Why write this up?
I wanted to blog about this work for two reasons - first, because it's been a fun hobby project and someone might benefit from what I've done to implement it, but also because it's been interesting to see how much the availability of AI tools has changed the execution of not just the coding parts of this project, but also the research and implementation. I think it's an interesting example of where we are at this moment in time.
A note on AI/LLM usage
While I used AI and LLMs extensively in this project, this blog post is 100% written by me, except for a bit of copyediting.
Background
When we moved into our current house, it came with a small infinity pool - about 12' x 7' with a pump at one end so you can swim against the current. We don't use it for exercise very often - as my wife says, "swimming in place is even more boring than swimming laps," but the kids like to play in it. We live in San Francisco, where it's never very hot outside so we keep the pool warm, which is fairly expensive - the pool just has a resistive electrical heater, so it adds to our electric bill considerably.
It clearly loses a lot of heat through the top, since its only covering is a heavy-duty fabric cover, and so I've long wanted to do something to improve the insulation and reduce the heating cost. My first thought was just to put insulating foam on top of the cover using the foil-backed rigid foam panels used to insulate houses. In doing so I realized that I had a lot of surface area to use, and that on warm sunny days solar heating of the top cover of the pool would heat it above the thermostat setpoint, so I began to think about adding a thermal solar heater to the top of the insulating panels.
Goals
There were a few implicit goals at the start of the work:
- Reduce the cost of heating the pool.
- The entire system should be self-contained and solar powered - I didn't want electrical cables running to the pool.
- The system needed to be modular and easy to remove from the pool for swimming.
I wasn't explicitly concerned about breakeven cost - my interest was in part to have fun and learn some new things so if it took a while for the heating system to break even, I was OK with that. I also don't have a current way to determine the exact part of our electric bill attributable to the pool heating, so this wasn't easy to calculate.
Initial work - the insulating panels
My first step was to get some insulating panels. I cut them to size and put them on top of the pool cover. This achieved the insulating goal but it didn't look that good - the panels rapidly bent under their own weight. My first thought to stiffen them was to glue a plywood skin to the top of the panels but the weight of the plywood quickly convinced me it was too heavy to make the panels easy to move. I turned to ChatGPT:
What could I use as a lightweight stiffener for foam insulating panels to keep them from bending? I want to make an insulating pool cover from them. Each panel is 7'x4'. Must be commercially available or easy to fabricate
A good way to keep 7'×4' foam panels from "potato-chipping" is to add a light beam along the long edges (or a couple across the width), or to turn the panel into a simple sandwich panel (foam + thin skins). Here are practical, easy-to-source options: ...
What about gluing strips of plywood at right angles to the foam - that is, gluing a strip of plywood to the edge of the foam board, say like a half-inch thick plywood, four inches wide, so that the four-inch-wide direction sticks up at right angles to the foam thickness?
That'll work very well. What you're describing is basically a torsion rib / edge beam: a vertical web bonded to the foam that raises the section's moment of inertia a lot, so it resists bending much more than foam alone.
Why it works
A 4" tall plywood rib acts like the web of an I-beam. Even if the plywood is only 1/2" thick, making the rib tall is what buys you stiffness.
...
Practical design suggestions
Put ribs on the underside if possible so they're not in the sun (UV) and not tripping you up handling the panel.
Two ribs along the 7' edges often beats one center rib.
If you can, cut a shallow slot (dado) into the foam edge so the plywood rib is partially embedded (even 1/2" deep helps a lot).
If you want the same concept but more "pool-proof"
Instead of plywood, use:
PVC trim board (Azek-type) ripped into strips (won't rot), or FRP pultruded angle/flat bar, or Aluminum angle/channel.
Those do the same vertical-rib trick with far less worry about water.
Bottom line — yes, plywood ribs are a legit stiffening method and can be very stiff, but for a pool cover you'll want to either (a) seal the plywood like you mean it, or (b) swap plywood for PVC/FRP/aluminum to avoid swelling/rot.
I went with the plywood plan (I already had the plywood) and glued it to the edges after embedding fender washers into the foam to hold the plywood in place with 1/4-20 bolts. I then painted the white side of the panels black, drilled some holes through the foam to let rain drain, and placed them foil-side down on the pool. That completed my insulating panels.