A matrix-based risk assessment approach for addressing linear hazards such as pipelines

https://doi.org/10.1016/j.jlp.2005.10.005Get rights and content

Abstract

Pipelines represent a linear risk source that can create unique challenges when assessing risks. In the past, risk has been managed by identifying construction requirements and setbacks based on population densities and types of land use. In the current risk assessment a matrix-based approach has been developed so as to determine the risks associated with high-vapor pressure liquids pipelines. The approach involved the development of a matrix representing each 100 m section of the reviewed pipeline along with approximately 30 risk factors that describe that section of the pipeline. Further, a receptor matrix was constructed to account for each hectare of land within 1 km of the reviewed pipeline system. This approach has allowed for the determination of risk as a function of location and separation from the pipeline and in turn has allowed for the determination of those areas where peak risks exist. In addition, this approach has ensured that the linear geometry related to pipeline risks has been accurately modeled. The resulting estimated risks have been evaluated against MIACC risk thresholds (geographic risk-based measures) and against proprietary internal corporate standards (societal risk-based measures). In this way the acceptability of the risk from the perspective of both the potentially impacted community and that of the pipeline operator can be measured. The net result is that the company has a clear picture of the risks associated with its pipeline and is better able to optimize its risk management and pipeline integrity programs.

Introduction

The Major Industrial Accident Council of Canada (MIACC, 1997) document “Risk Assessment Guidelines for Municipalities and IndustriesAn Initial Screening Tool” gives guidance as to reasonable risk thresholds for varying land uses. Further, the screening tool gives guidance as to how to estimate risk levels based on fixed facilities while stating:

Obtaining individual risk contours for a transport corridor is not straightforward. The issue arises though the ‘per unit length’ frequency units that are used in transportation problems.

A second MIACC document “Land Use Planning With Respect To PipelinesA Guideline for Local Authorities, Developers and Pipeline Operators” (MIACC, 1998) also avoids giving an approach as to how to address the issue of risk calculation with respect to linear risk sources such as pipelines. The document does however suggest that consultation should take place between the various parties when new development occurs based on a separation distance of 200 m. The document then goes on to suggest that this consultation distance should be expanded for high-vapor pressure liquids or sour gas pipelines. The goal of this consultation process is to help local authorities in establishing appropriate setbacks, which are not too restrictive or too conservative.

In the absence of clear guidance as to how to estimate pipeline risk organizations such as the Alberta Industrial Heartland Association (2003) have developed their own risk assessment approaches. These types of approaches suffer based on limited data and do not readily reflect the operation of the reviewed pipeline systems, as the operating companies are not involved in the assessment process. As a result these types of risk assessments tend to be overly conservative and have the potential to misinform the public regarding the potential risk levels. This potential overstatement of risks can be of particular concern when pipeline risks are compared to risks related to fixed facilities where more straightforward risk assessment approaches exist.

The lack of accessible guidance with respect to conducting pipeline risk assessments creates a situation where it is critical that pipeline operators readily assess their pipeline systems in a clear manner and be in a position to share this information when needed as per the approach suggested within the MIACC guidelines. In this way local governments can be provided with appropriate information for land use planning purposes allowing for informed and balanced decision making.

Risk can be defined as the product of the likelihood of an event and its consequence. Within the current risk assessment a collection of matrices has been defined in order to allow this relatively simple risk calculation to be computed numerous times so as to give a measure of risk at both a system level and across the system. The various matrices have been defined based on geography whereby 100 m lengths or 1 ha partitions (100 m×100 m) have been utilized. 100 m was selected as it represents a distance where factors associated with data quality, computational needs, and geometry can be best balanced. At shorter distances data quality and computational needs become an issue. At larger distances errors associated with geometry assumptions can become significant. In addition, assumptions as to the uniformity of the risk levels within the measured cell partitions can also become an issue for larger distances.

Through working with and defining the various matrices key information related to the pipeline system can be readily extracted from the risk assessment:

  • Failure rate as a function of location (leak and rupture rates were calculated for each 100 m of the pipeline based on 9 different failure modes).

  • Distribution of consequences following a release for each hectare of land within 1 km of the point of failure (multiple scenarios were considered and accounted for day and night conditions, varying wind directions, explosive yields…).

  • Geographic risk estimates for each hectare of land within 1 km of the pipeline (this data was also utilized to give risk contours relative to the pipeline).

  • Societal risk measures for each hectare of land within 1 km of the pipeline.

The net result is that the company can readily view the risk associated with its pipeline systems from a number of perspectives and is positioned to address those areas representing greatest concern. Further, through an iterative process the pipeline operator can review various risk reduction strategies so as to determine which activities yield the greatest value by balancing cost against the effectiveness of each strategy.

Section snippets

Failure rate determination

As was stated above, failure rates were calculated for each 100 m section of the reviewed pipeline with this calculation being based on the following failure models:

  • Internal corrosion

  • External corrosion

  • Stress corrosion cracking

  • Earth movement

  • Over-pressure

  • Valve/fitting failure

  • Construction/material defect

  • Third party damage

  • Other/miscellaneous failures

With the exception of the stress corrosion-cracking component the various models were developed based on Alberta Energy and Utility Board (EUB) (1998)

Consequence analysis

Data held by the US Department of Transportation—Office of Pipeline Safety (US DOT—OPS) (2003) related to natural gas liquids pipeline releases were reviewed allowing for a determination of the ignition probabilities and explosion probabilities (delayed ignition) as a function of release size. The resulting ignition probabilities were then modified based on the properties of the potentially released materials relative to the “average” natural gas liquids released and for land use types. For

Geographic risk

One common treatment utilized in conjunction with the MIACC methodology is to estimate pipeline risk as per the methodology given for a fixed facility. The resulting distances to the various risk contours estimated in this manner are then applied perpendicular to the pipeline to create risk corridors aligned with the pipeline. Typically people equate the MIACC failure rates with 1 km of pipeline and as such this approach results in situations where the distance to the risk contours is

Modeled outcomes

The MIACC guidelines strive to measure risk acceptability from the perspective of the potentially impacted communities. However, this risk measure does not indicate if an activity is acceptable from the perspective of the operating company. Both of these perspectives need to be reviewed in order for a risk assessment to be considered complete. In this particular case the company utilized a risk scoring methodology that assigns values to various outcomes across several categories when judging

Societal risk

Societal risk based measures, viewed from the perspective of a corporation, attempt to gauge the impact the surrounding community will have on the corporation as a result of the potential impact the corporation has on the community. The non-linear relationship between fatality number and risk value reflects the greater outrage a community will express as related to multiple fatality events compared to individual fatality events. This greater outrage has the potential for greater regulatory

Conclusions

Due to the linear nature of pipelines most risk assessment techniques fail to generate risk values that can be readily compared to fixed facilities. This difficulty in analysis combined with limited agreement on how best to conduct these types of assessments has resulted in a situation where various levels of government are being hampered with respect to land use planning. As a result it is critical that companies involved in pipeline operations conduct their own risk assessments and be in a

References (12)

  • Alberta Energy and Utility Board (1998). Pipeline performance in Alberta...
  • Alberta Industrial Heartland Association (2003). Regional pipeline corridor and setback...
  • Burgess, D. S., & Zabetakis, M. G. (1973). Detonation of a flammable cloud following a propane pipeline break. The...
  • CCPS (1996). Guidelines for evaluating process plant buildings for external explosions and fires, New York: American...
  • EPA (1999). Final Longhorn Pipeline Environmental...
  • European Gas Pipeline Incident Data Group (2002). 5th EGIG Report 1970–2001, Gas Pipeline...
There are more references available in the full text version of this article.

Cited by (40)

  • Safety analysis and risk assessment of a Micro-GtL Plant

    2023, Journal of Loss Prevention in the Process Industries
  • Measuring risk in fuel supply chains

    2021, Sustainable Production and Consumption
    Citation Excerpt :

    This approach, if used within a systematic method of screening its operations, provides an organisation with a technique for identifying its most serious risks, and is thus well-suited to the assessment of energy security. Quantitative risk matrices are widely used because they are perceived as easy to construct, explain, and score (Thomas et al., 2014), for example in electricity generation (Hammond and Waldron, 2008), future (smart) electricity networks (Rossebø et al., 2017), human health (Schleier and Peterson, 2010; Vatanpour et al., 2015), pipelines (Henselwood and Phillips, 2006), process safety (Whipple and Pitblado, 2010), project management (Hillson and Simon, 2007), shipping (Hsu et al., 2017; Marchenko et al., 2018), agricultural pollution (Hewett et al., 2004), and water recycling (West et al., 2016). Several organisations give guidelines for using risk matrices in their sectors (NPSA, 2008; IMO, 2013; IPIECA and IPIECA; IOGP, 2013).

View all citing articles on Scopus
View full text