Capability Study for Leak Tester

[shivam465]

Hello All, I am trying to do a capability study for the leak tester machine we use for testing the blow moulded plastic vehicle fuel tanks. The test is only done to detect any leaks in fuel tank and there is no customer specification as for pressure. Now, shall i do a GAGe study or Capability study? the data type is attribute since its only pass or fail in leak test. I am doing this for the first time and would appreciate any kind of help.!!
Thank you

 

[John C. Abnet]

Good day @shivam465 , and welcome to the site.
It sounds to me as if your measurement technique is objective (i.e detects LEAK or detects NO leaks and there is no subjectivity applied by the operator. Is this understanding correct?

I'm not sure how this test can only be attribute. In other words, how do we know if the equipment is detecting a loss of pressure (variable data) or not ?

Yes, the result of the measurement would be attribute i.e. PASS (no loss of pressure) vs FAIL (loss of pressure), but how is the gaging system verified to ensure it is detecting ANY loss of pressure over a set time?

Gage R&R for attributes is looking at "agreement". In other words does an operator(s) "agree" with themselves/each other over a particular set of data repeated?

Hope this helps.

 

[Johnnymo62]

I think there are requirements for vehicle fuel systems. 49 CFR 301 or 571, I think.

 

[bobdoering]

You can buy standard orifices to test your machine. You can use them for calibration and sometimes gage r&r. But, capability requires checking parts - and they need some variability or you cannot detect the "difference" between them.

 

[shivam465]

Good day @shivam465 , and welcome to the site.
It sounds to me as if your measurement technique is objective (i.e detects LEAK or detects NO leaks and there is no subjectivity applied by the operator. Is this understanding correct?

I'm not sure how this test can only be attribute. In other words, how do we know if the equipment is detecting a loss of pressure (variable data) or not ?

Yes, the result of the measurement would be attribute i.e. PASS (no loss of pressure) vs FAIL (loss of pressure), but how is the gaging system verified to ensure it is detecting ANY loss of pressure over a set time?

Gage R&R for attributes is looking at "agreement". In other words does an operator(s) "agree" with themselves/each other over a particular set of data repeated?

Hope this helps.

Thank You so much sir for your help. As far as i know, there is no detection of pressure loss data, the machine runs the test at a specified pressure and if any leak of pressure is found, it shows red signal. I think I can only do capability and not gage study since there is no evidencial proof or base for operators to agree within themselves / different operator.

I really appreciate your answer sir.

 

[shivam465]

I think there are requirements for vehicle fuel systems. 49 CFR 301 or 571, I think.

I will definitely check these requirements thank you for showing.

 

[shivam465]

You can buy standard orifices to test your machine. You can use them for calibration and sometimes gage r&r. But, capability requires checking parts - and they need some variability or you cannot detect the "difference" between them.

I am not sure if I can use orifiices to test the machine since there is no variable output coming from machine as in Loss in pressure or any other data. Thank you for your recommendation sir.

 

[optomist1]

Hi Shivam465 welcome...a couple of questions:
1) how does one (your firm) define a "leak",
2) what is the duration of equipment looking for a "leak"?
3) 'Leak" test uses air as a leak gas?

 

[shivam465]

Hi Shivam465 welcome...a couple of questions:
1) how does one (your firm) define a "leak",
2) what is the duration of equipment looking for a "leak"?
3) 'Leak" test uses air as a leak gas?

A1) We are using Differential Pressure decay for testing the leak in fuel tanks.
A2) The cycle time for leak test is around 8.3 minutes. The leak tester looks for pressure leak in the test specimen by filling in air uptill the pressure that of the reference/ standard fuel tank already mounted onto the machine ( used as Standard condition). If the test specimen shows no sign of pressure decrease untill the pressure inside reaches similar to that of Reference tank, it shows green light for Pass, If there is any noticeable pressure drop it will stop creating pressure and will show Red. The exact location of leak operator has to manually find by using soap water.
A3) Air is used as Leak gas..


Adding on some photos to help understand better in the attachments.

Please help me suggesting how should I do the capability test for this machine.
ThaNk you very much for your reply.

 

[It Was I Who]

Hi Shivam

I have just glanced through our question and let's see if i get it right.
You have a specified pressure that you pressurize the object with for a specified time. And after that time the light become green or red. (Is that right?)
If so you will probably have a reference volume somewhere. It could be a built in, in to the machine or it could be an equal reference as the volume you are testing. Never the less it will probably do the test similar.

Both volumes will get pressurized with the same pressure and then the valves will shut close. And in-between these two volumes ( The reference and the test object) there is an membrane in the machine that will detect the difference between the two. Ex. if the test object leaks the membrane will bend over to the reference side and will then registrat the deferens ( In Pascal most common.) there will be a Max pascal limit that will be triggering the red light to signal leak if it is exceeded.

So you have most probably a specified pressure for the object, a specified time and a max limit of pascal. ( The pascal will probably be recalculated into mm3/S Cubic mm per second. That is likely from the specifications for the tank
So if you need to make an R&R on this you need 5-10 pieces for the test that you will circulate one by one through the machine. ( you will need this due to that you need to ensure that the force in the test object is worn of between each test, so you don't have any left when you do the next test with that object) Now you need to get the reference Pascal that is set in the machine to be able to make the calculation.

I would say it is an variable gage and not an attribute. You will get an "Attribute answer" out from the variable test the machine is doing.

If not the test is conducted under water or with soap, to do a "Bubble test" then you have a attribute gage.