The objective of the pipeline incident database is to capture information that is relevant to the risk based approach of AS2885. Conclusions drawn from the data may be used for a variety of purposes, such as identifying key vulnerabilities, assisting focus on the most effective pipe protection measures, demonstrating to regulators and other parties that pipeline transportation is safe, and providing a basis for future revisions of AS2885.
The database contains 120 fields that provide information on the incident location, severity, causes and consequences, as well as details of the pipeline such as its dimensions, operating conditions and protective measures. Completing an incident report form is not meant to be an onerous task.
The data
The number of incidents in the database has grown dramatically since the last report provided to APIA in 2004. There is now a total of 420 entries, up from 118 previously. There are several reasons for this increase:
* Near misses are now included and occur much more frequently than damage incidents.
* Several pipeline operators have provided retrospective data.
* New Zealand data from the past couple of years is now included.
There is also substantially increased confidence that the data captures the majority of incidents, at least for recent years. Reporting is voluntary and hence vulnerable to omission.
Figure 1 shows the number of damage incidents per half-decade and the growth in pipeline length. From that the overall rate for damage incidents can be determined, expressed as incidents per 1,000 km and per year (as shown in the Figure 2).
Location, location, location
Pipeline location classification reflects an expectation that more incidents are likely in more populated areas (as well as the expectation that such incidents have the potential to be much more serious).
Current data indicates that approximately 60 per cent of incidents occur in the suburban and high density location classes. However the pattern of actual damage is quite different, with nearly 75 per cent in rural areas.
One speculative interpretation of this difference is that urban areas are the scene of a great deal of unauthorised excavation activity within pipeline easements, but the great majority of it is minor and does not result in contact with a pipeline, perhaps partly because the work is identified and stopped by frequent pipeline patrols.
Damage incidents are shown to incur a four-fold higher threat to urban pipelines compared to those in rural areas. Nevertheless, over 85 per cent of Australian pipelines are located in remote rural areas so it is not surprising that the greatest number of incidents occur there.
Causes
Figure 4 confirms that external interference is by far the dominant cause of pipeline damage. This data includes all pipeline damage incidents from both Australia and New Zealand (no near misses) but does exclude the many corrosion incidents on poorly managed pipelines as it is not representative of practice in recent decades.
Incident severity
The database classifies damage into six levels of severity: coating damage, stress corrosion cracking/corrosion (no leak), deformation, gouge, leak and rupture. Figures 5a and 5b show the distribution of severities, and the causes of those incidents which resulted in a leak or rupture.
Overall, only one in six damage incidents results in a loss of containment (leak or rupture). Of these, about 65 per cent are a result of external interference, which is interesting given that the previous graph showed external interference comprising some 85 per cent of damage incidents. One possible conclusion is that the other causes are more damaging, but it seems more likely that incidents involving those other causes simply do not come to light unless they are serious enough to result in a loss of containment.
Figure 5a is based on all damage incidents and does not tell the whole story about recent trends.
There has been a change over time in the distribution of reported damage severity, with deformation incidents almost completely replaced by coating damage and gouges. However it seems likely that this is largely due to variation in the basis for reporting, with gouges previously classed as deformation (suggested by some of the incident descriptions even where the damage is not explicitly described) and relatively minor incidents such as coating damage were possibly not reported in the past. Figure 5c shows the recent pattern of severities, and is consistent with the common-sense expectation that coating damage will be reasonably common relative to other levels of severity.
The culprits
It is of interest to understand which parties are responsible for causing pipeline incidents.
The incident database demonstrates that, on all incidents, property owners (and their contractors) are the principle culprits. However the relative contribution from property owners is much reduced when the data is restricted to damage incidents. In this case the largest source of damage is government agencies and utilities and their contractors. A plausible interpretation of this is that property owners (mostly rural) typically undertake relatively light construction at only shallow depths (fencing, surface drains, etc); no loss of containment incidents have been attributed to property owners or their contractors. On the other hand, government agencies and utilities are more likely to be involved in large construction or maintenance projects using heavy machinery and deeper excavation with a greater likelihood of impacting a pipeline.
Pipeline operators (and their contractors) are a disturbingly frequent cause (28 per cent) of pipeline damage, virtually always during excavation for pipeline maintenance. None of these cases resulted in a loss of containment, presumably because in each instance the perpetrators were well aware of the presence of the pipeline and dug with sufficient caution to cause only minor damage. Nevertheless, the frequency of damage by pipeline operators suggests that at least some operators take insufficient care. A review of pipeline excavation procedures may be beneficial, with consideration of additional physical and procedural pipe protection measures.
What equipment is used?
Information on the type of equipment involved in external interference incidents is relevant to the design of pipeline protection measures. Figure 7 shows the relative frequency of the main equipment types for all incidents (including near misses), damage incidents and incidents resulting in loss of containment.
The dominance of excavators and backhoes is evident, probably reflecting their presence in almost all construction work.
There is not much difference between the contributions of each equipment type to all incidents and actual damage, but the pattern changes for loss of containment incidents. The most interesting feature of Figure 7 is the observation that dozers and rippers are relatively minor contributors to incidents in general but cause
30 per cent of loss of containment events. This data supports the long-held suspicion that dozer/ripper equipment contacting a pipe has a high likelihood of penetrating it because its mass and stiffness mean it is less likely to be diverted from its course. For other equipment types only around 10 per cent of damage incidents result in loss of containment; that rises to nearly 40 per cent for dozers and rippers.
Drawing conclusions
External interference remains the dominant cause of pipeline incidents in Australia, and therefore should remain the focus for further improvement in safety performance. Incidents caused by corrosion or defects are insignificant in comparison.
Incidents are caused by a variety of perpetrators, among whom property owners are prominent, but government agencies and utilities bear more responsibility for actual pipe damage. An unexpected number of damage incidents are caused by pipeline operators themselves, suggesting scope for improved pipeline maintenance procedures.
External interference damage in urban location classes occurs at a rate four-fold higher than in rural areas. The main equipment involved is excavators and backhoes, but the data shows that where a dozer or ripper contacts the pipe it has a much higher likelihood of causing a leak or rupture.
By continuing with conscientious incident reporting it is expected that in future years sufficient data will be available for further aspects of pipeline incidents to be usefully analysed. It will also enable Australia to play its role in the International Gas Union project to harmonise pipeline incident data reporting around the world.
APIA intends that a regular report on pipeline incidents be produced every 3-5 years. The report will contain information generally similar to the above plus analysis of additional parameters such as effectiveness of pipe protection measures as sufficient data become available. A committee representing pipeline operators, regulators and risk consultants will determine the information to be included in the report. The presentation of data will of course be done in a way such that details of individual incidents will remain confidential.
This article is a summary of Craig Bonar and Peter Tuft’s paper entitled The Australian experience with the pipeline incident database presented at the 2009 APIA Convention.
An incident is defined as:
1. Any loss of containment (not including minor leaks at flanges). 2. Any damage to the coating or pipe caused by mechanical equipment. 3. Any other defect (eg. corrosion) which requires either MAOP reduction or pipe repair (eg. reinforcing sleeves, clock spring, or cut-out and replacement).
A near miss is any unauthorised third-party activity that does not damage the pipeline. This includes:
1. Activities covered by AS2885.3 Clause 6.5.3(b): “˜Land disturbance activities deeper than 300 mm…on the pipeline easement, or where no easement exists, a minimum of 3 m (but preferably 6 m) each side of the pipeline.’ This would include excavation, auguring and boring activities. 2. Seismic activity and use of explosives in the vicinity of the pipeline. See AS2885.3 Clauses 6.5.3(c) and 6.5.4.
Reducing risk
Even though Australian pipelines have an excellent safety record, the potential for catastrophe remains, so the industry should be seeking continuous improvement in risk reduction. The incident data helps to draw conclusions about where the greatest gains might be in improving pipeline safety.
86 per cent of pipeline incidents are due to external interference, and nearly half of those that cause damage are associated with infrastructure maintenance or construction. So, efforts to reduce incidents will be most effective if they have a focus on government authorities, utilities and contractors that build and maintain infrastructure.
Most pipeline operators are already doing pretty much all that can be expected of them under their pipeline awareness programs. Despite that, they have no direct influence on the activities of third parties. However, governments are likely to have much more influence, so there may be a case for stronger government regulation of third party excavation. The incident data helps support the argument that the greatest gains in risk reduction may come in this area.