Oil Cooler Internal Leak Detection Methods ? 1

[frns4ever]

Hi,
Is there a better procedure to detect an internal leak from an oil cooler? Internals are brazed. There is an inline air leak detection equipment set at 5psi to detect the leaks. Still, some are getting missed, as there found to be small gaps enough to let the cooler pass the air leak check. Those gaps are getting expanded during vehicle operation and a leak path is formed. Serial production couldn't find these.
Any recommendations?

 

[TugboatEng]

Why test at only 5 psi and not, say, 1.5x operating pressure like a typical hydro test? You want it to fail on the test stand so you need to abuse it on the stand.

As an operator, I have never experienced a leak from a brazed plate heat exchanger. It has always been the o-ring on the cooler flange. Most types of rubbers can't tolerate both a hot water and hot oil environment.

 

[frns4ever]

Hi. Thank you for your feedback. Just to clarify, this is for the inline serial production coolers which are close to 500 a day. Cannot have them tested to fail. Few samples/day are being tested to fail as you said.
An example of the internal leak is for the trans cooler which cannot see a drip outside. It can do damage for the trans as the oil and coolant get mixed up, leading to a bigger repair. Need to avoid that. The same applies to the engine oil cooler.

 

[TugboatEng]

I'm no suggesting to test all to failure. You want the weak units to fail on the test stand so they can be rejected. You need to test to a higher level than would be seen in normal operation on order to rule out failure. If the normal operating pressure is 100 psi, the coolers should be tested to 150 psi.

 

[Tmoose]

"Cannot have them tested to fail."

1.5x operating pressure should really NOT be "test (good parts) to fail."

 

[SwinnyGG]

In order to diagnose what's wrong with your test, we need to know more about your test.

Are you pressurizing only the hot or cold side of the exchanger? How long is the pressure held for a part to pass? Is the supply pressure from the equipment shut off during the measurement period (this is very important)? How are you detecting failures, and how are you certifying 'good' parts?[pre][/pre]

 

[EdStainless]

Is this being tested air underwater or are you pressurizing and looking at pressure drop?
We used to air underwater test SS tubing at 200psi and do 5,000 pcs per shift.
A tank of warm De-Ionized water (kept pure), quick release chucks, and some large accumulators for clean, dry compressed air.
I don't see why you wouldn't 100% test these. Sounds very short sighted.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Compositepro]

It does not sound like your problem is one of leak detection, but brazing quality control. You have weak joints that do not leak until they have seen some stress cycles. If that is the case, 100% leak testing using your current procedure will not help. That is why others have suggested a higher pressure test.

 

[enginesrus]

I would think the manufacture of the cooler would have specs on the proper pressure to use to check the unit. It should be at the proper temperature as well.

 

[SwinnyGG]

It does not sound like your problem is one of leak detection, but brazing quality control. You have weak joints that do not leak until they have seen some stress cycles. If that is the case, 100% leak testing using your current procedure will not help. That is why others have suggested a higher pressure test.

Respectfully... I couldn't disagree more.

It is a literal impossibility that a process which makes brazed heat exchangers will make perfect, leak-free heat exchangers 100% of the time, indefinitely. It cannot possibly happen.

What is possible, and what OP needs, is a test which is robust enough to catch all reasonable failure modes, regardless of how often they occur.

Test design is a oft-neglected component of part design, but it's just as important as all other aspects.

 

[moon161]

If there's a possibility of internal leakage, that implies multiple cavities. Run sequential cavity decay tests (important that they run sequentially).

 

[GregLocock]

The grown up answer is you write an FMEA and then generate a DVP and an in-production quality plan.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?