Opinions- hydrostatic test data

[Eric Myers]

Hi all, I'm looking for a second (or more) set of eyes on some pipeline test data. The pipe in question is approximately 600ft of 24", mostly exposed except for a 30' section.

Here is a chart of pipe temperature and pressure for the last hours of the test.

Some of my own observations:
[ul][li]Recorded pipe temperature reached a low of 59F at .4 hours, and rose back up to 60F near the 2 hour mark. I'm guessing the pipe reached an actual low around the 1.25 hour mark and temperature started climbing back up after that. Per my own understanding, it would be generous to say that the lowest pressure should have been 386F if everything was okay, and should have started climbing from there.[/li]
[li]As pipe temperature climbs, pressure continues to drop until 3.2 hours on. The temperature is 64F at this time, and my calculated pressure is around 468 psi. Actual recorded pressure is 362 psi.[/li]
[li]At the end of the test, pipe temperature has risen 10 degrees F, but the pressure has only now gotten back up to what it was at the time the temperature started going up.[/li][/ul]

Is anyone aware of a way that the data could look like this without it indicating a leak?

 

[hydtools]

Pipe dimensions are expanding faster than fluid volume with temperature.

Ted

 

[Eric Myers]

Thank you for the response Ted.

Is this something you've dealt with in the field, to the extent of being comfortable accounting for it in calculating pass/fail criteria on a test?

I haven't reviewed many of these, but have seen some. In other tests I've looked over, the response has been fairly immediate- when pipe temperature goes up, pressure goes up. When it goes down, pressure goes down. To my thinking, this isn't just a short duration discrepancy when something wiggles on the test, it's over a period of four hours. And it doesn't disagree with the calculations by just a little, it's off by 200 psi by the end of the test.

If this could be explained by a phenomenon that is solid enough to justify accepting test results, I would be very interested in being pointed toward how to quantify it in a way that would stand up in court if it ever came to that. In my current case I'm sure I've lost a customer, but for education's sake I'd like to get on top of it.

 

[1503-44]

Pressure dropping at constant temperature is not going to turn out well.
And with no leaks, the rise in pressure per degree should be the inverse of the drop in pressure per degree. That's obviously nowhere close.

 

[LittleInch]

Why did the temperature only increase after 2 1/2 hrs?

Did the sun come out?


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[LittleInch]

But that drop in pressure is extensive.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Eric Myers]

Correct, the test was started at night. The point where temperature reading rises to 60F is 7:30AM.

 

[Eric Myers]

Here is a look at the data over the full duration of the test. This includes the time before the last injection, while pipe temperature was measurably dropping. Over this period, the pressure drop rate is a little more steep than the allowable predicted, but not wildly off.

 

[1503-44]

So really it lost 150 psi.
How much water was it initially filled with and how much was added during each reinjection?
What is the wall thickness of the pipe?
I quess it is steel, right?

 

[Eric Myers]

Start to finish, disregarding the injections, yes, about 150psi sounds about right. Steel pipe, .375 wall.

My bad- I got lazy after seeing how hosed-up the overall data looks, and didn't work the details on the injections. The first injection was 7.78 gallons, second one was 6.2 gallons. Fill volume is 13716 gallons. So by my calculations this would put the final pressure closer to 300 psi low rather than 200 psi.

 

[1503-44]

Did you meter the initial fill of 13716 gal, or is that the calculated volume of the pipe?

 

[Eric Myers]

Calculated- 615ft of 24" pipe + 16ft of 16".

 

[Eric Myers]

Since so far it seems like there's still sufficient reason to not accept the test results, I'll invite opinions on another issue I'm having.

My client for this job is the contractor who performed the test. When I warned them I wouldn't be able to sign off on this as a passing test, I got a succession of calls protesting that the test must somehow have really been okay.

I was informed that the following day they came to the consensus that the pipe temperature data must have been wrong. They sent the records to another engineer who signed off on it. I haven't seen that report so I don't know if it addresses this question in any way.

I informed the contractor that I would need to bill for my time, and they said that is fine. When I do, I'll supply a report including my observations on why I didn't find the test acceptable.

I also believe I'm obliged to make sure that the asset owner is aware that the first person to see the data had a problem with it. I have something of a trust issue with the contractor after the way they received the news and how they handled it, and I am tempted to copy their customer in on the report without asking. That's obviously a harsh route to take, but in the end I don't want to be left with doubts about whether the owner will ever actually see my notes.

Thoughts?

 

[1503-44]

Initial fill should have been metered too, so that there is confirmation of absense of air.

I would be tempted, but just send your client your full report via registered letter or FDX. Keep your copy of the notification of receipt in your archive. That should at least be enough to keep you out of trouble if things head south. Technically your responsibility ends there with your client, so unless you believe there is eminent danger, leave it at that. Are there extenuating circumstances, such as a pipeline built on public land, or is it piping confined to plant limits?

 

[LittleInch]

Well, when you run the numbers it's not actually too far off for the fall off in pressure vs temperature change to it not being a leak, but it looks like the surface temperature rise has not occurred over a long enough period to allow the water temperature to catch up. Also there is no mention of what the initial WATER temperature was. The water has a very large thermal mass compared to the steel and will change slowly. If it wasn't exactly 65F at the start then all bets are off as this can't be accounted for properly.

From the graphs you can see than min pressure is 3 hours AFTER min temp. The pressure graph is starting to climb fast when the test ended and if allowed to finish or stabilise at 70F the pressure would have quickly got to the point where you needed to remove and measure the water to prevent overpressure.

This test is a farce - at no point has sufficient time been allowed for such a large volume of water to stabilise in temperature or pressure.

However you can say that the pipe has passed its "strength" test and clearly not failed, but you cannot show that it has passed the leak test as neither the temperature or pressure stabilised enough to be able to draw conclusions.

Doing pressure tests on above ground pipe is hard when exposed to sun and diurnal temperatures.

So the key question which you don't make clear is what sort of test was this? Strength or leak? To what code?
If 31.3 it requires you to examine the pipe and welds, even the underground one, during the test for leaks. Did anyone / did you?

I don't think sending your report to the client is a good idea. You haven't certified the test, someone else has. All the responsibility is now on that person / company, not you. I don't think there is any way with the violent changes in temperature and pressure that you can draw any conclusion about whether this pipe is leaking or not. All you can say is that it didn't fail.
What possible good is going to come from sending your report to the cleint with whom you have no contractual arrangement?

Only if the contractor forged your signature or said verbally or in an e-mail that YOU had said it had passed the test would I be getting upset.

Just send in your report and invoice and move on. IMHO.



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[1503-44]

I was too lazy to do any calcs, but LI, you could be right. If you assume it is stabilised at 3.0h, it could be OK. Does it track dP-dT closely for those 2 hrs? If so, then if the test was allowed to continue for 8h after stabilisation (another 6h), that might have proven as OK. The problem is getting anything to stabilise with 600ft not buried. But being unburied allows you to easily get visual confirmation of no leakage, so that kind of overrules calculations anyway. 8h is usual minimum for pipeline work, so not enough time to qualify for that.

 

[Eric Myers]

Thanks to all for your thoughts.

Agreed that the fill should have been metered.

There is no record of the initial water temperature, I assume it was somewhat warmer than 65F. I did take a look at average water temperature over time when I started trying to understand what might have been going on. I ran two cases, one with initial water temp at 65F and one at 80F. I can't say that I went in depth into it though. I'll go down that path some to see if it might give a plausible alternate explanation for these numbers.

The pipe is mostly exposed. It passes under the road for 30 feet, and in my thinking that is both an opportunity for corrosion or something else to occur, plus a way to conceal it from inattentive eyes.

The test was for leakage, 49 CFR 195 E. I will say this- if this data is acceptable in anyone's book, I don't really see the point in the tests. If they accept this, they'd accept anything, so why bother?

Regarding my reason to consider sending results to the owner: in my view, if I find that there is strong reason to believe there really is a leak, then I have a responsibility to see that the owner is at least aware that there was a concern, if the contractor doesn't take any action other than to shop around for someone else who's willing to say everything's great. The pipe is in a plant- but we all live on the same planet in the end, so the stuff that goes in the ground there is going to end up in the water table that the kid down the road drinks from.

 

[1503-44]

195 I think is a minimum of 8hr hold time (after stabilization).

 

[Eric Myers]

8 hours, but no mention of stabilization. It's very rare that I've seen any of these not move around some, but it's always been in the error bars. The temperature swing in this one is definitely more than I've encountered previously.

 

[1503-44]

8 hr with proof of no leaks. As you've seen, that is difficult to prove before stabilisation.