Towards a causal model from pipeline incident data analysis

https://doi.org/10.1016/j.psep.2020.06.047Get rights and content

Abstract

Pipeline network facilitates transportation of hazardous materials over long distances and is considered a relatively safer mode of transportation. However, a significant number of incidents occurred over the last few years that has increased the need for development of an accurate predictive model for incidents. Pipeline incident databases have been collecting a wide range of data regarding the incidents including the background operating conditions, the causes contributed to the incident, and the severity of consequences. Most academic studies analyzing these databases have developed predictive failure models focusing only on a narrow range of factors that played a role behind certain categories of incidents. They fail to provide a holistic understanding of how interaction of multiple factors contributing to an incident or how their combined effect eventually leads to failure. The amount of data collected in the databases are however huge. For example, the current reporting format of Pipeline and Hazardous Materials Safety Administration (PHMSA) collects over six hundred datapoints from each incident. Given the large amount of information available and current technologies, it is now possible to look into ways to understand the holistic effect of all contributing factors. The current study looks into how the databases can be utilized to derive such a complete understanding. The study compares the databases of US PHMSA, Canada National Energy Board (NEB), and European Gas Pipeline Incident Data Group (EGIG) to examine the frameworks used to classify the causal factors of the incident data and shows that the background factors, underlying factors and the causal factors identified in the various databases are inter-related and show varying degrees of dependency. This curves the way forward to development of a causal model that will help identify the important factors that needs to be addressed and a prediction of future failure given existing conditions of a pipeline system.

Introduction

Worldwide energy demand grew by 2.3 % in 2018 with natural gas emerging as the fuel of choice, showing the biggest gains and accounting for 45 % of the rise in energy consumption (The International Energy Agency, 2019). A worldwide wave of pipeline construction activity has been driven by this continuing global shift towards natural gas. In particular, the United States led the global increase for the first time in 20 years; the rapid growth of shale production, the lifting of an oil export ban, and the predicted growth in global LNG demand led to a massive infrastructure development, including oil and LNG export terminals, and the pipeline capacity to supply them (Awalt, 2019). The U.S. pipeline network which consists of two-third of world’s pipeline mileage, transports almost all the natural gas produced and used in the country, as well as over 90 % of crude oil and refined petroleum products. There has been a 44 % increase in transportation through pipelines in the last five years (Allison and Mandler, 2018; American Petroleum Institute and Associattion of Oil Pipelines, 2019; Central Intelligence Agency, 2019; Wikipedia, 2019a). The US pipeline network however is no different from the pipelines of the rest the world when it comes to pipeline incidents. Fig. 1 shows a few well-known pipeline incidents from across the world since 2010 (Abraham et al., 2019; Wikipedia, 2019b), including data of lives the claimed and the spillage they caused. It includes some major US pipeline incidents such as Kalamazoo River pipeline leak spilling 20,000 barrels of oil (National Transportation Safety Board, 2012), San Bruno pipeline explosion killing 8 people (National Transportation Safety Board, 2011), and Keystone pipeline spillage (Ramírez-Camacho, Carbone, Pastor, Bubbico, & Casal) releasing 5000 barrels of crude oil (National Transportation Safety Board, 2018). The occurrence of such major incidents at frequent intervals worldwide indicate that the challenges involving safe oil and gas transportation via pipelines are still magnanimous.

The Pipeline and Hazardous Materials Safety Administration (Pipeline and Hazardous Materials Safety Administration), which is responsible for safe operation of the pipeline network in the United States, publishes pipeline incident data regularly on their website (Pipeline and Hazardous Materials Safety Administration, 2019a). A set of safety performance parameters used by PHMSA is summarized in Table 1 and the corresponding trend lines are shown in Fig. 2. These parameters (number of incidents, significant incident, fatality, injury, asset damage, and spillage) are considered lagging indicators, which is “a retrospective set of metrics that are based on incidents that meet the threshold of severity that should be reported as part of the industry-wide process safety metric” (Center for Chemical Process Safety, 2011). The variation of parameter values especially number of injuries and spillage amount appears random as there can be multiple reasons behind it. They are used in calculating failure rates and might be useful to predict a future incident; but they do not provide much insight on how to improve the current performance or reduce the chances of future incidents. However, the sheer number of incidents, their trends and causes identified indicate that there exist ample opportunities for improvement. In the past, this has drawn a lot of researchers to this domain of pipeline research.

A brief summary of literature on cause analysis of pipeline incidents is shown in Table 2. The table contains the causal factors that the studies considered, background factors that were identified as having some association with the causal factors, and the source of data where the causes were reported (such as US DOT, CONCAWE, EGIG, and UKOPA). Here, the causal factor is defined as “a major unplanned, unintended contributor to an incident (a negative event or undesirable condition), that if eliminated would have either prevented the occurrence of the incident, or reduced its severity or frequency” (Center for Chemical Process Safety, 2019). Thus, causal factors are those that are directly responsible for causing an incident and these factors may act alone or with other factors to cause the incident or worsen the impact. The studies reported corrosion, mechanical failure, third-party damage, outside force etc. as causal factors. A causal factor like corrosion may contribute to an incident along, for example by causing pipe damage leading to a leakage, or it may weaken the pipeline and with slight third-party damage, may together cause a pipeline to break. A few studies have also linked other parameters such as commodity transported, pipeline location, diameter, wall thickness, installation year, depth of cover with the incidents. These parameters are defined as background factor. A background factor can be defined as one that seems to influence a causal factor under certain condition or value more than it does under other conditions, thus increasing the likelihood of failure due to that causal factor. They are generally inherent characteristics of the pipeline and transportation and do not have any direct influence on pipeline failure. For instance, pipelines with certain diameters seem to fail more than those with larger or smaller diameters and corrosion seems to be a leading causal factor behind such failures. Here, corrosion is considered as the causal factor with diameter of the pipeline being the background factor. It may be that the mechanism by which the background factors influence the causal factors is not yet perceived, and subsequently, why they influence the causal factors is not yet well understood. There is a lack of studies linking causal factors to the management system failures. Typically management, design, planning, organizational or operational failings are identified as root cause or underlying cause (National Energy Board, 2019; Occupational Safety and Health Administration, 2015), which is defined as “a fundamental, underlying, system-related reason why an incident occurred that identifies one or more correctable system failures” (Center for Chemical Process Safety, 2019). In this article, the term underlying cause has been used subsequently. They represent the system’s performance, and have a direct influence on a causal factor, but do not directly cause a pipeline failure. Similar studies in oil and gas production in offshore (Halim et al., 2018), onshore (Yu et al., 2017), and hazardous material transport (Quddus et al., 2018) incidents show a strong link between causal factors with underlying causes, as the negative events and undesirable conditions involve some of the active and latent failures that contributed to the incident.

The objective of the present study is set to understand how pipeline incident data was analyzed in the literature and identify the scope of improvement of the analysis. The study examines the frameworks used to classify the causal factors of the incident data from various databases and reports and found that the difference is minimal. Distribution of the causal factors and corresponding failure rates are compared for the datasets considered. Since one dataset allowed reporting of multiple causal factor for a single incident, it is studied further to understand its importance. Association between causal factors and a few selected background factors are investigated to identify the dependence of the factors. Influence of underlying factors on causal factors and their interdependence are also studied. Finally, the relative importance the different factors collected by the organizations and the limitations of the current analysis techniques are discussed. Associativity and causality of various types of factors and causes established to the pipeline incident are discussed. It is concluded that without a proper causal model, the understanding of the pipeline failure is partial and flawed.

Section snippets

Pipeline incident data

The current analysis investigates pipeline incident data from three data sources originating from three regions: US PHMSA, Canada National Energy Board (NEB), and European Gas Pipeline Incident Data Group (EGIG).

US PHMSA maintains four separate incident databases for hazardous liquid (HL), natural gas transmission and gathering (GTG), natural gas distribution (Wu, Zhou, Xu, & Wu), and liquified natural gas (LNG). PHMSA pipeline infrastructure has 347,020 km (215,628 miles) of pipeline for crude

Comparison of causal factors

PHMSA classifies causal factors into 7 categories (1. corrosion, 2. natural force damage, 3. excavation damage, 4. other outside force damage, 5. material/ weld/ equipment failure, 6. incorrect operation, and 7. all other causes). With mapped sub-causes, cause-classification forms a causal-tree with more detailed information. For instance, the mapped cause “Corrosion” is sub-divided into “Internal corrosion” and “External Corrosion" as Sub-cause. Another level of information may also be

Background factors in PHMSA database

In the incident report form, PHMSA collects information about location, facility, operating conditions, and consequence as well as apparent cause of failure. The data includes condition of pipeline network at the time of the incident and it does not change with the outcome of the incident. Some of the parameters can be associated with the causal factors shown in Table 2, such as, commodity transported, year of installation i.e., age of the pipeline network, pipeline diameter, pipeline

Distribution of underlying causes

Unlike PHMSA and EGIG, NEB collects underlying causes (“why it happened”) in addition to causal factors (“what happened”). Some organizational or management system elements are identified as underlying causes that may contribute to any causal factor. Distribution of underlying causes on how they affect the incidents are plotted in Fig. 5. The NEB allows identification of multiple causal factors in an incident. It also allows multiple underlying causes to be identified for each incident but does

Discussion

Pipelines have been operating as a moderately safe mode of transportation of petroleum and its derivatives compared with other modes of transport (Green and Jackson, 2015). Incidents of pipeline failure, however, keep on happening. With increasing transportation volume and pipeline mileage, it is becoming increasingly vital to get to the root of the problem, to understand and mitigate them in order to minimize the losses as much and as early as possible.

The past studies on pipeline causation

Conclusion

The current work analyzes three databases set on incident data collected from pipeline operations from three different regions: PHMSA from the USA, NEB from Canada and EGIG from European Union. All databases provide a large amount of information related to an even larger number of pipeline incidents. A review of the type of information gathered is provided and some analysis of data pertaining to the causal factors behind the failures are provided. PHMSA database also provides information on

Declaration of Competing Interest

There have been no conflict of interest of the authors regarding the research area or analyzed results.

Acknowledgement

This work is supported by the United States Pipeline and Hazardous Materials Safety Administration (PHMSA) Award No. 693JK31850011CAAP.

References (42)

  • H. Wang et al.

    Likelihood, causes, and consequences of focused leakage and rupture of US natural gas transmission pipelines

    J. Loss Prev. Process Ind.

    (2014)
  • J. Wu et al.

    Probabilistic analysis of natural gas pipeline network accident based on Bayesian network

    J. Loss Prev. Process Ind.

    (2017)
  • M. Yu et al.

    Development of a safety management system (SMS) for drilling and servicing operations within OSHA jurisdiction area of Texas

    J. Loss Prev. Process Ind.

    (2017)
  • W. Abraham et al.

    Large Pipeline Explosion and Oil Spill Accidents Around the World

    (2019)
  • E. Allison et al.

    Transportation of Oil, Gas, and Refined Products

    (2018)
  • American Petroleum Institute et al.

    Pipeline Safety Excellence Performance 2019 Annual Liquids Report

    (2019)
  • T. Andersen et al.

    Pipeline reliability: an investigation of pipeline failure characteristics and analysis of pipeline failure rates for submarine and cross-country pipelines

    J. Pet. Technol.

    (1983)
  • J. Awalt

    P&GJ’s global pipeline construction outlook

    Pipeline and Gas Journal

    (2019)
  • C. Bersani et al.

    Accident Occurrance Evaluation in the Pipeline Transport of Dangerous Goods

    (2010)
  • Canada Energy Regulator

    Incident Data

    (2019)
  • Center for Chemical Process Safety

    Process Safety Leading and Lagging Metrics

    (2011)
  • Cited by (39)

    View all citing articles on Scopus
    View full text