Pipelines OZ

Location classification

I thought I might start a serious of posts on location classification, since it’s an important concept in AS 2885 and some aspects of it cause confusion.
To start with the basics, location classification is a way of categorising land uses to reflect the threats to and from the pipeline. Where there are lots of people there is not only likely to be an increased level of activity that might result in pipeline damage, but the consequences of a serious pipeline failure will probably also be greater. So in more populated areas we need more stringent protection against pipeline failure.
The primary purpose of the AS 2885 location classes is to draw attention to the greater risk level in more populated areas so that the pipeline can be designed (or managed) appropriately. The standard doesn’t actually have many fixed rules that vary between location classes. The most important set of rules are in Section 4.7, Special Provisions for High Consequence Areas. Apart from that it has some fairly common-sense requirements for things like depth of cover and sign spacing, although even those often make little distinction between R1 and R2, or between T1 and T2.
My view is that the most important function of location classification is to help focus the safety management study on the areas where the likelihood and/or consequences of a failure may be greatest. For that reason I don’t care much about getting the location classification exactly right. As long as the chosen classification will guide you to an appropriate design (or risk management measures) then it doesn’t matter whether you’ve adopted, say, R2/I or T1 for an industrial area.
Over the next few posts I hope to work my way through each of the primary and secondary location classes with some guidance material and observations on pitfalls.

T1 and T2 location classes

These urban location classes are fairly straightforward. T1 is for ordinary suburban areas and they are allowed to include small patches of higher density (up to 10%). It also includes other land uses with similar population density. T2 is for multi-storey areas or large commercial centres where there are normally many people present. I don’t think there is much difficulty with any of that.
One point worth considering is the borderline between R2 and T1. As discussed in the previous post on R2 areas, the number of people affected by a pipeline rupture depends not only on population density but also on the measurement length and hence pipeline size. So an area of 2 ha blocks around a very large pipeline might present a risk to a larger number of people than a small pipeline in a T1 area, even though strictly speaking it would fit the R2 definition (blocks of 1 – 5 ha). For this reason in some low-density urban fringe areas it might be worth extending the T1 area beyond the extent required by a black-letter interpretation of the Standard.
Another grey area concerns places that can accumulate large crowds but only infrequently or for short periods. Playing fields are a classic example. I don’t have any definitive suggestions for these cases. It can be worth considering whether the activities that might cause a pipeline failure could occur at the same time that the people are present – eg. construction work on weekdays, local footie game on the weekend, no overlap between threat and crowd.
A variation on this temporary crowd issue is busy roads in urban areas. If traffic congestion means that roads become virtual parking lots at peak hour then there is the potential to have a very large number of people concentrated along the road. Think about a three lane road with a car every 7 m. Even with only single occupants that’s a couple of hundred people who might be within the measurement length of a DN 300 pipeline.
As I’ve said before, I think actual location classification is less important than recognition of the number of people who might be affected. This is particularly important if a threat needs a full risk evaluation. Risk evaluation provides the tool to deal with the intermittent crowds because the likelihood of the failure occurring at the same time as the crowd can be included in the frequency estimate – failure with a couple of people present might have frequency F, failure next to the traffic jam might have frequency 0.25 F, failure with the crowd at the footie field might have frequency only 0.01 F (examples only, real cases of course depending on the exact circumstances).

R1 location class

The R1 (Rural) location classification in AS 2885 is pretty straightforward: areas that are either unpopulated or have only occasional isolated residences.
When the code committee was considering location classification there was some discussion of an even lower location class (informally known as R0) for truly remote outback regions. However in the end it didn’t seem worthwhile. In both R1 and “R0″ areas the most likely victim of a pipeline failure will be the person who created it by digging where they shouldn’t. In other words, the threat brings its own consequences with it regardless of the fixed population in the surroundings.
The only situation in which an “R0″ definition might be useful is the unusual case of aboveground pipelines. Section 5.8.3 of AS 2885.1 (and Fig 5.8.3) define the conditions under which high pressure pipelines can be installed without normal burial. Gas pipelines can be installed aboveground only in locations with zero population and managed access, and then only subject to a special safety management study. But that is such an rare situation that it wasn’t worth defining a whole new location class to cover it.
A common question about R1 areas is how to treat isolated houses that unavoidably fall within the measurement length. Do they need extra protection? That depends on the safety management study. Clause 4.11.2 of AS 2885.1 says that “localised increased protection against external interference should be provided, including increased penetration resistance where appropriate.” One common practice is to provide increased depth of cover (say 1200 mm) where the pipeline is within the 4.7 kW/m2 distance from a house and also to increase wall thickness when the pipeline is within the 12.6 kW/m2 distance from a house. But there isn’t much point increasing the thickness further if the pipe already has reasonable penetration resistance.
For existing pipelines undergoing safety management study review there is not much that can be done about isolated houses. About the only suggestion I’ve seen, which is a little half-baked, is that in-line inspection data should be given closer scrutiny for defects in the vicinity of isolated residences.
Anyway, R1 areas don’t raise many issues. It’s the other location classes that can get slightly more complicated.

R2 location class

Location class R2 (rural residential) has potential to be mis-applied, although I suspect that many people don’t realise it. Let me illustrate with a comparison of two very different pipelines, both with MAOP 10.2 MPa: DN 100 (common supply to small towns) and DN 1050 (now being designed en masse in Queensland). The “measurement length” is about 100 m and about 1000 m respectively.
The key part of the R2 definition is the reference to “single residence blocks typically in the range 1 ha to 5 ha”. If the blocks are square then they would be have sides of 100 m to 225 m. Our DN 100 pipeline though such an environment would at worst affect one or two houses if it was to suffer a full bore rupture, and may well affect none. You would be pretty unlucky to have any injuries or fatalities among the residents unless the pipeline happened to be very close to a house at the point of failure.
On the other hand the DN 1050 pipeline could conceivably have a couple of hundred houses within the measurement length (if the blocks are the minimum 1 ha), or more realistically might have a few dozen houses if the blocks are 5 ha. Either way the consequences of a full bore rupture would be much more serious than failure of the little DN 100 pipeline in a T1 (suburban area); multiple fatalities and many injuries seem possible.
To me this looks inconsistent. The examples are at the extreme ends of the range but serve to illustrate the point.
It’s also worth thinking about the note at the end of the R2 definition (AS 2885.1 Clause 4.3.4(b)):
NOTE: In Rural Residential societal risk (the risk of multiple fatalities associated with a loss of containment) is not a dominant design consideration.
This reinforces my view that it doesn’t seem right to apply R2 to a situation where pipeline failure can result in multiple fatalities.
For a couple of projects I’ve suggested that the R2 definition be interpreted conservatively so that R2 is applied anywhere there are more than a dozen or so houses within the measurement length, regardless of lot size. So for the DN 1050 pipeline the lot size within the preferred R2 area may be substantially larger than 5 ha. Conversely, for small pipelines (our DN 100 example) one could quite defensibly decide that R2 doesn’t apply for lot sizes larger than say 1 ha, and even that is quite conservative.
I’m not suggesting that the 1-5 ha guideline is redundant, but rather than it be applied with some thought about how many houses might actually be affected by a pipeline failure.
Having said all that, the design rules for R2 are not very different to those for R1, yet again reinforcing the view that R2 areas are not expected to have potential for really serious failure consequences. I think the main value in assigning R2 location class (or any location class for that matter) is in helping the participants in the safety management study appreciate the general magnitude of consequences if the worst should happen.

Secondary location classes

(I’m back. Things still seem to be incredibly busy but I’m hoping to find enough time to resume reasonably regular posts here.)
I still haven’t finished the series of posts on location classification so it seems appropriate to resume with a brief overview of the secondary location classes. To establish the basics:
Every point on a pipeline must have a primary location class (R1, R2, T1 or T2). But there are some land uses and features that don’t fit neatly into those primary classes. So AS 2885 has provision for secondary location classes that draw attention to these special locations.
Secondary location classes are optional – they only need to apply where they are relevant. As noted previously, the main purpose of location classification is to help the safety management study identify and focus on areas of higher risk and in some cases to require increased risk reduction measures. That applies equally to primary and secondary location classes. If a secondary location class requires additional risk reduction then that overrides the requirements of the primary location class.
There are five secondary location classes (see Clause 4.3.5 of AS 2885.1-2007): ▪ Sensitive (S) ▪ Industrial (I) ▪ Heavy industrial (HI) ▪ Common infrastructure corridor (CIC) ▪ Submerged (W)
That should all be pretty straightforward. In subsequent posts I’ll expand a bit of the first four secondary classes. There isn’t much to say about the W class because it doesn’t impose any particular requirements, other than the obvious need to ensure that the design is appropriate for the submerged (or equivalent) installation conditions.

Industrial location classes (I and HI)

The two industrial location classes are provided to allow recognition of land uses that don’t fit neatly into the four primary location classes. Unfortunately it seems that the definitions of Industrial (I) and Heavy Industrial (HI) are not always well understood, perhaps because the definitions in the Standard could be refined a bit.
Industrial (I) should be simple – it’s meant to cover urban industrial areas where there are multiple industries set in an ordinary streetscape. This is usually (but not always) light industry – warehouses, small factories, car yards, industry-related retail outlets, etc. Usually there will be quite a few offices as well as take-away food outlets scattered through the area if it is large. The area is generally accessible to the public. This image of Laverton North in Melbourne is typical of a large Industrial area, but they can be smaller too.
Industrial areas are reasonably well populated by workers and the public, and they are accessible to all the service providers who might pose a threat to pipelines by digging in the streets for maintenance or new installation of their utilities. For these reasons AS 2885 requires that they be treated the same as T1 (Residential) areas.
Heavy Industrial (HI) areas are different. These tend to be locations where there is only a single large industrial plant (or perhaps a few at most). As implied by the description in the Standard these areas may have quite diverse characteristics. Examples might include the land near a waste dump, a mine, a major mineral or coal processing facility, a process plant such as a refinery or a petrochemical facility, or a tank farm. I also once encountered a plant that manufactured aluminium power (highly flammable).
Because HI areas may be diverse there is no simple rule for them – they may need to be treated as R2, T1 or T2 depending on circumstances. The point of this location classification is to highlight that there is an unusual land use and make you think about it. The Standard highlights that HI areas may contain either unusual threats, or present unusual consequences in the event of a failure, the latter meaning escalation such as might occur if there is a major pipeline failure next to a tank farm and the tanks catch fire too.
To clarify this last point, I’ve seen confusion over whether a small store of drums containing flammable fluids would require a classification of HI and be treated as T2 because of the potential for escalation. In that particular case the fire from an ignited pipeline rupture would have been much bigger than any possible fire from the drums, so the escalation would have been insignificant. On the other hand another couple of kilometres along the same pipeline was a large refinery tank farm, and that was definitely a case of potential escalation.
At the other end of the scale mineral processing plants or waste dumps would only need to treated as if they are R2, and even that might be conservative in some cases.
As I’ve said before, getting the location classification exactly right isn’t critical because it’s only an intermediate stage in getting the design and risk assessment right. Sometimes there might be more than one way of classifying a particular location, which doesn’t matter as long as the design and safety management study identify all the relevant threats and treat them and the possible consequences properly.

Common Infrastructure Corridor (CIC)

Last (?) in this intermittent series of posts on location classification – Common Infrastructure Corridor.
The CIC location class highlights the special set of threats, and corresponding special mitigation measures, that arise when a pipeline is installed closely parallel to other infrastructure. That other infrastructure might include other pipelines, power lines, roads, railways or even a long conveyor belt. It doesn’t much matter whether the other stuff is in the same easement. The key point is that it is parallel and close enough that there is potential for interaction with the pipeline. ”Close enough” is likely to vary depending on the nature of the other infrastructure.
For example, a pipeline within a road reserve definitely requires CIC classification, but a pipeline parallel to a road and outside the road reserve probably doesn’t because neither the road authority nor anyone else who builds services in roads is permitted to touch the adjoining private land. But few rules are universal and if there is evidence that the road does have extended influence (eg. cut-off drains that extend far into the adjoining land) then CIC might apply after all.
Assigning CIC location class has two main implications: ▪ Agreements addressing mutual protection of all the parallel infrastructure should be negotiated with the other parties if possible ( AS 2885.1-2007 Clauses 5.5.4(e) and 5.5.6(d) ) ▪ Consideration should be given to the possibly increased consequences of pipeline failure should the other infrastructure be affected (eg. a busy road, other pipelines, etc)
CIC does not apply where a pipeline crosses other infrastructure. The key feature of parallel infrastructure is that the overall exposure to increased threats and consequences is increased by the extent of close association between the pipeline and other stuff.
I’ll repeat (for the last time) that it’s not worth getting obsessed about the details of location classification, as long as it works to achieve its objective of highlighting features of the pipeline surroundings so they get appropriate treatment in the design and safety management study.

S (Sensitive) location class

The Sensitive (S) location class doesn’t seem to cause much confusion (with one minor exception).
Its intent is obvious: to identify locations where there are people who are less able to protect themselves against a pipeline failure. This includes particularly pre-schools, schools, hospitals, aged care facilities, etc. Prisons are are another instance – the inmates may be able-bodied but are nevertheless limited in their movement for other reasons. (But every case has to be considered on its merits – I know of an old 19th century gaol whose massive 6 m high stone walls provide superb radiation protection for anyone inside!)
AS 2885 requires that Sensitive areas be treated the same as T2 (High Density) areas. So if your rural pipeline passes by a little isolated country school then the section of pipeline within the measurement length from the school must be designed the same as if it is in a densely populated urban area.
Nothing complicated about that I think.
However confusion sometimes arises because of this statement at the end of the description of Sensitive location class in AS 2885:
It shall also include locations of high environmental sensitivity.
The key to correct interpretation of this is that the safety management study process in AS 2885 is about failure of pipelines while in service. It has nothing whatsoever to do with construction impacts – they are important, but are addressed through different processes. A better form of words might have been:
It shall also include locations of high environmental sensitivity to pipeline failure.
With that clarification, it should be apparent that gas pipelines will rarely if ever receive an S classification for environmental reasons (I’ve certainly never encountered one). However it may be entirely reasonable for an oil pipeline to have a number of Sensitive locations, depending on where spilled oil may spread and the environmental damage that could result.

Location classification for facilities

(Been busy lately – rather shocked that it’s nearly four weeks since my last post.)
I had a question the other day of a type that comes up from time to time: What location classification do you assign to pipeline facilities? It’s a question based on a misunderstanding of location classification principles, but an easily forgiven misunderstanding because AS 2885 is silent on the matter.
The basics of location classification were outlined here, where I said “The primary purpose of the AS 2885 location classes is to draw attention to the greater risk level in more populated areas so that the pipeline can be designed (or managed) appropriately.” In higher location classes there tend to be both more threats to the pipeline (because of the higher level of human activity) and more serious consequences of pipeline failure (because there are more people around to be hurt).
How do you apply that principle to a pipeline facility? The short answer, in my view, is that you don’t. Most pipeline facilities are neither an origin of threats to the pipeline nor part of the consequences of a pipeline failure. Rather, they are part of the source of the hazard (ie. the pipeline). So my starting point in responding to the question of location classification for pipeline facilities is that is often a meaningless question.
Having said that, there are always exceptions. A pipeline site that is normally manned by a few people might be treated in the same way as an isolated house in a rural area because the population density is similar. That doesn’t need a change in location class but might require some additional physical protection of the nearby pipeline (eg. extra wall thickness and depth of cover, which is common practice anyway).
Another argument that might be applied is where the facility is a large compressor station complex. If a pipeline failure at some remote location causes serious supply disruption for a few days pending repair, that’s bad. If a pipeline failure immediately adjacent to a critical compressor station disables the station and cuts off supply to a city for months until the station can be reconstructed, that’s extra bad. I could be persuaded that in this situation the compressor station vicinity should be classified as Heavy Industrial (HI) on the basis that a pipeline failure can have serious flow-on effects (or escalation, using the terminology from AS 2885).
So far this discussion has been about sites that are dedicated to the pipeline. The situation might be different if the facility is, say, a gas processing plant operated by another party, or a power station to which the pipeline is delivering gas, or a refinery to which the pipeline is delivering oil. Those sites are different partly because they are outside the control of the pipeline operator (usually), and partly because a pipeline failure may result in escalation similar to the compressor station example above.
A final comment: No pipeline facility, and no adjacent plant such as a gas processing site, power station or refinery, should give rise to the Industrial (I) location class. As discussed in this post the Industrial location class is for light industry set in a typical urban streetscape, usually with a lot of people around. It has essentially the same population density and level of public access as T1 (Suburban) except the buildings are industrial/commercial/retail rather than residences. It would clearly be wrong to assign this classification to the surroundings of a compressor station, gas plant, power station or refinery unless it is one of the few older facilities that really are located in a heavily built-up urban area (and I can think of just one of those).
The location classification system has its limits. With only four primarily location classes and five secondary classes there will be situations that don’t fit nearly into any of them. As I’ve said repeatedly, the point is to highlight features of the pipeline surroundings that can affect the types of threat and the consequences of failure. If you focus on those things you will get the design right no matter what location classification is assigned.