Pipeline hydrotesting in winter temperatures.

[mjh250473]

80% of pipeline is backfilled, remaining 20% is exposed. we use weighted values for start and end temps to give us overall DT that is used for allowable DP for the overall DT in the Shell Depp formula.

On all previous tests, as normal we see a pressure drop when temperature drops and calculate our DT as follows:

Start temp = (Tground*0.8)+(Texposed*0.2)
Start temp = (7.24*0.8)+(2.2*0.2)
Start temp = 6.232

End temp = (Tground*0.8)+(Texposed*0.2)
End temp = (7.18*0.8)+(1.6*0.2)
End temp = 6.064

DT = 0.168 Drop

The most recent test was carried out in very cold conditions and the relationship between P and T inversed on the exposed pipe. In general when the ambient got very cold so did the exposed pipe and the pressure went up. When ambient went up so did the exposed pipe and the pressure went down. Tground remains almost constant but behaves in the normal fashion (slight Tdrop for drop in pressure).

This left us concerned on how to calculate the DT and i question whether to use the same method as before. ie

Start temp = (Tground*0.8)+(Texposed*0.2)
Start temp = (7.24*0.8)+(0.73*0.2)
Start temp = 5.938

End temp = (Tground*0.8)+(Texposed*0.2)
End temp = (7.18*0.8)+(1.99*0.2)
End temp = 6.142

DT = 0.204 Gain

The effect of the Exposed pipe has caused the overall DT to be a gain, when the pressure has actually dropped (Pstart 155,465 Pend 155,380). This causes obvious problems.

Now as the inverse relationship is occuring would it be correct to say that as the exposed pipe temp increase should be used as a drop i.e.

Start temp = (Tground*0.8)+(Texposed*0.2)
Start temp = (7.24*0.8)+(0.73*0.2)
Start temp = 5.938

End temp = (Tground*0.8)+(Texposed*0.2)
End temp = (7.18*0.8)-(1.99*0.2)
End temp = 5.346

DT = 0.592 Drop

Which is preferable, but is it correct?
Thankyou in advance

 

[rconner]

Was curious exactly how/on what/with what your are measuring these various temperatures -- also, is this a welded steel pipeline, testing with unblocked end blank/bulkeheads or flanges, or what? Sounds like the length of the pipeline could perhaps be contracting with cold, differentially compressing further the contents?

 

[dicksewerrat]

the backfilled portion of the pipeline is probaly warmer than you think. when you are filling with ambient tempfluids the backfilled section warmed up the fluids.

Richard A. Cornelius, P.E.

http://WWW.amlinereast.com

 

[mjh250473]

welded X70 pipeline, welded high presure test headers. 3km test section. Buried section temp data loggers - to fit probes - dig down on pipe connect probe to pipe and backfill. Exposed pipe same technique without the back fill.

Were going to try and cover the exposed sections with rockwool/hay to stop them freezing. We are still unsure on exactly how to handle the previous test.

 

[CRG]

First, I am not a big fan of using a hydro test as a sensitive leak test for buried pipelines. I have seen many engineering firms specify temperature compensation similar to what you are doing with unrealistic assumptions about the accuracy. A company called “Vista Leak Detection” has developed test methods and math algorithms to compensate for temperature and pressure change. Take a look at their web site and review their publications. They used to have quite a bit of information regarding their test methods; however, the last time I looked for one of their better papers, it was no longer listed for download. I think it was pulled because it gave away too much proprietary information.

I have no idea what a “Shell Depp formula” is. Please post the formula with units.

I assume that the temperatures listed are in Celsius. If this is so, then you are working in a very non linear area of temp-density relationship for water. Around 36 degF the density of water starts to decrease as the temperature drops. Above this temp the same is true. There is an inflection point around 36 degF in the curve. This is why most lakes do not freeze all the way to the bottom in cold climates (that is another topic that I will not expand on here). Your assumptions of using (Tground*0.8)+(Texposed*0.2) for an equivalent water temperature will only be accurate when your fluid density is linear with respect to temperature.

 

[Drazen]

Uf, i saw many formulas in design specs, but never something that was applicable in practice.

We agree to cleverly call pipe testing "hydrostatic testing". That means put test pressures on pipes, but note the pressure for only limited amount of time, 2 hours maximum. Insulation is good, just to prevent temperature change during these two ours (between 1 and 3 PM is the best during the winter). Temperature measuring is good, just to make sure there are no temperature changes during this period.

This is good enough for reliable test results. If you are inclined to burst safety, you can hold pipe on pressure as long as you want having freeze protection implemented - just you don't need to officialy note pressure drop! If you hold pipes on pressure during say 12 hours, any leakage will show off as so big pressure drop that you will not need official formula to check with - common sense is reliable enough.

 

[rconner]

More precise hydrostatic testing of a long section of piping is no doubt a potentially complicated subject, particularly with changes in temperature. Thinking long and hard about this, "the pressure went up" is probably generally lesser of a worry than the other way around, at least for a freestanding, blanked section of piping! To analyze this further, it seems one would have to have more data or make assumptions e.g. with regard to the temperature/changes and maybe also amount of air/trapped etc. also in the pipeline contents (I have heard folks in the past say they had seen pressures go up when a pipeline with exposed section heated up, however I had not previously heard the other way around - there is always something to learn!)

 

[Drazen]

Generally in life, if you decide to hold filled pipe for 12 hours or similar, most often end of your testing period appear to be in colder part of day (evening or morning) and of course that is sole cause of the pressure drop of there is no leakage - if there is leakage apparent large pressure drop appears over such period.

The point is not to stick to official noting of pressure over such period. Short period of 1-2 hours should be sufficiant for noting pressure differences while avoiding influence of temperature changes.