az, and 29 deg. tilt or so. A single tilt also makes for easier construction (read less $$'s) and perhaps improved aesthetics.

I suspect the temp. diff. due to the difference in tilts and wind exposure and thus panel/string) voltages would have been more pronounced at the solstices and least at the equinoxes.

[bcroe]

My theory of installing panels A and B in two different angle does not work. I used PVWatt calculator and assumed panel of capacity 500 watts. Reduced all the default inefficiencies, increased efficiency of inverter to max it supported. Panel A tilted at 16 degrees in blue line in the graph below. Panel B tilted at 40 degrees in orange. They produce different outputs as expected meeting at two points during spring and fall. Output from A and B total is in gray line. What is shocking is total output (orange line) when both panels are installed in the same angle of 28 degrees.

I would not back away from your theory so quickly. The problem with PVwatt, is it throws in every possible variable. The big
clinker is THE WEATHER, which puts in a variable much more significant than the effect your are trying to optimize. Since
you can't turn off this wild card, I concluded this simulation is USELESS for fine tuning optimization.
Bruce (uses 3 different orientations) Roe

[J.P.M.]

I would not back away from your theory so quickly. The problem with PVwatt, is it throws in every possible variable. The big
clinker is THE WEATHER, which puts in a variable much more significant than the effect your are trying to optimize. Since
you can't turn off this wild card, I concluded this simulation is USELESS for fine tuning optimization.
Bruce (uses 3 different orientations) Roe

Bruce: I think I respectfully disagree with part of you statement about PVWatts.

As the PVWatts help info screens say on the getting started page under "Important Note": "PVWatts is suitable for preliminary (underline added) studies .......The production estimates that PVWatts calculates do not (underline added) account for many factors that are important in the design of photovoltaic systems..............". That would seem to run counter to your statement that PVWatts "throws in every possible variable".

As one example of any number of variables not accounted for in PVWatts, and perhaps of some importance to you in particular: Terrestrial albedo, or other possible reflections from surroundings, including fresh snow are not considered. Many of the more important variables that make up the system loss parameter are adjustable in the model.

Accounted for in the model or not, a perhaps reasonable 1st approximation of a way to match PVWatts estimated output to actual might be to compare actual "clear" day output to the model estimated output on clear days for days when the temp./wind profile is similar. Doing so for several days at different times of the year may allow the system loss parameter to be adjusted until the model estimate and actual output are similar. That's sort of after the fact, and a bastardization of model use, but it might be useful for some things including model verification or system deterioration estimates to name two.

PVWatts also uses weather in a big way, which, unless I misunderstand your meaning, you seem to be saying is not adequately or correctly accounted for in PVWatts, at least to your way of thinking. PVWatts uses the same weather data bases as most other models. Those data are for the most part, but not entirely, composed of synthetically generated data. However, over time, they have shown to provide reasonable approximations that seem useful to estimate long term, average output.

While I believe there are perhaps better models available, I've found that, with some experience based massaging, PVWatts to be fit for purposecan produce what appear to be reasonably acceptable estimates of long term average that seem to more or less agree with other, more rigorous models such as SAM or TRNSYS.

In spite of what some, perhaps most folks think about PVWatts, it, like any model, is a guess, albeit educated, of system performance based on a number of variables using methods and approximations that are understood and used by folks who know much more about solar and solar modeling than I do.

Respectfully,

J.P.M.

[bcroe]

Bruce: I think I respectfully disagree with part of you statement about PVWatts.

As the PVWatts help info screens say on the getting started page under "Important Note": "PVWatts is suitable for preliminary (underline added) studies .......The production estimates that PVWatts calculates do not (underline added) account for many factors that are important in the design of photovoltaic systems..............". That would seem to run counter to your statement that PVWatts "throws in every possible variable".

As one example of any number of variables not accounted for in PVWatts, and perhaps of some importance to you in particular: Terrestrial albedo, or other possible reflections from surroundings, including fresh snow are not considered. Many of the more important variables that make up the system loss parameter are adjustable in the model.

Accounted for in the model or not, a perhaps reasonable 1st approximation of a way to match PVWatts estimated output to actual might be to compare actual "clear" day output to the model estimated output on clear days for days when the temp./wind profile is similar. Doing so for several days at different times of the year may allow the system loss parameter to be adjusted until the model estimate and actual output are similar. That's sort of after the fact, and a bastardization of model use, but it might be useful for some things including model verification or system deterioration estimates to name two.

PVWatts also uses weather in a big way, which, unless I misunderstand your meaning, you seem to be saying is not adequately or correctly accounted for in PVWatts, at least to your way of thinking. PVWatts uses the same weather data bases as most other models. Those data are for the most part, but not entirely, composed of synthetically generated data. However, over time, they have shown to provide reasonable approximations that seem useful to estimate long term, average output.

While I believe there are perhaps better models available, I've found that, with some experience based massaging, PVWatts to be fit for purposecan produce what appear to be reasonably acceptable estimates of long term average that seem to more or less agree with other, more rigorous models such as SAM or TRNSYS.

In spite of what some, perhaps most folks think about PVWatts, it, like any model, is a guess, albeit educated, of system performance based on a number of variables using methods and approximations that are understood and used by folks who know much more about solar and solar modeling than I do.

Respectfully, J.P.M.

I'm not today arguing the applications of PVWatts you mentioned. It isn't LITERALLY true that it considers
"every possible variable", but it hits most of the ones that matter for its claimed accuracy (what is the phase
of Jupiters 3rd moon?). I'm not changing my opinion (based on my limited experience) that its not the right
tool for the problem presented by the OP.

When running a simulation, I expect the exact same inputs to yield the exact same result. And a sufficiently
small change in inputs should produce a minor output change, at least for this type of problem. This I don't
get from PVWatts. Bruce

[J.P.M.]

I'm not today arguing the applications of PVWatts you mentioned. It isn't LITERALLY true that it considers
"every possible variable", but it hits most of the ones that matter for its claimed accuracy (what is the phase
of Jupiters 3rd moon?). I'm not changing my opinion (based on my limited experience) that its not the right
tool for the problem presented by the OP.

When running a simulation, I expect the exact same inputs to yield the exact same result. And a sufficiently
small change in inputs should produce a minor output change, at least for this type of problem. This I don't
get from PVWatts. Bruce

Understood. While respecting your opinion, my experience with simulations, including and particularly PVWatts, is that the same inputs will yield the same results every time. If not, something is amiss. Minor changes in inputs will likely change results, sometimes quite a bit.

While I believe what you write, my guess is you are changing some inputs, perhaps without knowing it. In any case, you are certainly entitled to your opinion.

The last word is yours.

J.P.M.