Measurement length and “No Rupture” pipelines

(The busyness has got to me again lately, with no posts for nearly 4 weeks – apologies.)

Location classes in AS 2885 are based on the land use within the “measurement length”, defined as the distance at which the radiation from an ignited full bore rupture is 4.7 kW/m2 (the level at which an unprotected person will receive very serious burns in under a minute).  Because this definition is based on full bore rupture there are occasional questions about whether this definition applies to a “No Rupture” pipeline (one that meets the  requirement of Clause 4.7.2 in Part 1 of the Standard).

The answer is that the measurement length applies to ALL pipelines, regardless of their potential failure mode, because the possibility of catastrophic consequences from full bore rupture might have been the reason for selecting a “No Rupture” design in the first place.  You can’t adjust the measurement length based on failure mode because that upsets the logic behind this.

It is perhaps easy to confuse two separate distances:  the measurement length (used to determine location class) and the consequence distance for failure from a specific threat (used for risk evaluation).  There is nothing inconsistent in the following (in fact this should be the norm):

  1. Assign the location class based on the measurement length (full bore rupture radiation distance)
  2. Establish a “No Rupture” design if required by the resulting location classification
  3. Evaluate risks based on whatever radiation distances are applicable to the failure modes from any uncontrolled threats that exist at that location (e.g. pinhole due to corrosion, puncture due to mechanical damage puncture, rupture only if it is a credible failure mode)

Of course much same logic applies to reviewing the risks to an existing pipeline that happens to meet the “No Rupture” requirement in a high consequence area.

Advertisements
This entry was posted in External interference, Pipeline design, Risk assessment. Bookmark the permalink.

7 Responses to Measurement length and “No Rupture” pipelines

  1. Chris Hughes says:

    It should be blazingly (painfully?) obvious that the reason behind having to extend the T1 (say) requirements for a measurement length into an adjoining R1/R2 location is that a full-bore rupture in the R1 area which occurs closer than a measurement length to the T1 will still affect the T1 location.
    However I think we need to make it clear that there is a difference between the determination of location classes on the ground (where they are based on land use and population density) and along the pipeline (where it is based on the requirements to ensure the safety of the population.
    So we determine the location class by studying the maps and aerial photography to determine whether the count of houses per hectare or the type of development such as hospitals (Sensitive) or factories (Industrial) – in the latter two cases the location class will only refer to an area of a few hundred square metres. We then use the measurement length to determine the design requirements for the pipeline in the vicinity of each location class (e.g. No Rupture). What we should not do is define the location class as extending outside the line where the land use changes, as by doing so we fail to distinguish between the method of determining location class and the requirements for pipeline design.

  2. Chris Hughes says:

    Peter pointed out to me that my original comment was unclear. The point I was trying to get across is that I believe Peter is wrong – the actual location class has nothing to do with the measurement length, which only affects the way you have to treat the pipeline design at the transition between two location classes.

    When the pipeline moves from suburbia into the country, the transition from T1 to R1 (or R2) occurs as you pass the last houses on the outskirts of the town – that is fixed by the residential development and is completely independent of the pipeline. The design of the pipeline has then to maintain the T1 requirements for a measurement length into the R1 area, and how far that is depends on the size and operating pressure of the pipeline: so two parallel pipelines will have different measurement lengths but the point at which they transition from T1 to R1 will be the same for both pipes. For example, the Roma-Brisbane pipeline was originally a 10″ pipe which has since been looped with 16″: the different pipes will have different measurement lengths but this does not affect the point where the location class changes.

    Hope that makes it clearer

    • petertuft says:

      Despite our off-line exchange I’m still not completely sure that I understand where Chris is coming from. However his comments highlight a remarkable drafting oversight in the AS 2885.1. Chris comments “the actual location class has nothing to do with the measurement length, which only affects the way you have to treat the pipeline design at the transition between two location classes.” On a literal reading of Sections 4.3.2 and 4.3.3 of the Standard this is absolutely correct, but it was never intended to be that way.

      When the committee was drafting the Standard the whole point of the measurement length was that the location class IS determined by the land use within that distance either side of the pipeline. Probably most people understand that, but the extraordinary thing is that the Standard never actually states it explicitly (and I for one never noticed the omission). The use of the measurement length in Clause 4.3.3.(b) to determine how far a higher location class extends beyond the boundary of a lower land use was a necessary corollary of the basic principle but was never meant to be the sole use of the measurement length. We will consider clarifying this in the revision of the standard that following the one due to be published real soon.

      On the example of two parallel pipelines of different size, yes it is strictly correct that they have different measurement lengths. Therefore (if you are really pedantic) the location classification change at the edge of a town would be at a different location for each line. However for practical purposes this is a pointless refinement and my approach would always be to use the measurement length and resulting location classification based on the larger pipeline. You have little or no idea which pipeline is going to fail so the only prudent thing to do is adopt the more conservative alternative. I think Chris and I end up at the same point via slightly different reasoning here.

      • Chris Hughes says:

        Peter, I still dislike your interpretation. You say that the fact that parallel pipelines will have differing measurement lengths is a “pointless refinement”, and in the case I quoted of the RBP you are correct as both pipelines have the same owner and operator – but consider Braemar 1 & 2, or the area in Victoria where pipelines with different owners (used to be Gasnet and Santos, no idea who owns them now!) share a common easement – or any CIC pipelines for that matter. Using your definition of location class each pipeline will have a different boundary between location classes: using mine each will use the same boundary. Each will be required to extend the higher class design features for a measurement length into the lower class: these distances will be different for each pipeline. The result in terms of pipeline safety will be the same whichever definition you use but I honestly believe that my definition is easier to understand.

        The only reason I can think of for wanting to use the longest measurement length in a parallel pipeline situation would be if the rupture failure of the smaller pipeline could result in the rupture failure of the larger pipeline, in which case the measurement length would have to take into account the combined energy release from both pipelines. I don’t think this has ever been considered to be a credible event.

  3. Lynndon Harnell says:

    Chris, I agree with your logic I think. Different pipelines in size and or pressure have different impacts (measurement lengths).

    Btw there was a study done as to what criteria would have to be met for one (quite large) pipeline rupture would damage an adjacent one in a parallel CIC ROW, and it concluded that it the 2 lines would have to be extremely (unreasonably) close. Thus I do not believe that a “triggered” or cascade rupture event caused by another adjacent ruture is credible.

    • petertuft says:

      There have been at least a couple of studies of the vulnerability of pipelines to failure of an adjacent pipeline, and it seems to be generally true that a cascade failure is at worst extremely unlikely, if not non-credible as you suggest. I have seen (but can not immediately lay hands on) a semi-standard method for calculating the crater size due to pipeline rupture. If the adjacent parallel pipe is outside the crater it is immune. And even if it is inside the crater it may still be OK, as illustrated by pictures of the 2004 Belgian incident which clearly show another pipe in the crater wall, no longer with any coating but still intact.

  4. petertuft says:

    The discussion on interpretation of measurement length opened an unexpected can of worms. Rather than pursue it further here it is likely that the AS 2885 Part 1 committee will have a think about it and there may be something further coming out of that.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s