Pressure drop during hydro DIGITAL VS ANALOG 1

[viper3982]

What is an acceptable pressure drop rate for a hydro test?

We test units produced at our shop to both Sec 1 and 8. We perform PED, IBR, GOST jobs. We provide CRN when required...we are very experienced in dealing with code hydros.

Typically we hydro at pressures around 4KSI to 6KSI. Typical volume for our units are ~1gal of water for the pressure test.
Our analog gauges are appropriately sized for the code requirements.

Recently we switched from analog pressure gauges to a digital gauge. The digital gauge is much more accurate than the analog (as is required by code).

We are noticing a very minor pressure drop. The gauge (which reads accurate to 1/10 psi) is showing ~2-3psi drop per minute.
We attribute this to either ambient loss, any very minor change in pressure for that small of a volume will show a drastic pressure change.

Per our corporate safety group we are not allowed near equipment during hydro-test (an argument for another day)...but more importantly our hydro test pump/kit has many hoses, fittings, adapters, etc. It meets ASME requirements for non-interference and what have you...but point being: it has many places for super-minor leaks.

The 3rd party (NOT AI) inspector for our project is rejecting the hydro saying that we are unable to hold pressure.

If we were to swap to an analog gauge...this pressure change would not even be noticeable on the scale over the course of an hour. (~100psi increments on a 4" scale)

Does anybody know of any tolerance that would allow for either ambient loss of a hydro system?
Can anybody suggest methods of talking to/conveying this concept across...shall we say...individuals less technically trained?
Thoughts?

 

[Don56]

Could something be yielding, a hose? You say it drops 2-3 PSI/minute. Does it stop dropping? If you pump it back up will it hold?

 

[viper3982]

There are no evident leaks. We check the unit, the hoses, the only thing we can't see inside is the hydropump but we do not find anything dripping.

We haven't held it long enough for it to stop. We have had it up for about 30 min and it continues to slowly drop.
Still no evident leaks.

 

[JR97]

Viper,

Just to be clear. The digital gages are not more accurate, they have better resolution. I suspect there is nothing new going on during the test except the resolution increase in the instrumentation. We run hydro tests in the same range you do but with about 3,000 gallon fill volume. What we do is increase the pressure incrementally with 5 minute pauses at different levels and then when reaching the test pressure do a 10 to 15 minute stabilization period. We find when the pump is isolated there will be a decrease in pressure, as you noted. I attribute it to temperature equalization plus some elastic growth in the unit or piping. 2 psi in a 6 Ksi test is only a .03% decrease. The rate of decease diminishes. After the 10-15 minute stabilization period the pressure is almost constant. We then begin the official hold period. If the rate of decrease does not lessen during the stabilization period, there is a problem that must be investigated. I do not know of any magical pressure drop that constitutes a leak. We occasionally have issues with customers as well. They don't understand a hydrotest is a proof test, not a leak test. The Code acceptance criteria is "evidence of leakage or permanent deformation". I do not consider minute drop on the gage display to constitute evidence of leakage, especially if the rate is decreasing . Can you re-program the digital readout to display in Ksi instead of psi, and then set for one or two decimal places? That would duplicate the resolution of the analog gage everyone was happy with. I have resorted to that when logic wasn't working.

JR97

 

[moltenmetal]

You are suffering from digital display bias- the tendency to treat a measurement with more displayed figures as if it has more SIGNIFICANT figures.

An analog display automatically gives the viewer a visceral appreciation of the readability/resolution of the measurement. That is completely destroyed by putting a digital display on a device- now the brain trusts the numbers even if the numbers are meaningless, or just noise.

Are you observing a real expansion of the parts? Maybe, or maybe the temperature of the essentially incompressible liquid you've filled the text fixture with has changed infinitessimally. If you're doing a really good hydrotest, which means ZERO gas padding anywhere, TINY changes in temperature or volume result in profound changes in pressure. But as JR97 points out, that's not the point of the test- the point of the test is to see if the component withstands the applied pressure without discharging any noticeable amount of material through joints or parent material. Your components are passing the test. Move on.

 

[viper3982]

JR/Moltenmetal,

That is effectively what we have come up with and the AI agrees. The trick is trying to get that info across to the customer. You both provided the same answer but with a little better wording than I was originally using so thank you, I will try to re-phrase it with the customer in that manner. Hopefully it will help.

/r