Introduction

[dante]

Hi, I'm a math/software engineer, and am down the road somewhat in a 144 250 Watt panel design.

I'm looking for a formula to estimate a string's wattage output given the wattage of each panel, and with optimizers on each panel.

I'm trying to optimize an off-grid solution for the winter months. At my location, the azimuth varies 150 degrees at the winter solstice, but elevation only varies 29 degrees, and am seeking to provide continuous solar panel power as best as possible during the deep winter months, as balanced by a relatively large battery backup.

My particular setup accommodates 12 strings with 12 250-watt panels.

My intent is to write a program to perform a search algorithm that will determine a result that is reasonably close to optimal panel orientation. Note, I'm aware that "optimal" has many factors, such as weather, overcast days, rainy days, temperature, day of the year, varying power requirements over the year, etc., and am capable of writing a search algorithm. I do have the information on the panels. What I am seeking is how to model the output of the string given different wattages at different panels.

Having that, I will be able to create models for weather, and determine how to orient the panels in a string to meet energy requirements of the loads. Thanks for any insights!

 

[was-ja]

If you live in an area where you need to heat your house in the winter and not where you need air conditioning in the summer, the optimal orientation is the orientation that produces the most power on a sunny day on December 21st. In your case - almost vertically aligned.

If you can change the position of the panels throughout the day, that's a different story.

 

[dante]

If you live in an area where you need to heat your house in the winter and not where you need air conditioning in the summer, the optimal orientation is the orientation that produces the most power on a sunny day on December 21st. In your case - almost vertically aligned.

If you can change the position of the panels throughout the day, that's a different story.

Understood. The question is the optimal orientation for 144 panels, with 12 strings, year-round. In my estimation, it's not trivial, but the important question is how much difference in wattage per panel optimizers can provide, which will indicate the freedom to orient fixed panels of a string. Thanks for your note!

 

[OzSolar]

Welcome!

Have you already run different simulations using NREL's PV Watts? Other's can probably add more but I don't think there's anything you can do to squeeze out more than 2% or 3% of what it tells you. You can change tilt and azimuth to run it again.

Side note: I have an young engineer that works for me. He too is preparing to build an off grid-house. I often have to remind him that "Perfect is the enemy of good".

 

[dante]

Welcome!

Have you already run different simulations using NREL's PV Watts? Other's can probably add more but I don't think there's anything you can do to squeeze out more than 2% or 3% of what it tells you. You can change tilt and azimuth to run it again.

Side note: I have an young engineer that works for me. He too is preparing to build an off grid-house. I often have to remind him that "Perfect is the enemy of good".

"Perfect is the enemy of good". OMG, that is absolutely wonderful! I have learned that lesson very many times, and continue to learn it due to my nature. On the other hand, as a mathematician and software engineer, I know full well that missing even one bit of information can be catastrophic! I suppose we must seek the essential elements of a solution. Folks with deep experience will be able to guess better, but people like myself don't have that innate intuitive sense from many years of experience.

I have quite a bit of experience writing search algorithms, I suspect I will be able to write a program, with some assumptions in it, that can find a reasonable solution for panel orientation rather than trying to guess myself. I do note that these problems are incredibly challenging for utilities, and in Germany they have added billions of "Fast Start" fossil fuel systems in order to make up for the vicissitudes of solar power. Winter must be significantly overprovisioned for this reason. My intent is to make an engine that produces a probabilistic set of weather conditions as one input to the search algorithm. I suppose, ultimately, I'll discover if 144 250 watt panels will allow me to disconnect the utility completely.

Meanwhile, I realized I could contact a few optimizer companies, and adjust their numbers down a bit from their optimal to find the right orientations for all panels and all strings.

Thanks for that note! I love the insight, but unfortunately, it presupposes an intuition I do not have. Perhaps I'll post on how this goes, though the future is always uncertain.

 

[John Frum]

Just reading this thread is tweaking my OCD.

 

[John Frum]

Are solar trackers out of the question?

 

[dante]

Are solar trackers out of the question?

Well, yes. they are expensive, active, and complicated. I tend to prefer passive systems! Thanks for the thought!

 

[was-ja]

Another comment of mine. It also depends a lot on which system you will be using. In my case, I have a lot of trees around, and some panels are covered now, and some are still in the sun. If they work completely in parallel (as I have now), then it's still fine, but I have a lot of losses from the fact that the batteries are running at 12V. I am now moving to a higher voltage system (48V) for which I will need to assemble long chains of 8 panels, and just placing the elements inside the circuit in my case can be even much more difficult than the general alignment of the system relative to the sun.

 

[DougfromdaUP]

If you have a "relatively large" battery, wouldn't the optimum angle be that which produces the most output over the day? I get the impression that you want to equalize power production throughout the day. If you have a large battery, wouldn't the most power available, regardless of the time of day, serve you better?

 

[OzSolar]

"Perfect is the enemy of good". OMG, that is absolutely wonderful! I have learned that lesson very many times, and continue to learn it due to my nature.

How do you think I recognized it so quickly? It's sure helped me. Probably should get it tattooed on my forearm. "the first step is admitting you have a problem"

 

[dante]

If you have a "relatively large" battery, wouldn't the optimum angle be that which produces the most output over the day? I get the impression that you want to equalize power production throughout the day. If you have a large battery, wouldn't the most power available, regardless of the time of day, serve you better?

Not necessarily. Suppose the batteries are full, and the system overgenerates power. In that case, if the energy comes predominately around noon, and later in the day there is insufficient solar power for the load, the batteries are tapped earlier in the afternoon.

On the other hand, if the panels are oriented such that there is sufficient power around noon, AND energy later in the day as panels are oriented to also capture afternoon energy, the batteries are tapped later.

 

[dante]

How do you think I recognized it so quickly? It's sure helped me. Probably should get it tattooed on my forearm. "the first step is admitting you have a problem"

In my neck of the world, we use "KISS": Keep it simple, stupid! I continue to have to relearn that lesson. Some dogs simply cannot be taught!

 

[OzSolar]

In my neck of the world, we use "KISS": Keep it simple, stupid! I continue to have to relearn that lesson. Some dogs simply cannot be taught!

It has a few more names.

-The pareto rule

-Diminishing returns.

Side note: Edwards Deming is my hero.

 

[KITROBASKIN]

Sometimes the mind of a child has a solution...

If you link to a grid, does that allow more energy collection out of your panels? Batteries get full but the sun can still be shining. Perhaps you are really wanting to disconnect from your power company? If we had a grid connection, I would let the panels produce whenever possible, and get credit.

 

[curiouscarbon]

Welcome to forum?

Happy to see a new visitor with computer science background.

Best of luck with this build.

 

[400bird]

First, welcome!

Second, I think you're over complicating this.
If you have room for a full south facing array, it will make more power over the course of the day vs a south west facing array that will make more in the afternoon but at a significant cost of morning production.

What's your energy consumption? Desired days of autonomy (not starting the gen or reconnecting to the grid)?

36,000 watts is a lot for panels and a lot of juice.
750 amps of charging (at 48 volts) if you get the full array charging at full tilt.

If you need to tame down the 750 amps of charging, you could split your array into SW and SE. It would cost total daily production, but lower peak output by spreading it across more hours.