Combining water barrels for greenhouse thermal mass

Hello everybody! I stumbled upon this forum by chance when looking for greenhouse ideas, and I was overwhelmed by the creative technical designs I could find here.

I have an 8x12 greenhouse, where I am trying to add thermal mass. In an attempt to experiment and limit cost, I have sourced for very cheap four 55-gallon poly drums, which I filled with water, an old car radiator, a radiator fan, and an aquarium water pump. I strapped some pipes through a combination of adapters, and got the system running (few photos attached to this post).

The system works well: on my temperature sensors I see a nice little bump at night when I release heat stored in the water, and it reduces my need to provide supplemental heat (which I still need). However, while with a single barrel the system is easy to operate, I am having a very hard time using all 4 barrels at once. I tried using garden-hose 4x1 splitters to both retrieve water from the barrels and return it to them (barrels in parallel), but the predictable result is that each barrel ends up supplying and receiving water at a slightly different rate, which results sooner or later in having a barrel which overflows and one that gets progressively empty. I have thought of some solutions, but all of them seem to have their cons:

1) Leave the barrels connected in parallel as they are now, but only operate the pump for a few minutes a time. The water level will become uneven, but when the pump shuts off, it can even out via the siphon created by the outlets and their common connection at the 4x1 splitter. That would require experimentation on how long to pause, and how often, and what if my timer doesn't work well enough? Or what if the time it takes for the water to even out greatly exceeds the time it takes to become uneven? I'd operate my pump a fraction of the potential time, making the whole system not very useful.

2) Daisy-chain the barrels, i.e., instead of connecting them in parallel as I have them now, connect them in series. Water enters in barrel 1, moves to barrels 2, then 3, then 4, where it is pumped out to the radiator. But how can I let water move across barrels without either being too slow for the rate at which the pump fills #1 and empties #4? In other words, it doesn't seem to me that pure siphon effect between barrels 1->2, 2->3, and 3->4 can keep up with the rate at which water is being pumped into #1 and out of #4... Even if I were to connect the barrels with 4 large hoses at the bottom, how can I avoid that only the water at the bottom of the barrels move through, while most water just rests at the top? Or is that even a problem?

3) Use solenoid valves and water sensors to monitor the barrel levels. Talk about over-engineering...

Would anybody have any recommendation? Any comment of the solutions I could think of, or other solutions that I am missing altogether? I can provide additional clarifying information if needed. Thank you all very much in advance, this community looks amazing!

-Victor

You're on the right track with your general plan. If you link the barrels together at the bottoms then they will act as a single reservoir:

 ----------          ----------          ----------          ----------
|          |        |          |        |          |        |          |
|~~~~~~~~~~|        |~~~~~~~~~~|        |~~~~~~~~~~|        |~~~~~~~~~~|  water level
|          |        |          |        |          |        |          |
|          |        |          |        |          |        |          |
|          |        |          |        |          |        |          |
|          |        |          |        |          |        |          |
|          |        |          |        |          |        |          |
 ----------          ----------          ----------          ----------
     |                   |                    |                   |
     +===================+====================+===================+=========(/) tap

Then you can pump from the header line back into the tops of the barrels to equalise the temperatures. It won't matter if the split on the return is inexact, because the system is self-leveling.

Paint your barrels a dark color. They'll heat up faster.

Very cool build!
I was going to suggest float valves to prevent over fill, but
I think Phil is spot on in his advice.

You might want to add a solar thermal water heating coil in line with your air to water heat exchanger.
An off the shelf solar pool heater could do the trick.
Running black hose back and forth on the wall above the barrels might work.

Some insulation on the north wall and roof of the greenhouse,with reflective facing could really help retain heat.

Jim Fry wrote:Paint your barrels a dark color. They'll heat up faster.

Thank you, will do!

Phil Stevens wrote:You're on the right track with your general plan. If you link the barrels together at the bottoms then they will act as a single reservoir:

 ----------          ----------          ----------          ----------
|          |        |          |        |          |        |          |
|~~~~~~~~~~|        |~~~~~~~~~~|        |~~~~~~~~~~|        |~~~~~~~~~~|  water level
|          |        |          |        |          |        |          |
|          |        |          |        |          |        |          |
|          |        |          |        |          |        |          |
|          |        |          |        |          |        |          |
|          |        |          |        |          |        |          |
 ----------          ----------          ----------          ----------
     |                   |                    |                   |
     +===================+====================+===================+=========(/) tap

Then you can pump from the header line back into the tops of the barrels to equalise the temperatures. It won't matter if the split on the return is inexact, because the system is self-leveling.

Thank you all for the advice. Even though my system seems to be set up in a similar way as Phil described, it does not seem to self-level. I attached a simple diagram of my connections. With that system I definitely see the water becoming uneven, which eventually results in at least one barrel overflowing (often #2 or #3). The way I see it it's either because of some issue with the returns or the supplies, or both.

As far as the returns to the top of the barrels are concerned, I can visually see that the water flow is not even. Some hoses spit out way more water than others. I believe it's due to the pipe lengths, with water preferring shorter pipe runs over the longer ones (I am an electrical engineer, I imagine it to work like electrical resistance. Maybe I'm wrong). With the little valves on the 4x1 splitter I can adjust the flow and that alleviates the situation but it's basically impossible to make it even. I can see how it wouldn't matter if my supplies from the bottom of the barrels would even out the system anyway.

I cannot visually inspect the flow level of the supplies from the bottom of the barrels, and I assume water is also being pulled unevenly like it happens at the top. (or maybe not?). But is it possible that the self-leveling, being driven by gravity/siphon just can't keep up with the flow of the pump? I can see how the system would be self-levelling eventually, but I could also see that there is a limit to the speed at which the self-leveling occurs, and if the pump exceeds that rate, it would become a problem. Or is there something else at play, like my way of making the 4x1 connections?

Should I downsize the pump? Or should I try something else? Thanks!

How large is the tubing at the bottom of the barrels? If there's high rate of flow going into one or two of them, it might be overwhelming the system's ability to self level. The bottleneck could be that 4x1 manifold splitter...swapping that out for larger diameter hose and tee fittings will help.

If you think about it like an electrical circuit, we want low impedance on the manifold supplying the pump and higher impedance on the return from the pump to the tops of the barrels.

Seems like a float valve in each barrel might actually help.

Phil Stevens wrote:How large is the tubing at the bottom of the barrels? If there's high rate of flow going into one or two of them, it might be overwhelming the system's ability to self level. The bottleneck could be that 4x1 manifold splitter...swapping that out for larger diameter hose and tee fittings will help.

If you think about it like an electrical circuit, we want low impedance on the manifold supplying the pump and higher impedance on the return from the pump to the tops of the barrels.

Everything is standard US garden hoses. I'll look into improving there. Thanks for the suggestion!

William Bronson wrote: Seems like a float valve in each barrel might actually help.

Thanks for the suggestion. I have not used them before. How are you recommending I use them? Are there specific parts you would recommend?

William Bronson wrote: Seems like a float valve in each barrel might actually help.

Oh, something like these? That would be so easy!

Yep, something like those.
They work just like the float valve in a toilet tank.
If each barrel has a such a valve, I think the system will be forced to balance out.

Another thing that should work is connecting the barrels in series with the top of each barrel connecting to the bottom of the next.
With a single path of water flow, there should be no imbalance between the barrels.



Looking at your set up has me thinking about doing something similar with water heater tanks.
Mine would be wood fired, but otherwise very similar.

It looks like if you can connect the top inlets of the barrels into a single header pipe (pvc?) that is connected to the barrels in a water tight manner, and then put the pump output into that header, it would even out the flow.  The water level would end up being in that header, not down in the barrels.

It might be that you're pushing more water to one hose or another, or that you're sucking more water from one barrel or another.  Or likely a combination.

Another option might be to replace the barrels with an ICB tote.  That will hold 6 barrels worth of water and be easier to manage.  

Update to my post with the final build, in case people are interested. I ended up ordering 4 floater valves, and they work great. Got a bunch of tubing going and cleaned it all up. Pictures are attached. Next, I'll try to collect some temperature data to see how the system behaves. Thanks all for the advice!