Glycol Losses

Canuck_3

I need some assistance here Wallies. We're working with a general contractor who is in a legal broil. There is some bad blood between them and the owner of the buildings that they just built. The entire heating system for 5-6 buildings is in-slab radiant. We're not the installing contractor - we were asked to come in for some technical assistance. One of the issues is the fact that the system is losing glycol as can be seen from lower and lower levels in the reservoir tanks. The tanks are poly with lids and have been marked in pen on the outside of the tank, over time. Here's the kicker - losses are in the neighbourhood of 5-8 gal in the course of about three years. They've pressure tested the Rehau and have had some loops pass, then fail, then pass again. They've completed some thermal imaging, which was inconclusive. I'm saying that a loss of 1/2 to 1-1/2 oz/day/building is extremely small (certainly tough to find) and may be due to evaporation from the reservoir.

I appreciate that if there's a leak, well then, there's a leak...but what I'd like a better understanding of is what many of you may have experienced and what you see as fairly normal or abnormal, in the way of glycol loss in a system.

Thanks for the help!



Canuck

SpeyFitter

Wood frame?

You said 5-6 buildings, do they have one system tha feeds each building with a common mechanical room, or is it one system per building?

What is the percentage of glycol to water? I'm assuming its inhibited propylene glycol?

Did they pressure test with just water, or glycol? Sometimes with just water leaks won't appear as readily as they do with a higher percentage of glycol - glycol is a smaller molecule as I understsand it and will exploit areas water might not.

Are these buildings wood frame with a concrete subfloor that the radiant pipes are embedded in? Sometimes if a screw penetrates the wood subfloor into a radiant pipe embedded in concrete it can let out a very small amount of water.  

I'm assuming these tanks are some sort of chemical pot feeder, or are they an axiom type system that has a resorvoir and it pressure pumps the system and determines the static pressure of the system (along with the expansion tank). Are they open to the atomsphere or closed with a seal? Is there some way of isolating this tank from the system so you can see if your theory is correct?  

I'm assuming they're using all Rehau Pro-Balance manifolds with actuators? Have you investigated and seen/inspected each and every manifold? Around the actuators (remove the actuator physically), around the GPM meters, etc. - there all sorts of O-rings in these meters. Also inspect and make sure the supply/return valves are on tight on these manifolds as they use a fibergasket that can also leak if not completely tight.

Canuck_3

Not wood

Buildings are commercial construction with block walls. Glycol is inhibited propylene at about 30%. Reservoir is connected to an Albany, excess pressure (jockey) pump, which is tied into a pressure switch on the system. The tanks have lids, which are a loose fit. The glycol has a pink colourant added to it so that any external leak is readily visible. None have been noted. Pressure tests were done with air on isolated, glycol filled loops. All loops were tested and retested as results varied from one test to another with regard to leak/no leak. They may have had air leaks in their test apparatus. Evidently, all systems, (there is one for each building), were pressure tested, and passed, at time of, during and subsequent to the concrete pour.

Jean-David Beyer

I am a homeowner, not a contractor.

It seems to me that if the problem were due to evaporation, that only the water would evaporate because the boiling point of propylene glycol is around 370F, and water is around 212F. So if the concentration of the glycol in the water goes up, evaporation could be the cause, but if it remains the same, that leakage is more likely to be the cause. (Remember that ethylene glycol in car radiators not only lowers the freezing point, but raises the boiling point.)

SpeyFitter

Water versus Air

I have seen radiant piping leak that was pressure tested with air to about 60 PSI before they poured the slab (so that if the concrete guys damaged the radiant tube while pouring they would know but it wouldn't flood the place and compromise their concrete pour). It held for the air test but began to leak when filled with water down the line after the concrete was poured and they wanted to get temporary heat going. It happened as a result of a staple that was put through the pipe while laying on a wood subfloor (we used a wide crown Bostitch stapler with a modified safety - effectively a jig - it's fairly reliable for not hitting pipe but sometimes it happens if you aren't paying attention to your movements).

hot_rod

leaks when hot?

I have seen system leak a bit when they are warmed up, but not when the system is off or cold.



Crimp rings or connections are a common place to find leaks that only appear when the system is running. Look for a blueish stain around threaded joints to indicate glycol leaks.



Also the fill glycol should be in a sealed container to prevent air (O2) from depleting the inhibitor. You really should not be loosing any to evaporation.



hr

Canuck_3

The fish that got away

I think that's what this "leak" is going to be. I don't believe that the water evaporation is a viable argument, because we're not talking about the liquid boiling off. The alcohol in the glycol is far more volatile than water and would naturally evaporate faster.........and I realize that there isn't any evaporation in the system piping, perhaps only at the reservoir, which is open to atmosphere, except for a loose fitting lid. We've scoured the visible piping without seeing any signs of leakage. Penetration damage is also remote as the concrete is 8" thick with the tubing around center slab, and floor finish is either polished concrete or heavy ceramic overlay



What's frustrating is that system losses are so low, (1/2 to 1-1/2 oz/day, depending on which building, with total system capacity in the area of 600-700 gal/building). We can't see a leak with the naked eye, we can't see it with thermal imaging, Doppler testing is of no use and I'm not even confident of the consulting engineer's testing as some of the results flip-flopped. In the end, if we can't give a viable reason for the glycol loss, (an alternate reason than an U/G leak in the pex), the GC will need to tear out millions of dollars of floor.



 What I find unusual is that all of the buildings leak the same. I'd understand one building having an extremely slow leak, but all of them, and all in the same range?? Too much of a coincidence for my liking.



By the way HR - I agree about the inhibitor, I guess that's why they have a contract with Ashland to monitor the system chemicals.  

Jean-David Beyer

The alcohol in the glycol is far more volatile than water and

I do not believe there is any alcohol in propylene glycol solutions.

Mark Eatherton

The simplest of tests...

During the non heating season, isolate the expansion tank, shut off the glycol make up and watch your pressure. If it drops to zero there is a leak. If it remains stable, there is no leak.



Test done.



If the pressure drops, start doing isolated hydrostatic tests to fine tune the area where the actual alleged leak is occurring, and pump the pressure up and go back to the Doppler leak corelator test.



If it is in the slab, locate all lines using infra red technology, and then zero in on it with the Doppler system.



Before you go to that extent tho, check out all of the usual suspects, like automatic air vents with tail pieces run to the sanitary sewer, relief valve with unsupervised terminations etc.



I concur, that the chances of having EXACTLY the same amount of leakage in EVERY building, is slim to none.



Almost sounds like a monetary dispute between an owner and the GC, and the owner is trying to force some kind of deal in their favor.



If worse comes to worse, I would hire someone to "Negotiate" or "Arbitrate" a deal whereby the GC agrees to pay for the disappearing glycol, and that IF a leak does present itself, that he will accept total responsibility for the required repairs and any consequential damages (loss of business, rentable space, etc) if they occur.



Keep it out of court...



ME

rich pickering

system testing

Do you have a boiler inspector that tests the lwco and the relief?  2 gal a year sounds about right.

Canuck_3

C3H8O2

Propylene glycol, also called 1,2-propanediol or propane-1,2-diol, is an organic compound (a diol or double alcohol) with formula C3H8O2 or HO-CH2-CHOH-CH3. It is a colorless, nearly odorless, clear, viscous liquid with a faintly sweet taste, hygroscopic and miscible with water, acetone, and chloroform.

The compound is sometimes called α-propylene glycol to distinguish it from the isomer propane-1,3-diol HO-(CH2)3-OH, also called β-propylene glycol.

Canuck_3

You've hit the nail on the head

The issues are definitely monetary and court action has already begun. They're in 'exams' right now, where they ask questions and try to unravel each other's arguments. I had mentioned the static test to the GC on previous occasions but they're worried that the results may not be in their favour. You are absolutely correct though, that this would answer the question once and forever. I guess I need to convince them that this is the route to go. Who knows, it may also show that there is no leak in the system and then they're off the hook.



Thanks Mark.



Thanks to all!!