Leak Testing A Heat Exchanger 2

[Casimo5]

If it is suspected that a tube or tubes are leaking inside a shell/tube heat exchanger, what impact, if any, do the testing conditions have if they are not the same as the operating conditions.

In our case, we have an exchanger that is suspect. The exchanger has cooling water in the tubes, and vapors going to a vacuum jet on the shell side. We are seeing a large amount of water (greater than normal) in the shell side. At this point, there are mainly non-condensibles in the vapor side (this is the 3rd condenser in the vacuum system), so the tubes are suspected of leaking.

When checking for leakes, we took both heads off the condenser (a two pass shell/tube exchanger) and filled the shell side with water and pressured it up to 80 psig. Then we dried the tubes as much as we could by blowing them with high pressure nitrogen, each tube individually. We then waited for 30 minutes and then blew the tubes out again, no water came out of any of them.

The conclusion is that there are no leaks. However, the condenser is normally above 200 deg F, and for the test, it was significantly cooler, say around 80 deg F. Could it be possible that any small cracks in a tube would close up when the metal contracts as it cools? Then when it heats up because of the process, they would expand and allow water to leak through? The tubes are stainless steel.

 

[doct9960]

Why not do an in-service ultrasonic leak detection?

Some companies offer this kind of service like this one...

http://www.uesystems.com/why_up_apps_heat_exchangers.asp

 

[metengr]

I would suggest going to a more sensitive exam like a helium leak test or see if you can flood the shell side of the heat exchanger. We have had problems in the past with MIC (microbiologically influenced corrosion) in austenitic stainless steel, main steam condenser tubes in service. MIC results in pits that are so small you can barely see them, visually.

We used two practical methods for leak detection;

If designed for this, flood the shell side of the condenser and check for water running or dribbling out of the heat exchanger tubes.

Or - use helium leak detection to check for very small leaks like MIC attack. It is expensive but it will find even the smallest of leaks in heat exchanger tubes.

 

[Casimo5]

metengr,

What you described is exactly what we did, but found no leaks. I agree that a helium leak test would give us a more definitive answer, but I think we were suspecting a gusher, not a dripper. ;-)

Thanks for the info about MIC. It'll give me something to research on a slow Friday.

 

[vpl]

I know some companies are doing eddy current testing on heat exchanger tubes to look for corrosion.

Patricia Lougheed

Please see FAQ731-376 for tips on how to make the best use of the Eng-Tips Forums.

 

[rzrbk]

Given a sample of the water from the shell side, a lab should be able to tell you whether it is cooling water, condensate, or process water.

 

[JR97]

Casimo5,

I suggest you first determine approximately what size leak you have. How much extra water is present in the shell? Over what length of time with water in the tubes? These answers should give you an approximate leak rate in gal/hr. Leak testing generally uses the units of cc(cubic centimeters) / sec. Once you know the magnitude of the leak you are looking for, you can decide on an appropriate leak test method; i.e. pressure drop, ultrasonic, helium, etc.

Post back your anticipated leak size and we should be able to match a leak test method that would be appropriate.

JR97

 

[Casimo5]

Well, the leak was thought to be a big one because when the down leg of the condenser was isolated at the hot well, roughly 80 ft. below the condenser, it didn't take long for them to fill up completely with water. Like I said before, it is the 3rd condenser in the system, so the amount of condensibles should be pretty low.

Since my original posting, we discovered that the steam jacket on the discharge of the jet directly before the condenser was also full of water. The condensate drain/steam trap was not working properly. Now we will have to see if somehow water is leaking into the discharge of the jet from the jacket.

vpl,

We were going to do eddy current testing, but the maintenance crew was in a rush to get the heads back on and get the unit back up and running.

rzrbk,

The lab LOVES it when I bring them stuff. They usually turn to me and say "Analyze THIS!"

Seriously though folks, my original question was whether or not the testing conditions have to be the same as the operating conditions when looking for a leak, and if they're not, does that affect the outcome?

 

[vpl]

Seriously though folks, my original question was whether or not the testing conditions have to be the same as the operating conditions when looking for a leak, and if they're not, does that affect the outcome?

I wouldn't think they need to be exactly the same. However, I think there's a good possibility that your test method wouldn't show anything other than a large hole in a tube. From what I understand, normally the water is in the tube and air is in the shell. The way it was tested was the shell was filled with air, the tubes were dryed and then air blown through them to look for water accumulation. This doesn't put the tube under the same conditions and any "crack" might well close up. From what I can tell about your test set-up, there would be no driving force for the water to move from the shell to the tube. However, under normal conditions the water is being moved and finds the path of least resistance. This may well be a hairline fracture. Additionally, you may have loose pieces of "crud" (to use the technical term) in your tubes that are heavy enough that the water wouldn't seep by if the fluids were stationary, but that could be jostled by the flow enough to allow leakage under normal conditions. Another thing with MIC is that you can have multiple small holes making up the leak.

You know you have a leak, or you wouldn’t be getting water into your vacuum system. It might “fix itself” for a while (and sometimes corrosion-driven leaks do that for a while, due to the crud getting fixed in place and acting like a plug). But you need to be coming up with a plan on what you’re going to do, whether it’s just retubing the whole condenser or identifying and isolating the leak. I can understand your maintenance people wanting to get the heads back on and get the unit back up and running... well, actually, I can understand plant management wanting that; usually the maintenance crew doesn’t care. However, you might want to glance at thread391-184788 as to the consequences of ignoring a known leak.

Patricia Lougheed

Please see FAQ731-376 for tips on how to make the best use of the Eng-Tips Forums.

 

[Casimo5]

VPL,

You're right, management wanted to get the unit running, but maintenance was also told that if the unit was down for much longer, they might as well work on a pump that was scheduled for maintenance the following week to keep from brining the unit down again, and it was a Friday.

Thanks for the link, I will check out the other thread.

 

[unclesyd]

When everything else fails back to the basics, check each tube, a partial tube, or just a section of tubes with either pressure or vacuum.
Give Expansion Seal a call and get their catalogue.

http://www.expansionseal.com/html/hepg7.htm