Fracture control

I realised recently that my understanding of pipeline fracture control had become out of date; in fact it was probably never up to date.  Fracture control is a difficult topic.  The main concepts are deceptively simple but as soon as you get into the detail it becomes complex and potentially confusing, and many pipelines require the level of detail that opens up that scope for confusion.  The situation is not helped by the fact that a proper fracture control plan is required only infrequently (how many new pipelines do you design every day?) so it is not a routine activity.

I’m pretty sure I’m not alone in struggling to get my mind around a lot of this stuff.  In fact I have feeling that very few people indeed have a confident understanding but few want to admit that they struggle with it (including me).  So it seems a suitable topic for a series of posts.  I’m taking a risk of exposing myself as lacking expertise here, but I’m sure someone will correct me if I get anything seriously wrong and perhaps we will all get a better grasp of this infuriating subject.

I’ll start with the real basics, perhaps well-known to almost everyone but worth setting out anyway.  There is not much below that is not in AS 2885.1, but expressed somewhat differently.

Steel is usually ductile, but not always.  I was alerted to that at an early age when my father (who was a naval architect) talked about the WWII liberty ships, mass-produced cargo carriers that suffered brittle fractures in the cold North Atlantic.  Wikipedia says about 2700 of these ships were built and 1500 suffered cracking problems including 12 that snapped in half without warning.  So the first step in preventing fracture is to ensure that the steel remains ductile at operating temperature, but there is a lot more to it than that.

If a pipeline suffers a some insult that causes a through-wall defect (hole or crack) there may be three outcomes depending on the length of the defect, the operating conditions and the properties of the pipe:

  • Leak, when the defect remains stable in size.
  • Rupture with arrest, which occurs when the defect is too large for the remaining steel to resist the hoop stress and the pipe peels open for its full diameter but the fracture does not propagate very far.
  • Propagating fracture, when the rupture is not arrested but the crack continues to run for a considerable distance, up to many kilometres in the worst cases; propagating fractures may be either brittle or ductile.

(I have already glossed over through-wall versus part-wall defects but for the time being let’s keep things simple.)

The objectives of fracture control for pipelines include:

  • Maximise the likelihood that a defect will remain stable and not progress to a rupture; this is achieved by providing adequate toughness in both the pipe body and the longitudinal weld (fracture initiation control).
  • Ensure that brittle fracture can never occur by providing sufficient toughness in the steel at the minimum temperature the pipe will encounter (brittle fracture control); the drop weight tear test is used to assess whether fracture at the minimum design temperature is brittle or tearing (ductile).
  • Ensure that a ductile fracture will be arrested rather than propagating, again by providing sufficient toughness (tearing fracture control); toughness for tearing fracture control is assessed by the energy absorbed in the Charpy impact test.

In future posts I hope to address some of these areas in more detail.

This entry was posted in Pipeline design. Bookmark the permalink.

6 Responses to Fracture control

  1. Chris Hughes says:

    I have to agree that the lack of knowledge on this subject amongst pipeline engineers has to be of concern within the industry, particularly with regard to the subject of brittle v. ductile failure. This has been well researched with many PRC and PRCI papers on the subject over the years and yet it still does not seem to be well understood.
    In particular the current linepipe specifications being issued for pipeline projects are generally lacking in detail regarding low temperature testing. Linepipe specs will call for Charpy or DWTT testing to be carried out at minimum design temperature or 5-10 degrees below minimum: this confirms that the pipe will remain ductile at that temperature but gives no information as to its behaviour at the lower temperatures which may occur during special events such as blowdown or rpressurising. Consequently unecessary restrictions are placed on such activities.

    It adds very little to the linepipe cost to specify a series of DWTT tests to be carried out on a selection of heats at a range of temperatures down to -70C or -80C: these tests will show whether the pipe transitions from ductile failure to brittle failure within that temperature range and can consequently be used when drafting procedures for venting and pressurising. A few years ago we were asked to look at such parameters for a Victorian pipeline built decades ago: when we looked at the mill certificates for the pipe we discovered that Charpy tests had been carried out that indicated ductile failure mode down to -70C. How many engineers nowadays put this detail into their linepipe specs?

    Linepipe and pipe steel manufacturers will not guarante to supply pipe with a transition temperature below -10C or so: however they are perfectly happy to provide the data which tells you where the transition temperature is and it will certainly be well below the minimum design temperature and may well be below the lowest transient temperature that could occur. IT PAYS TO KNOW!!!

  2. Chris Hughes says:

    By the way Peter, fracture control plans are not just for new pipelines: just about every existing pipeline needs one as well. All pipelines must have an SMS review every 5 years: an SMS cannot be performed without some knowledge of puncture resistance and pipe rupture characteristics, and rupture characteristics are what the fracture control plan determines.

    • petertuft says:

      True that most existing pipelines will sooner or later need a fracture control plan, even if only at the design life review stage, but in-depth preparation of a fracture control plan remains very much a non-routine activity for the vast majority of pipeline engineers and that is the point I was making.

      • Chris Hughes says:

        Unfortunately you may be right – which is of great concern as I don’t know how anyone can be considered a ‘pipeline engineer’ without a basic understanding of fracture mechanics and the AS2885 Fracture Control Plan.

  3. Chris Hughes says:

    And a third point from me – you state in your third dot point at the end of your original post that “toughness for tearing fracture control is assessed by the energy absorbed in the Charpy impact test.” Unfortunately this is a true statement even though there have been many research papers published since as early as 1974 indicating that for modern high-strength steels the correlation between Charpy values and crack arrest is erratic, and more sophisticated tests such as crack tip opening angle (CTOA) give predictions which are far closer to empirical results. However the Charpy (and DWTT) test is cheap and cheerful and well understood and so we have not pushed for anything better to become industry practice.
    AS2885.1 recognises this to some extent by
    a) requiring the calculated Charpy value to be multiplied by 1.4 for X80 pipe and
    b) requiring independent expert verification of the fracture control requirements if the calculated Charpy requirement exceeds 100 joules.
    This goes part-way to solving the problem but, in my opinion, not really far enough.

    I have other concerns regarding the use of the Batelle Charpy equation which I have taken up with other senior industry figures and which I don’t intend to reiterate here: suffice it to say that I think we may be lucky in not having had a pipeline suffer a running fracture in this country when the design to the standard indicated that fracture should have been arrested.

  4. petertuft says:

    I might leave a detailed technical response to others who are closer to the minutiae of fracture mechanics. However my understanding is the the Battelle two-curve model (which I might describe in more detail in a later post) has been quite well validated for the range of pipelines commonly constructed overseas. The major uncertainty that occupies the minds of Australian researchers is that the validation has not been extended to the smaller diameters and higher pressures that we typically build in this country, and that uncertainty is mainly to do with the effects of friction on the decompression velocity rather than crack propagation. There is very active research into refining both sides of the model within the Energy Pipelines CRC.

Leave a Reply

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

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

Google+ photo

You are commenting using your Google+ 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 )

Connecting to %s