EcoRenovator (
-   Solar Heating (
-   -   Designing & building a solar hot water tank (

Daox 06-16-11 02:28 PM

Designing & building a solar hot water tank
1 Attachment(s)
With the solar panel racks designed to my liking, I am moving on to designing the tank for the solar hot water setup. I think I'm actually going to build the water tank first as I can use it as a temperting tank for my domestic hot water until I get the solar panels installed and working.

The tank will be made of wood, and hold around 400 (updated to 650 now, see later posts) gallons of water. It will be built similar to Gary's DIY solar hot water tank. The main differences are that mine is going to be smaller, it lined with a different material, and insulated mostly with cellulose instead of foam board. These changes are mostly to decrease cost. The smaller tank size matches my 200 sqft of collector with a rule of thumb that you need 2 gallons per sqft of collector. I'll go through the heat gain calculations in another post to see exactly why more than 400 gallons of water is not needed and might actually be bad for the system depending on how you intend to use it.

Anyway, I got it somewhat modeled up over lunch today. As you can see, its very similar to Gary's tank. I will be doing some FEA to see if the plywood thickness can be reduced. This could save a decent amount of money.

Daox 06-20-11 12:56 PM

4 Attachment(s)
I ran a few simulations over lunch today. The weight of just over 400 gallons of water is roughly 3,600 lbs. That is more weight than most cars, so its nothing to play around with.

The one thing I struggled with was getting the water weight forces to come out exactly how they would in the real world. The bottom of the tank is obviously going to see the full 3,600 lb force of the water, but what about the sides? At the bottom they'll see the most pressure, but at the top it will be less. So, I'm still searching for a formula that will help me figure that out. For now, I used 1 psi as my baseline. The base of the tank is 3,400 sq/in, so roughly 1 psi. I used this for the sides as well. Here is what I got:

This is with 15/32" thick plywood. Try to ignore the colors as they don't truely tell you the whole story. The graph on the right shows the peak stress and the scaling.

This is with 19/32" thick plywood. Again make sure to reference the graph as the scale varies.

Now with 23/32".

And once again with 3/4" thick plywood.

As you can see, the 3/4" plywood showed quite a bit more strength than the other two thicknesses. I'm not quite sure why the huge reduction in stress. But, its not really consequential. Plywood has a tensile strength of 4500 psi. So, even the 15/32" can hold up to the stress of the water (assuming my simulated pressure isn't way off). I'll guess I'll run a few more simulations with even thinner plywood. However, there is a point of diminishing returns I'm sure. The 19/32 4x8 sheet only costs $3 more than the 15/32 sheet, so there isn't huge cost savings to be had.

Daox 06-20-11 01:19 PM

So, I talked with an engineer at work and he cleared up the numbers I should be using and the forces involved. I was using 1psi where I should have been using 1.3 psi. So, the stresses should go up about 30%. I'll rerun the simulations with the new numbers to make sure this is the case. I'll also incorperate the correct loading on the sides of the tank.

RobertSmalls 06-20-11 09:15 PM

Interesting. I just figured tanks were something you had to buy, but it looks like Gary has pulled it off.

I would caution that wood, ceramics, and other materials with brittle failure modes don't always behave the way the FEA models suggest. Minor cracks and porosities that would have little effect on a ductile material will make the strength of brittle materials more unpredictable. Stress concentrations, such as at fasteners, will be significant.

I wonder... could you use steel hoops (or fiberglass) for their tensile strength, to allow you to reduce the amount of plywood required?

My intuition tells me a square tank would be better than rectangular, in terms of both cost and heat retention.

Btw, what software are you using?

Daox 06-21-11 07:11 AM

Yep, Gary has a few different ideas on DIY tank designs on his site thankfully. However, I believe his is the cheapest and easiest to build and insulate. A 120 gallon tank retail for $1000+ and I'd need a few of them! My current cost estimate for the tank and liner (no plumbing) is about $250 in materials.

I agree, a square tank would be more thermally efficient. The currently modeled tank can hold about 650 gallons of water. Since I only need 400 I will be shortening it up a bit length wise. However, I still have a rectangular footprint to put the tank in, so that will also be determining the shape of it to some extent.

The tank (and the rack) is modeled in SolidWorks and analysed with CosmosWorks.

Daox 06-21-11 12:08 PM

I redid my calculations on the tank sizing and decided that going with a larger tank will actually be worth it. The end result is the lower temperature increases collector output enough to be worth the reduced tank temperature.

Here is a quote from another post where I explained the calculation:


Thats all quite true AC Hacker. It made me rethink things (as well as S-F's comments). So, I ran my own numbers on using a 400 gallon vs 650 gallon tank. There are a few effects to consider.

First off, this change lowers the tank temperature. The reduced tank temperature reduces the heat loss of the tank due to the water being cooler. However, the tank is also physically larger, so it has more surface area to loose heat from. In the end, the larger tank does loose more heat, but the effect is very minimal. All in all, the 650 gallon tank looses ~300 btu per day more than the 400 gallon tank. This is nothing when you consider even in December (worst month for solar), my average solar heat input is ~79,000 btu.

Second, the tank change also increases the collector efficiency. This is because cycling hot water through a collector makes it picks up less energy. The panel also stays cooler and looses less heat to the environment. This is where most of the benefit comes from.

Lastly, going to a larger tank supplys you with cooler water. In my case, going from 400 to 650 gallons will likely lower the average tank temperature about 10 degrees (from 115 to 105F), and lower the peak temperature from 140F to 120F. 105F is just barely warm enough to shower, so some additional energy may be needed for DHW. However, 105F is fine for hydronic heating. Considering that space heating is by far the larger load and energy hog, I should primarily design to its needs. In addition for 6 months of the year I'll be collecting more heat than will be used, so hot water won't be an issue then.

All in all, going from a 400 to 650 gallon tank will allow me to collect about 6.7% more heat from the panels. I just need to be aware that I'll be dealing wih a lower temperature water and design accordinly for it.

Daox 06-21-11 12:28 PM

1 Attachment(s)
So, I just finished up running the next simulation with the 650 gallon tank and 15/32 plywood and things look pretty good. A max stress of just under 3000 psi puts us in pretty good shape. I didn't bother correcting the load distribution on the sidewalls. If its good now, it'll only be better with less loading on it.

Here is the stress distribution. I'll probably add some corner supports like Gary did with his, especially since I'll be using thinner plywood. Without something solid to screw into, it'll be hard to make a strong joint.

strider3700 06-21-11 01:38 PM

I'm curious what are the dimensions of the tank you're thinking about?

Daox 06-21-11 02:32 PM

The base of the tank is a 4x8 sheet of plywood. There is 2x4 framing around it and the side sheets are 48" tall. The dimensions of the inside of the tank are 40"x88"x48". If filled to the brim, it'll hold about 730 gallons. With the fill level 6" below the top, you get 640 gallons.

AC_Hacker 06-21-11 02:58 PM


Originally Posted by Daox (Post 14182)
...Here is the stress distribution...

The stress distribution looks the way I would expect it to look if the box had a lid and was filled with a compressed gas...

I would expect that an open box filled with water would show some indication of greater stress toward the bottom of the sides and end, and a reduction in stress to zero toward the top of the box.


All times are GMT -5. The time now is 02:14 AM.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, vBulletin Solutions Inc.
Ad Management by RedTyger