Leakage from welded diaphragm gasket of Heat Exchanger

[franklin55]

We have a leakage of heat exchanger from the diaphragm.The pressure inside is 180 kg/cm2. The diaphragm configuration is attached herewith. Can any body advice: will the tightening be helpful in stopping the leakage?

 

[desertfox]

Hi franklin55

Tightening the bolts might help then again it might make it worse, it depends on whether the bolts are already tightened to the correct preload, if so then tightening them more might damage the diaphragm and make the leak worse.
How difficult is it to replace the diaphragm or at least have a new one standing by if you attempt tightening the bolts further.
If you use a torque wrench to loosen a bolt by incrementing the torque up in stages it might give you an idea of how tight the bolts are at present.
Finally you need to get some info on the diaphragm in terms of what clamping force it needs etc, can you contact the maker?

desertfox

 

[SnTMan]

franklin55, I kind of doubt tighteneng the bolts wil help, unless they are VERY loose. This is not likely a gasketed joint, more often it is a welded seal aound the OD of the diaphram. I would guess the weld has a flaw. With the unit OFFLINE you might pull every other stud, inspect the weld and repeat for the other half of the studs. Repair as required.

For sure, get the manufacturers' recommendations.

Regards,

Mike

 

[rmw]

I agree with SnTMan. You have a leak in either the diaphragm or the diaphragm weld, probably the weld. Loose bolts might have contributed to that, but tighten them only to the specified torque from the OEM. He has also given you a good suggestion with respect to getting access to the defective weld. You should be able to determine the general vicinity of the leak with the unit on line in order to narrow down your area of search.

While doubtful, should it turn out to be a cracked diaphragm, replace it. Again, consult the OEM regarding the replacement as this is part of the pressure boundary and comes under the jurisdiction of any governing codes.

rmw

 

[franklin55]

We have tried to tighten the bolts in the vicinity of the leak point, and the leakage actually increased. We have a spare diaphragm but for replacing the diaphragm we have to shutdown the plant (2-3 days outage) which we are trying to avoid. Now in case if the leakage further increases and we have to shutdown the plant, what should be our strategy:
1. to replace the diaphragm (more repair time)
OR
2. to replace only the welding of the existing diaphragm from the point of leakage(less repair time).

 

[rmw]

Not surprising.

I am assuming by the nature of your question that isolating the leaking heater or train that the heater is in is not an option. You may want to make double sure of that as if this is an older plant, way back when, heater strings were often designed to allow heater repairs without having to take the plant completely off line.

If this is not the case, then based on the supposition that the higher probability is that your leak is a defective diaphragm seal weld, option 2 will be your quicker solution.

For those who are experienced with heaters with welded diaphragms, this type of leak is not terribly uncommon and the procedure outlined by SnTMan above is utilized often.

Even with the plant off line, hopefully you can isolate the heater and pressurize it through a manual vent in order to locate the leak while doing the procedure of removing and swapping the bolts (studs) around. But, once located, drain the heater and make sure the area to be weld repaired is good and dry.

Hopefully too, you won't have any ancillary grief - galled stud threads - to deal with while trying to fix the leak. Be careful there too. An ounce of prevention will pay off dividends. The studs are part of the pressure boundary code calculations too, so don't monkey around with them.

Not knowing where you are located in the world, there may be available to you companies who specialize in this type of emergency repair and who will come to the site with all the necessary tools and equipment to get this done in short order. If you own in-house maintenance people haven't done this before or have any similar experience, I'd recommend going outside to someone who has already done this many times.

rmw

 

[jte]

Yep. That's why I avoid welded diaphram designs! They're great for new construction - the capital projects types like 'em. The hidden costs of reliability and maintenance are not included in the capital project.

jt

 

[rmw]

jte,

In lieu of what? Most welded diaphragm heaters I have had to deal with were well out of the class of gasket sealing, pressure wise so what then, the hated torus ring?

rmw

 

[franklin55]

Thanks rmw for advice.

Your assumption is right, the heater cannot be isolated from the rest of the plant. It needs shutdown of the whole plant (Ammonia Plant. The exchanger is Synthesis converter effluent converter and cools raw synthesis gas with cooling water.

We will do this job with our own resources (in-house). One problem is the pressure test after repair...I want to know whether the pressure test is necessary.....If so it will take very long time as the piping spools need to be dismantled and blinds to be inserted there. I want know if only the DPT of the weld is sufficient without any extra tests (hydro test)

 

[rmw]

Test my old memory. Is that a 123C heater?

You pose an interesting question to which I don't know the answer. I'll have to lean on my colleagues more code savvy. I don't, however, remember ever pressure testing after closing a diaphragm sealed heater (after repairs). That may just be old memory too.

rmw

 

[jte]

rmw-

There are ways...

franklin-

I would not expect a hydrotest. I do not believe that the welded diaphram exchangers in our system are routinely hydrotested. For a bit more formal backup, I'd suggest taking a look at ASME PCC-2 Article 5.1 Fig. 1. This Figure is a flow chart for determining what type of test to perform on equipment which is being placed back in to service. Based on what I know of your situation, I wind up in the "Perform NDE per applicable code" box, not the "Conduct hydrostatic pressur test" box. Weld 'er up and PT and finish closing the channel.

PCC-2 is available at

http://catalog.asme.org/Codes/PrintBook/PCC2_2008_Repair_Pressure.cfm

but be aware that the 2010 edition should hit the streets around April. There will be no changes to the Figure... I think no changes to the article.

jt

 

[franklin55]

Thanks alot rmw...