can crack appear on valve body after pressure test? 5

[portialiao]

We have processed strict pressure test on a C.S ball valve with DN300,600#, during the test there's no leakage. While 6 months later on site customer find a crack with 20mm length. Believe us we did do all that need to do to prevent such thing, we wonder if the crack appear after pressure test? is it possible, do you have any similar experience? thanks for your help!

 

[portialiao]

Please not, the body material is casting ,WCB.

 

[KoachCSR]

It is possible that a crack in the valve body casting could develop following the pressure test. I've witnessed personally a crack developing in the casting (from a minute trickle) to a 2-3 inch crack (with severe leakage) over a time span of approximately 18 months.

It is suspected that a minor defect in the casting at the crack location was quickly deteriorated due to particulates in the flow (the working fluid was raw water -strained but not heavily filtered).

I have pictures, but am on assignment at the moment and am unable to post them. Should have access to them again in a couple of weeks.

 

[MikeMet]

It is very possible for a crack to form after hydrotest. There could be a crack present during hydrotest and the pressure was simply not high enough to cause the crack to open during test. However, this same crack could open due to thermal or pressure induced stresses, or environmental effects, while in service.

 

[gerhardl]

Compere to a crack in a car front window: an impact (stone or grit flying) can give an invisible, or almost invisible material damage, and because of internal stress and variation in stress (pressure and temperature) widen to be a crack across the complete window.

A crack in a valve body (or a ball) , without any other damage in the system, you wold strongly suspect to have started from some material imperfection and be widened from stress.

Metallurgic microscopic sample would probably be able to confirm the material fault.

A particular material imperfection may well have passed all previous inspections, even thorough different first-class types especially if they occur on 'non suspect' sections of the valve(cases seen personally).

 

[bcd]

Some things to consider.

WCB is a fairly ductile matrial and as such, would be resistant to cracking. In the failure analysis, you should look at carbon content, verify proper heat treatment, and look for possible weld repair somewhere around the crack.

Also, there is a chance pipeline loads could have deformed the body and caused it to crack. Again, for a ductile material like WCB, when it is properly produced, would tend to yield and deform before cracking. But if the material is not processed correctly and is not ductile, it could crack if the valve body was used to jack mismatched pieces of piping together.

Other possible source of cracking is low temperature. If the material was not produced correctly (checmical composition, heat treament, and weld repair), the nil-ductility temperature could be up near ambient and the valve could crack when stressed. Doubt nil-ductility is the case, but should not be removed from the list of possible cuases until the inestigation data proves otherwise.

Best of luck in your root-cause analysis

 

[portialiao]

Thanks for all your prompt response! KoachCSR, waiting for your photos!

 

[portialiao]

by the way, this C.S ball valve with DN300,600# applied for Natrual Gas pipeline was tested as per API598, in order to find crack during test in future we intend to test valves with 600# as per API6D, which reqiures much longer duration of pressure.

Most of the cracks appear on 600# valve bodies, we seldom find cracks on valves with lower Class.

Obviously reqiurement for pressure test of API6D is higher than that of API598, but I wonder why? can any body help me?

 

[JimCasey]

Sometimes there is a small internal fault that may not be exposed by machining. It can lurk just below the surface, and the valve can pass testing. Then when the valve goes into service, the inclusion can be exposed by solution/erosion/crack propogation. Once it begins to leak, the flowing process fluid washes material out of it and causes the leak to grow.

If you notice a pattern with similar leaks, you should conference with your foundryman and have him look at the sprues and risers in the pattern. It is entirely possible for the shrinking metal to tear as it begins to solidify, and skillful analysis of the pattern can correct this.

 

[portialiao]

JimCasey, thanks for your response, I need to clarify that the valve I mentioned is a new vavle that has not been in comercial use. Firstly we performed pressure test strictly as per API 598 and found no leakage, after 6 months on site after the valve was installed on line the pressure test was performed again and a carck about 20mm on body was found, see picture attached.

We will check the root reason when the valve is sent back to us.