Thursday, June 18, 2009

Another insulated pool

Back when I posted about the insulated in-ground pool that I'm building, I asked if anyone else is building such a pool. I've received a few answers:
  • One reader in Melbourne is building such a pool.
  • Several have been built in the United States, but only one of the ones I've heard of is residential. The rest are all commercial facilities.
  • Insulated pools are standard when the pool sits on top of a parking structure. Apparently installations like these are simply impossible to heat if the pool is not insulated, and there are structural isolation benefits as well.
Up until now, though, no pictures! Thankfully, the reader from Melbourne has recently written in to share a few pictures of his insulated pool. Here's the standard picture of the dig:

The pool is 46 feet long, which is exactly the same length as mine. His is skinnier (10 feet wide) and more shallow (max 6 feet), which is appropriate for a lap pool. Below, it looks like they are installing an in-floor cleaning system. Very nice.

Below is a pic of the insulation going in. He is using Dow Highload 100, sold there as Dow HD300, in the same thickness that I used (2 inch). He says:
The insulation I'm using is Dow HD300 in 50mm boards. This product is made for insulating under coolroom floors with trucks driving on top, and is overkill given its compressive strength specs of 2% compression (1mm) after 20 years of 250 kPa or around 25 tons per sq meter. However, the pool contractor and engineers had never seen pool insulation done before and through an abundance of caution over-specified for the highest compressive strength product they could find to be sure it wasn't going to settle. With the loads from this pool of only around 2 tons per square meter, we have more than an order of magnitude margin of safety. In the end, the cost differential between this and lesser rated products was so small that in the interests of getting the pool contractor comfortable with signing off we went with the HD300.

The contractors didn't glue the boards to the soil with foam, instead they used the rather unsubtle method of nailing it through with steel rod. I had two concerns about this:
  • This will mean there's some heat conduction losses through the steel rod from the soil to the concrete, though the total surface area of steel in contract with the cement shell would still be minimal so this probably isn't a big deal.
  • A risk of the rod eventually rusting and applying pressure to the concrete shell, but the foam will (I hope) compress enough to accommodate any rust expansion and prevent concrete spalling off the shell were this ever an issue.
The upside is at least I don't have to worry about the compression issues for the expanded foam glue you'd used and hence avoids the risk you mention in your blog that this may place extra strain on the shell as it settled, and from the photos it seems the contractors have got a good solid base without the rocking problems you'd mentioned.

The steel rod seems like a good idea. I tried to find an equivalent product here and failed, which is why I ended up with the polyurethane foam. One other contractor I've talked with in the U.S. also used foam, but I neglected to ask him if he chose not to use steel nails for some reason.

Here in California we use Dobies to seperate the rebar from the ground/insulation. Dobies are simple 3" x 3" x 3" concrete cubes with a wire in them. Check out the much snazzier looking rebar spacers they use in Australia. The wall does not appear to have a bond beam at the top, but instead is pretty thick the whole way up.

Insulating the piping has been a major effort on my project. It's not clear in these pictures if this pool's piping is insulated.

Gunite going in:

His pool is in basically the same condition as mine right now. Note the clever combination of bench seat and stairs at the right hand side of the pool. Very nice. The pool looks deeper than it is because the lot slopes up to the left, and the left hand side of the pool is a retaining wall (raised bond beam).

It's a nice looking project, and I'm very curious to see how it turns out. Thanks a lot, Melbourne!

Sunday, June 14, 2009

A professional look at The Day After

Here is a set of essays on the calculus of nuclear war, written by someone who used to plan nuclear war.  They are short, funny in places, reassuring in places, and generally scary.

Of course, no mention of nuclear weapons is complete without directing readers to the Nuclear Weapons Archive, by Carey Sublette.  I remember first reading the FAQ in 1996 or so, and being astounded.  It changed the way I thought about The Bomb.

It's the physics bit that got me.  I had previously though of fusion bombs as being somewhat like the Sun, only, here.  But it turns out that fusion in the Sun proceeds along quite slowly, at comparatively low temperatures and pressures.  Fusion bombs operate at much higher pressures and temperatures than stars do, and (obviously) on much shorter timescales.  It turns out to be almost completely different physics.

For some reason that really bothers me.  The notion that we use physics that can't even be observed anywhere in the natural world seems odd.  Perhaps I'm succumbing to nuclear hocus pocus, since I can't think of anywhere in the natural world that we can observe hydrocarbon-oxygen combustion at dozens of atmospheres of pressure, and yet our cars and airplanes do that all the time.

Sunday, June 07, 2009

Reynolds number

It looks like one of the problems with the fountain is that I'm pushing slightly too much water through the flow straightener.

At very low velocities, flow through a pipe is laminar.  I wanted laminar flow in the flow straightener because laminar flow has no turbulence which can then break up the output jet.  It turns out that the flow velocity in the pipe has to be incredibly slow, and it turns out that I managed to design my fountain to be right in the transition region between turbulent and laminar flow.

Here is the Engineering Toolbox link on Reynold's numbers.

At full flow, I'm pushing about 180 gallons/minute through 16 of those flow straighteners.  Each has an internal diameter of 15.3 cm, so that the flow rate is 3.85 cm/sec.  Plug that into the handy calculator (the one using kinematic viscosity) and you get a Reynold's number of 5213.  That's turbulent flow.

At the flow tested in January (which worked properly), I was going up about 33 inches instead of 65 inches, so my jet velocity was 71% of full flow now.  Also, the cross section of the jets was .41 inches instead of 0.5 inches as it is now, so that the velocity inside the flow straightener was 48% of what it is now.  Plug 1.84 cm/s into that Reynold's number calculator and I get... 2491.  That's transient flow, but quite close to the 2300 needed for laminar flow.

If this is really the only problem with the fountain, then I ought to be able to slow down the flow enough to get the Reynold's number down to something around 2300, and see laminar flow at the output.  How slow?  To get half as much flow, the jet velocity is halved, and the arc height goes to 1/4 of what it is now, or 16.5 inches.  In fact, at that velocity, I do indeed get laminar flow:

Note that the impact here is on the first step into the hot tub, which is a little lower than the nominal water surface, and the arc is about 20 inches above the nozzle rather than 16.

The jet is well behaved until it gets to the top of the arc, where the bottom of the jet interferes with the top of the jet, and the result is that is spreads out laterally. That lateral spread then turns into an oscillation in the flow until it hits the step.

Anya demonstrates that the jet is 18 inches above the bond beam, or about 20 inches above the nozzle.

At this point the default setting for the fountain is to throttle back to 40 inches throw height, which clears the occupants of the hot tub and isn't too noisy.

If we wanted to get the tall jets to behave properly, it appears we'd need to cut the flow rate approximately in half, which means we'd have to reduce the jet diameter to 0.350 inches instead of 0.500 as it is now (so the finished hole diameter would be 0.440 inches).  That means I'd have to pull the stainless steel nozzles (recall they are epoxied into the PVC heads right now), get new nozzle made (probably $300), and epoxy them back in.  That all sounds possible, and certainly cheap enough, and probably can be done fast enough given that it's going to take 5 weeks to get the tile delivered.

However, there's a good chance I'd just destroy the PVC heads in the process, and there is also a good chance I'd get the nozzles glued back in crooked.  I don't think we're going to try.

We're getting more comfortable with how it looks.

Thursday, June 04, 2009

Camera Guy at work

One of the nice things about working here is that when we need stuff, we get it.