Hydrotest Exceeds ASTM Pressure for Flanges

[PipelineChic]

Hi all!

I am reviewing some hydrotest data to confirm that the tests were acceptable and to sign off on the design. My concern is that the 600# flanges were tested to a max pressure of about 2300 psig in some instances, whereas the ASTM guidelines identify the max test pressure to be 2225 psig for 600# flanges.

I've read a few other threads on here and this degree of over pressuring does not seem to be a concern in real-life applications (since there were no leaks identified during the test)... but I need to be able to prove on paper, via codes & calcs, that this is acceptable.

So, question 1:
Am I missing something in the code that might specify 2225 as a loose suggestion instead of an absolute max?

2:
Is there a way I could calculate the actual 100% SMYS of an individual flange based on the actual properties (ex. grade of steel used) identified on the MTR's? Would it be calculated the same way as pipe (2*SMYS*WT/OD)? Or is there a different equation to use for flanges?

3:
I’ve identified the flanges as an issue, but are there any other fittings/components I should look into further that may have been over pressured (ex. valves, o-lets, plugs, etc.)??

(Note: pipe/flange sizes range from 1” to 30” in this system)

Thanks in advance for the help!

 

[LittleInch]

1) You need to look at the flange rating in ASME B 16.5, which varies by material and temperature. For class 600 at ambient temperatures which I assume your hydro was, this can be as low as 85 bar! or as high as 103.4. Allowable test pressures are 1.5 x rating rounded UP to the nearest bar or 25 psi increment. For the 103.4 material this is 2266 psig

2) You would need to go into the design of the flange to work this out, but I think you're barking up the wrong tree here. It's completely different to pipe.

3) Most piping components are good for 1.5 x the rating.. This varies by material so you need to find the worst material. Valves are only an issue if you've tested against a closed valve. The seats are only good for 1.1 x design pressure, but the body should be good for 1.5.

it is probably not "acceptable", but given that is has apparently occurred, it is "allowable". What are your alternatives? Scrap all the flanges? No Didn't think so.

when you say the flanges were tested, I'm assuming you meant that the piping system to which they were welded was tested to xxxxpsi??

Would be good to understand why they exceeded the 2225 - 2250psi range.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[PipelineChic]

Hey LittleInch! Thanks for your response!

To answer your questions, I'm doing an engineering evaluation for a client... So I'm not sure what the repercussions of this will be on their end, but I would love to find them an acceptable solution to justify not scrapping them.
Yes, the flanges were welded to the piping system which was being tested.
And, I, like you, would love to understand why the target test pressure was so high

 

[moltenmetal]

Testing a component like a flange to 2300 psig when it is good to 2250 or so is a non-issue- if the joints didn't leak and show no obvious deformation, put the piping in service.

 

[LittleInch]

I agree with mm. In theory you probably won't find a numerically acceptable method of making 2250 = 2300. However in practice this is a value that is not critical and it would be a very brave mn to recommend scrapping a system for this minor over pressure.

The thing you really really need to do is check whether the flanges are actually your weakest point.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Duwe6]

Bottom line is: They passed the hydro test w/o leakage. Yes, that test was 2.2% higher than what the flanges were specified to be able to withstand. You have proved that these flanges are at least 102.2% of the minimum specified strength. Period.

And please note that the hydro pressure was specified as 1½ or 1.5 x Design. It was not specified as 1.5000 x Design. Engineers tend to try to stay with digital items in an analog world, i.e. 1.4998 and 1.502 are both 1½. Yes, I'd NEVER get rid of my digital calculators, but you have to keep in mind the accuracy of the data points, and the required accuracy of a calculation. Most times, very close to 'perfect' is fully acceptable.

 

[SNORGY]

To me, if this crossed my desk, this would be one of those "ya, whatever" moments. I would pretend that (or, maybe *check if*) the hydrostatic test pressure recorder was poorly calibrated and just call it good; no harm no foul. But I can be a cowboy sometimes.

The above stated, let's assume the pressure test was charted with a two or three pen recorder. My first recommendation for PipelineChic is to examine the calibration results for the recording instruments, if available, to see if the excess pressure can be rationalized on the basis of falling within the accuracy of those devices. I think 3.37% error will be somewhat greater than the expected error after proper calibration, so the next thing I would be inclined to do is see if the temperature was charted coincident with pressure. If the temperature went up slightly during the test duration, this might have been enough to explain the overpressure. There are detailed flange calculations that can be performed to examine them for theoretical "failure", but for the sake of 3.37%, I wouldn't deem them to be warranted.

 

[SNORGY]

The following links might, or might not, be of value...

thread292-345697

http://engineering.moonish.biz/blog...ii-division-1-flange-calculation-spreadsheet/

http://pagetequipment.com/pdf/engineering_example_calculation.pdf

http://pipemill.com/pipe-flange-design-asme.php

Or, simply, refer to ASME Section VIII Division 1 Mandatory Appendix 2...although I would suggest, as above, the ensuing tedium is not warranted for the sake of 3.37%.

 

[PipelineChic]

Hey guys, thank you for all this information! This is super helpful!

SNORGY, I'll take a look at the calibration info - that's good to know, thanks!

LittleInch, What should I look at that might be weaker than my flanges? Would it be the threaded fittings? How would I determine the max test pressure on those?

In general, I understand that this thread/concern is pretty silly bc obviously the over pressure is minimal and it passed the hydrotest without leakage, failure, or any other pressure loss... but my hesitation is putting my PE stamp on the report and saying it's 'good to go!', ya know?

In digging in further, CFR 195 only discusses required minimum pressures for hydrotest... I haven't found anywhere in Subpart E (or any other Subparts referenced) that says we can't test over the max pressure. So is the max test pressure even important to my report? Or is it just a manufacturer/vendor issue - like if a component failed down the road it's not the manufacturer's fault bc it had been over pressured...?

 

[LittleInch]

That's the way out of it - even B 16.5 says "no less than 1.5 times" so it has passed that.... There is an implied maximum of 100% SMYS in many cases and in any event no more than 2% permanent strain to comply with 31.3, but the 2-3% you were over the minimum test pressure is not worth talking about.

The other components all have pressure ratings for the different size and thickness of pipe and screwed components. The biggest issue usually comes with "special" elements which are designed to a pressure, not a class rating. Most valves and fittings will all be #600, but something like say a meter could be 85 bar. All the rest of your items should be covered under B 31.3 design so as long as it passes that to the correct pressure then you're good to go IMHO.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[lsone]

I dont want to start a new thread as I feel we can add to this thread. More information, one spot...

Code of construction is B31.3.

We have a piping system that is Class 150 flanges, System test pressure is 2940kpa (426PSI). System operates at 10kpa, design is 1034pka.

When the system was being brought up to pressure the operators took the test to 4481kpa (650psi). The test had no failures or leaks. We know the root cause as to why. However acceptability of the system remains in question. Finding max yield strength acceptability is difficult for a reference point. Our engineers have done calcs based on ASME BPVC VII Appendix 2, flanges would have reached 175% yeild stress however this is more design related and they understand this. Piping and fittings are a non issue. The valves and flanges are the major concern.

It is difficult to find grounds to accept/reject the situation at hand. The owner is leaning towards cutting off roughly 15 flanges and replacing 5 valves.

Thoughts?

 

[racookpe1978]

What were temperatures of the hydro test and the calculated max design service conditions?

2. What was corosion allowance for the design conditions? That is - the hydro was done on "new" pipe (and new flanges), right? So, there was no corrosion in the pipes (or flanges) and thus the walls of the pipes, flanges and valves were thicker than worse case design assumptions and margins.

3. Measure the offended valves and flanges. Check them against actual pressure. The "code" works for nominal wall and nominal flange thickness, not actual as-built thicknesses - which will be larger (thicker) and thus less stressed. Don't tell your customer that "They are good." Show him the calc's and "... And this shows the actual flanges were not over-stresed by this particular high-pressure condition at this temperature at this time of line."

 

[BigInch]

You were still far from yielding the flanges.
195 has no maximum test pressure.

Testing to yield is not a problem.
Note that when converting service to hazardous liquids and CO2 it is one method of qualification,
"§195.5 Conversion to service subject to this part.
(i) Testing the pipeline in accordance with ASME B31.8, Appendix N, to produce a stress equal to the yield strength; and .."

What is surprising is that no pipe yielded at that pressure.

 

[LittleInch]

I'm kind of guessing here that these are quite small diameter pipes (6-8"?) as otherwise I think the pipe would have got into trouble at more than twice the design pressure.

To be honest if the test didn't create a failure then it should be ok. I can't see flanges and valves yielding at twice design - there's simply too much metal.

I might be tempted to replace the stud bolts and nuts because they might have been over tightened and stressed and it's quite cheap, but I agree -how do you prove there is no damage. The only thing I can suggest is you get one, make it up and test to destruction or yield to see where that point is. Or an FEA analysis?

Otherwise maybe x ray, UT or MPI them to see if there are any cracks?

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

 

[lsone]

What were temperatures of the hydro test and the calculated max design service conditions? 15 degrees C, tested with water. Max design isn't listed on our Line Designation Table, Just design pressure, temp, operating conditions and hydrotest pressure

2. What was corosion allowance for the design conditions? That is - the hydro was done on "new" pipe (and new flanges), right? So, there was no corrosion in the pipes (or flanges) and thus the walls of the pipes, flanges and valves were thicker than worse case design assumptions and margins. New pipe/fittings. 1.6mm C.A. when in service.

3. Measure the offended valves and flanges. Check them against actual pressure. The "code" works for nominal wall and nominal flange thickness, not actual as-built thicknesses - which will be larger (thicker) and thus less stressed. Don't tell your customer that "They are good." Show him the calc's and "... And this shows the actual flanges were not over-stresed by this particular high-pressure condition at this temperature at this time of line." I agree with this statement, actual measurements will have to be verifed by a Field Engineer. No biggy there

I'm kind of guessing here that these are quite small diameter pipes (6-8"?) correct, smallest being 2" largest being 8". Flare header as otherwise I think the pipe would have got into trouble at more than twice the design pressure.

To be honest if the test didn't create a failure then it should be ok. I can't see flanges and valves yielding at twice design - there's simply too much metal. Not disagreeing with you but to prove it on paper rather than a personal opinion is required

I might be tempted to replace the stud bolts and nuts because they might have been over tightened and stressed and it's quite cheap, but I agree -how do you prove there is no damage. The only thing I can suggest is you get one, make it up and test to destruction or yield to see where that point is. Or an FEA analysis? The bolt up was actually verified and found to be only taken to 25% yield based on ASME section VII appendix 2, so they aren't too concerned there. Gaskets will be replaced

Otherwise maybe x ray, UT or MPI them to see if there are any cracks? It is an option, but trying to find calcs or something to either get us to this point or another

 

[BigInch]

isone
Please delete your posts herein and start a new thread.
It's really not polite to hijack this one, especially in bold red and black.

 

[lsone]

Big inch, if you read my post you would see my intentions. The last contribution to this thread was Dec-11-2014, perhaps you missed that detail. As I bumped it with a very similar situation but far worse, I figured maybe we could finally get a more concrete answer seeing how that was not provided and with a very extensive search my findings support a lack of conclusion.

Further explanation... As I cannot find a "quote" feature, I bolded what others have posted and put my answer in red. I do not understand how that is offensive especially seeing the lack of attention this thread has seen in 5 months. Also, please look at the original posters profile, they have not logged on since Dec-14-2014. The only reason you posted again was soley due to me bringing a 5 month old post back to the top, but that is just my opinion.

Back on topic please and lets work together to come to an agreement.

 

[LittleInch]

Well the only other point that you could do is to measure against a straight edge and se if there has actually been any yielding. If the flange did yield, which I very much doubt, then it will no be to the same dimensions or will be ever so slightly "bent" where the bolts are.

Ultimately how far do you want to go and how much time/money compared to just cutting them all off and replacing them?

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

 

[moltenmetal]

Size matters here- as does the service.

Is the service nonhazardous? What are the consequences of a failure?

Nonhazardous service? You've merely done a test which has proven, by absence of leakage, that the system has integrity to that pressure- a pressure well beyond what it will experience in service. Continue to use it unaltered, unless there are parts that show obvious damage- those need replacement.

Hazardous service? Risk of loss of life if a failure were to occur? Consider that risk against the cost of the rework and proceed accordingly. Chances are, you'll decide to do the replacement purely as a paperwork risk mitigation exercise, because that will be cheaper than doing a thorough examination to prove that it is undamaged.

 

[BigInch]

Attached PHMSA doc discussing high test pressure.