Mystery Occasional Flood from Hot Water Tank

[DC70]

Hi all,

I'm intrigued by an odd leak and wondered if anyone can shed light on this...

Twice, we have gone away from home for some weeks and so shut off the water supply at the stopcock and shut off the immersion heaters to our hot water tank in our apartment. Around three days after we leave, a substantial flood is reported by our neighbours coming from our flat.

On returning home, it is clear that a large flood has taken place, with evidence of flooding several metres from the hot water tank and into several rooms.

Turning a tap on proves that the stopcock is working as no water comes out.

Turning the water stopcock back on reveals no faults and no further flooding occurs. Other than the evidence of the recent flood, nothing seems to be wrong.

This happens in an apartment with neighbours on one side and above. There is no evidence that the water came from them - i.e no ceiling damage, no reported leaks, no water from the side the neighbour is on etc.

The only slight clue is that the jacket to the hot water tank is wet from the top down. So it could be that the temperature and pressure relief valve is letting go in spectacular fashion. But why would it if both the water and the heaters are off - where could excess pressure come from?

This has happened twice, albeit 5 years apart. I'm just a bit baffled what's going on!

Any clues would be greatly appreciated.

DC

 

[deg]

You mention a jacket being on the cylinder but also a tpr valve. Do you have an unvented cylinder?

 

[ShaunCorbs]

Any chance of a few pics please?

 

[DC70]

You mention a jacket being on the cylinder but also a tpr valve. Do you have an unvented cylinder?

Thanks for taking the time to answer.

I assume it is an unvented hot water cylinder as it is fed directly from the cold mains water supply. There is no other water tank. Hot water stops pretty immediately after shutting off the main stopcock. There is what I assume to be an expansion vessel on top of the main cylinder.

DC

 

[Stoney Ground]

Could be a leak that opens up when cold but seals itself when hot.

 

[Best]

Need a few photos. I know that corroded copper cylinders can leak when people turn their heating off when they go on holidays, but seal again when the cylinder is later heated.

 

[justlead1]

Unless there is an obvious visual clue, dampness, corrosion etc. It can be difficult spotting the problem. As you live in the property i would play the long game. Try using say talcum powder, dust a bit around any areas that are suspect.
And check on a daily basis, looking for water tracks. That may dry out with the heat under working conditions. If evidence is found it will be a good starting point for rectifying problems. Good Luck

 

[DC70]

Thanks very much for the replies.

The issue doesn't seem to be directly temperature related, because when we returned home after a couple of weeks, the system was stone cold. We then turned the water back on and there were no leaks. We then turned the immersion heaters on and no leaks were formed through the heating up process. In the last week while the system has been up to temperature, there have been no leaks.

The flood seemed to be quite substantial, with water travelling several metres across carpets and into multiple rooms in all directions. It feels like it must have been a catastrophic leak rather than a drip-drip leak. The flood was reported by neighbours when a visible pool of water reached a metre outside our front door into the corridor.

The last time this happened, a plumber replaced the TPR valve on top of the cylinder. That was about 5 years ago and the problem didn't occur again until a couple of weeks ago. The "jacket" (by which I mean the insulating foam around the cylinder) is wet near to where the TPR valve is, which is the only visual clue that exists. I recall this from last time, which is perhaps why the TPR valve was considered the culprit. I'm just unsure why a TPR valve would trigger / fail when there shouldn't be any ability to create pressure. Unless something specific happens when a sealed hot system goes cold and presumably creates compression, rather than expansion?

Thanks for all of your help.

DC

 

[Best]

After you turn your mains stopcock off, plus run a mains cold tap to confirm water completely off (no continual drips) then turn a hot tap on to release pressure.
After all that there should be no possibility of leaks except in storage - toilet cisterns or hot cylinder and no possibility at all of pressure leaks, unless mains pipe to stopcock.

 

[DC70]

I did run the taps after shutting off the stopcock to confirm the water was off. I am wondering, in a sealed system, is it sensible to leave a hot tap open when the stopcock is off, just to create some breathing room? I am thinking that if a cooling system creates compression and the stopcock is closed, what happens when the cylinder can't access the mains intake?
DC

 

[Staffy85]

get everything dry and reenact the situation as best as possible with getting that cylinder cold..your probably notice something as soon as its cold.

 

[Staffy85]

I am thinking that if a cooling system creates compression and the stopcock is closed, what happens when the cylinder can't access the mains intake?
DC

it doesnt anyway as theres a non return valve to stop water/expansion going that way.....or what happens to a cylinder full of hot water { fully expanded} thats allowed to cool down {contract} in a sealed enviroment, creates a vacum ? food for thought..

 

[Chuck]

The last time this happened, a plumber replaced the TPR valve on top of the cylinder. That was about 5 years ago and the problem didn't occur again until a couple of weeks ago.

The TPR valve should discharge via a tundish and the pipe work should be designed to ensure that it can handle the discharge without causing a flood. This should be tested annually by a G3 qualified competent person.

My guess, based on the limited information provided (no model number, no photographs, etc.) is that the joint (flange/boss) between the TPR valve and tank has a crack in the weld. A specialist leak-finding company might be able to prove this using a tracer gas technique but it will be quite expensive and there's no guarantee of a definitive result.

Personally I would simply replace the whole lot (tank and all pipework and fittings that could possibly be to blame). You have now had two warnings that something is wrong and were fortunate to get away with limited damage. Next time you could get back to find your home destroyed by water damage. When that happens you can look forward to your insurance company trying to avoid paying the claim because of what they'll call 'contributory negligence' on your part for failing to get a known fault promptly and effectively remedied.

 

[DC70]

Thanks for all the further replies.

If there were a crack in the weld, I would expect there to be a continual leak, especially when the pressure builds back up. But there is nothing, either when the system is hot or cold.

The last time this happened, the attending plumber blamed the TPR valve and replaced it, but I am not entirely convinced this was to blame. After all, where would excessive pressure and temperature come from in a system that is off? Also, the TPR valve does release into an "overflow" pipe, so apart from any splashes out of the inspection hole, this still shouldn't cause a sudden flood. Manually triggering the TPR valve shows no signs of unusual leaking.

The reason I posted here was to find out if I am missing an obvious cause which I could then seek a plumber to correct. It doesn't seem there is an obvious culprit so I think Staffy85 has the most logical solution - re-enact the shutdown and see what happens. Then I'll be able to see what's going on and brief a plumber efficiently.

Thanks everyone for your time helping with this. Much appreciated.

DC

 

[Chuck]

If there were a crack in the weld, I would expect there to be a continual leak, especially when the pressure builds back up. But there is nothing, either when the system is hot or cold.

Cracks can remain normally closed and only leak under the influence of transient stresses caused by differential expansion/contraction as the system heat/cools. Tends to be a problem where two objects with different thermal time-constants are joined. One expands/contracts faster than the other, which stresses the joint.

I'm not suggesting that it's the only explanation for the fault, just one possibility that explains why you see it so rarely.

 

[ShaunCorbs]

Any reason why you haven't posted any pictures?

Also to repair unvented cylinder you needs to be G3 registered, also they need servicing every year else a leak from them that causes abit of damage isn't covered by your home insurance

 

[DC70]

Nothing really to post pictures of - no evidence of recent leak etc.

The plumbing throughout our apartment is subject to annual checks and services as part of the estate-wide service charge. This was put in place a few years ago to combat the rising cost of insurance, which is built into the service charge, which was being adversely affected by flood related claims. Obviously when you have neighbours below you, insurance claims related to floods can multiply!

When I get to the bottom of what's causing this, I will be sure to get a G3 registered plumber if the leak is indeed related to the hot water system- thanks very much for that advice.

DC

 

[Vee]

Just a thought. Anybody checked the PRV / Combi Valve on the incoming water supply. Could be excess external cold main pressure when system not in use due to fluctuations and PRV not regulating down? Just coincident that it happens whilst you are away?