We have a few ball check valves with SW ends. The valve body and the piping are heat treated after welding. Now the check valves seem not seal. Would anyone advise the acceptance criteria for the leakage rate of the valves?

Personally I detest backflow on a check valve, its not really serving its function is it?
I have however seen plants leave ball check valves in with contaminants in the seat causing backflow for periods and didn't seem to care.

You need to look at time, cost, application, and decide.
I've had SW ball check valves installed and not had a problem with internals.....what makes you think the heat treat casued something?

It is really subjective.  The big question is always "why is the check valve there at all?"  If the answer is "to prevent spinning a centrifugal pump backwards" then you can possibly accept quite a bit of back flow.  If it is there to protect low-pressure piping from attached high pressure piping during an equipment outage (e.g., the compressor valves on a recip compressor), then you can probably accept much less back flow.  Ask the question and evaluate your backflow based on the answer.

In support of Zdas04, you have to perform a risk assesment.  What are the probabilty of small, medium, large leaks versus the consequences of small medium large. If the leak was like I described, a (relatively) medium size gas leak meant a big disaster. If its just water back feeding or draining a vessel, I'd accept the risk.  You state these are on maintenance lines, the probability of leak is small if there is a hand valve nearby (an assumption or better yet a must).  In this case I'd look at a carseal on the maintence valve to get close to zero leak while in operation and then a permit to open hand valve and a proceedure to assure that you do have a leak while in the maintenance mode.

MORE NEWS