Manual Valve Testing Plan 1

[Pavis1968]

Hi All,
Could somebody help me to find out which Code or Standard is specifying the valve percent to be hydrotested. Our valve supplier is replied as per below:
Type Qty Rating Percentage
Gate 840 150, 300, 600 20% persize/per rating
Globe 90 150, 300, 600 20% persize/per rating
Check 100 150, 300, 600 20% persize/per rating
Gate 650 150, 300, 600 10% persize/per rating

I couldn't find in our Spec the criteria for this and I will very appreciate if sombody help me to find out this.

Thanks and Regards,

 

[bcd]

For hydrotesting at 150% of the rated pressure, every code or standard I have ever seen for valves over the last 30 years requires 100%

If you are requiring a special test that is in addition to normal valve shell (hydrotest) and seat tests, then some end users allow testing of 10% of the lot but usually state something that if one fails, then the entire lot must be tested.

In short, all valves should be hydrtotested to 150% sometime before being put into service. Could be in the valve factory or when the entire piping system is tested. All valves should have a seat test to verify quality. Seat test conditions should be the most challenging for the valve to pass (i.e., don't do a high pressure seat test at 1.1 times the rated pressure on a soft-seated floating ball valve as it will easily pass. Use a low pressure air test).

 

[reverman]

The typical valve test standards are as follows:

Bronze valves: MSS SP80
Iron Gate: MSS SP70
Iron Globe: MSS SP85
Iron Check: MSS SP71

Steel Valves: API 598 or MSS SP61 ( the differences between the two are duration of test and acceptable leakage crteria, SP61 also doesn't call for a backseat test)

Typically you do a shell test at 150% the rated pressure, the a seat test at either 60-100 psi Air (gate valves) or 110% the rated pressure with water (globes and checks). The high pressure seat test is done on high pressure class gate valves in lieu of the low pressure closure test.


ASME B16.34 also has test criteria listed but it does not have acceptance criteria for the seat (closure) test.

 

[NGiLuzzu]

I agree with bcd!

Other reference standards: ISO 5208 and EN 12266-1.

For more detailed information, search this Forum for the topic in object and you'll find many discussions, like thread408-208032 and thread408-210466 (just to give a couple of examples).


Hope this helps, 'NGL

 

[Gorman3]

All valves must be tested to API 598, that's an absolute standard. What they are probably referring to is the random selection performed after wards, which can vary greatly. Testing alone can cause damage, so doing a decent random selection after the initial test is important.

You really can't assemble a valve properly if you don't test it. It'd be like building an engine and not checking if it starts.

 

[greg1950]

The original valve design standard will stipulate which test must be performed at time of manufacture. In order for the valve to meet its design standard (API 600, ISO 10434, API 603, etc..) EVERY valve must be tested at the factory at the time of final assembly. The most common test standards in use today are API 598 and ISO 5208.

However you might have an end-user requirement for re-testing a certain percentage of valves before they are installed. If this is the case, there are no standards to cover the percentage of re-test. Generally a small percentage (2-5%, sometimes 10%) are tested and if any fail, additional valves are retested.

 

[proinwv2]

If the testing described by greg1950 is in fact what is happening, then there are statistical testing methods that can be referred to that will tell the random sampling to be done based on population and on the accuracy obtainable. I don't have those standard numbers handy.

But regardless, all valves are to be tested 100% at final assembly by the manufacturer as attested to by the others here. I have worked for several major manufacturers and we always did this.

Paul

http://www.ostand.com