Вы находитесь на странице: 1из 7

Proceedings of the 2012 9th International Pipeline Conference IPC2012 September 24-28, 2012, Calgary, Alberta, Canada

IPC2012-90473

A GIS-BASED SYSTEM TO ASSESS THE ENVIRONMENTAL CONSEQUENCE OF A LIQUID PIPELINE RUPTURE AT WATERCOURSE CROSSINGS
Neil Ripley, M.Sc., GISP Geospatial Specialist BGC Engineering Inc. 500-1045 Howe Street Vancouver, BC, Canada, V6Z 2A9 nripley@bgcengineering.ca Elisa Scordo, M.Sc., P.Ag. Water Resources Specialist BGC Engineering Inc. 500-1045 Howe Street Vancouver, BC, Canada, V6Z 2A9 escordo@bgcengineering.ca Alex Baumgard, Ph.D, P.Eng, P.Geo Senior Geotechnical/ Environmental Engineer BGC Engineering Inc. 500-1045 Howe Street Vancouver, BC, Canada, V6Z 2A9 abaumgard@bgcengineering.ca

ABSTRACT BGC Engineering Inc. (BGC) was retained by a large pipeline operator to develop a GIS-based system to assess and rank the environmental consequence of a pipeline rupture on watercourse crossings within their pipeline system. Several physical, biological and socio-economic factors contribute to the environmental consequence of a pipeline rupture on a watercourse. This study examined select spatial and vulnerability factors, and did not consider biologic or economic impacts. Three factors were selected as part of the initial study to prioritize the pipeline watercourse crossings according to: (1) size of the watercourse at the pipeline crossing, (2) proximity of each individual crossing to larger downstream watercourses, and (3) pipeline liquid flow rate volume. A spatial analysis was conducted to determine the first two factors, while input for the third factor was provided by the pipeline operator. Watercourse size was determined using Strahlers stream order classification (Strahler 1952), while proximity to larger downstream watercourses was assessed using a Geographic Information System (GIS). This paper presents an overview of the data sources and methods used to develop an initial screening tool for identifying high consequence crossings within a pipeline system, and highlights the challenges encountered with acquiring and processing data to include in a consequence rating system. As with other pipeline risk assessments, the main challenges of this work include data availability, data integrity and resource limitations. This system is intended to fit within the pipeline operators current geohazard integrity management program and direct resources for a multi-year baseline field inspection program. INTRODUCTION BGC is developing a consequence rating system for a large pipeline operator to assess the environmental susceptibility of watercourse crossings to potential liquid content pipeline

ruptures. The environmental consequence of a pipeline rupture on a watercourse is influenced by several physical, biological and socio-economic factors. Several spatial and vulnerability factors were selected for the analysis, while the potential biological and economic consequence of a pipeline rupture was not included. Environmental sensitivities can be mapped spatially, as done by Gundlach and Hayes (1978). For example, NOAAs (National Oceanic and Atmospheric Administration) Environmental Sensitivity Index (ESI) approach uses maps to identify and protect sensitive habitats and locations if impacted by an oil spill by providing a concise inventory of coastal resources at risk (NOAA 2008). In an event of an oil spill, ESI maps are intended to help responders minimize the environmental consequences of the spill. ESI maps have been an integral component of oil-spill contingency planning since the late 1970s and are widely used internationally. Sensitivity mapping and analyses have been conducted for inland pipelines using standard GIS techniques to catalogue and geographically-reference sensitive resources, such as rivers, streams, wetlands, community drinking supplies, aquifers or protected areas (e.g., Jafari et al 2010, Paez and Roy 2010). Intersecting the environmentally sensitive areas with the pipeline system can be used to determine where to allocate additional risk reduction resources such as remote block values. The concept of pipeline risk assessment is well established in the industry. However, formal risk assessment techniques have evolved in response to regulatory requirements and also for the need for detailed integrity management programs. For example, the US Department of Transportation (US DoT) requires the identification and management of High Consequence Areas (HCA) located on any interstate pipeline. HCAs are defined as an area where transmission pipeline accidents could have a greater consequence to health and safety

Copyright 2012 by ASME

of the environment (US DoT 2002). The aim of an HCA analysis is to identify all of the pipeline segments within a network that have the potential to directly or indirectly affect an HCA, often using a GIS-based approach (e.g., Odegard and Humber 2005). Pipeline ruptures caused by geotechnical (landslides), hydrotechnical (river and stream erosion) or seismic geohazards can result in substantial environmental and economic consequences (Porter et al 2004). This study focuses on the consequences of potential ruptures caused by hydrotechnical hazards at watercourse crossings. However, this assessment system is scalable to other geohazards and corresponding sensitive areas. This system is intended to fit within the pipeline operators current geohazard integrity management program (Ripley et al 2010, Zaleski et al 2010). TERMINOLOGY The following terminology related to hazard and risk is largely adapted from Wise et al (2004), Fell et al (2004) and Leir et al (2004). Failure: An unplanned release of service fluid due to failure of a pipe or component by leak or rupture. Environmental Sensitivity: The response of an environmental unit to a given stress such as a pipeline rupture (Buckley 1982). Probability of Hazard, P(H): The probability or likelihood of occurrence of a natural phenomenon that could lead to damage. The probability is expressed as a value between 0 (no likelihood of occurrence) and 1 (absolute certainty of occurrence). Spatial Probability of Impact P(S:H): The likelihood that an element at risk is located within the area affected by a specific hazard. Spatial impact depends on the location of the hazard relative to the element as well as the potential travel distance, and velocity of the hazard. Probability of Temporal Impact, P(T:S): The probability of the element being present when the impact occurs. Temporal impact depends on how long the element is in the path of the hazard. If the element is a pipeline, the probability of temporal impact is set at 1 (certain), because the pipeline does not move. Consequence, C: The result of hazard occurrence, expressed in terms of the value of the elements damaged, destroyed or interfered with. Vulnerability, V: The degree of potential loss of an element or elements resulting from hazard occurrence, ranging from no loss to total loss. This can also be expressed as a probability of total loss. Elements at Risk, E: Population, structures, environmental features and/or economic activities in the area affected by a hazard. Can be expressed as a cost in dollars for the replacement value of the element. Risk, R: The product of hazard probability and consequence. Partial risk refers to the product of hazard and spatial probability of impact without an evaluation of vulnerability or the value of elements at risk.

CONSEQUENCE RATING COMPONENTS The goal of this program was to systematically rank the consequence of potential liquid content pipeline ruptures at watercourse crossing sites using readily available data for selected spatial probability and vulnerability factors. Risk may be expressed using this formula: R = P(H) x C (1)

The Probability of Hazard (P(H)) and Consequence (C) terms may be expanded into the sub-terms previously defined: R = P(H) x P(S:H) x P(T:S) x V x E (2)

There are two spatial probability of impact factors considered in this model. : 1. 2. Watercourse size, P(S:H)1; Distance from the potential rupture location to a downstream watercourse of a larger size class (i.e., crossing proximity), P(S:H)2

The vulnerability factor (V) is a function of the volume of liquid product flowing within the pipeline at a particular watercourse crossing. It is possible to re-organize Equation (2) by making some assumptions and setting some parameters. P(H) may be assessed through the adoption of a hazard management system such as CambioTM. In this model the hazard is assumed to be present and therefore P(H) is set to 1. Since the pipeline is a fixed asset that is assumed to be present, the temporal impact term P(T:S) is set to 1. The Element at Risk term E is to be determined by the pipeline operator and therefore is set aside. The equation now becomes an assessment of select spatial and vulnerability consequence factors and is expressed as: Rc = P(H) x P(S:H)1 x P(S:H)2 x P(T:S) x V x E Rc = P(S:H)1 x P(S:H)2 x V (3)

A consequence rating (Rc) is generated using a combination of a watercourse sensitivity index [P(S:H)1 and P(S:H)2] and vulnerability index as shown in Figure 1. Should a pipeline operator wish to conduct a more comprehensive Risk evaluation, additional factors including the Probability of Hazard and Elements at Risk could be combined with the methodology outlined herein. Since the model presented does not account for Elements at Risk and the hazard is always assumed to be present it should not be considered a full risk model but rather a Consequence model.

Copyright 2012 by ASME

greater number of tributaries contributing to the channel flow. Stream order increases when two streams of the same order converge. For example, when two first-order streams join together, they form a second-order stream. When two secondorder streams join together, they form a third-order stream, and so on. For this assessment watercourses were separated into three categories based on Strahler order to simplify the model: headwater stream, medium, or large river (Table 1). Strahler Order 13 Figure 1. Pipeline consequence rating flow chart. Description Small, headwater stream generally on steeper slopes. Channel may be dry for a portion of the year. Tributaries to larger streams. Typically unnamed crossings. Medium river. Generally less steep and lower flow velocity than headwater streams. Large river. Larger volumes of runoff and debris conveyed from smaller waterways.

46 6+

DATA SOURCES Data constraints and resource limitations are significant challenges to overcome in the development of an applied consequence rating model. A digital copy of the pipeline network currently operated by the pipeline company was acquired from IHSs AccuMap Enhanced Pipeline Dataset. The AccuMap dataset included a record of pipeline license and line numbers, which were used to establish a link between the digital pipeline network and pipeline operators internal database. With the link established, it was then possible to assign pipeline flow rates to individual pipelines. Publicly available digital hydrography data for the pipeline network were obtained from GeoBase (www.geobase.ca). Geobase is an initiative of the federal, provincial and territorial governments of Canada that provides a centralized internet accessible repository of geospatial data. The hydrography data obtained from GeoBase was part of the National Hydro Network (NHN) dataset. This data was intersected with the digital pipeline network to identify each of the pipeline watercourse crossings at a scale of 1:50,000. Results of the analysis were determined to be scale dependent. WATERCOURSE SENSITIVITY INDEX A watercourse sensitivity index was developed based on two criteria: (1) watercourse size, and (2) downstream proximity to a larger stream or river (i.e., crossing proximity). Watercourse size was determined using Strahlers stream order classification (Strahler 1952), while crossing proximity was assessed using a GIS. All stream channels within the GeoBase dataset were assigned Strahler stream orders using GIS techniques discussed in Holloway (2009). The Strahler stream order hierarchy is a method to classify the relative size of a perennial or ephemeral stream. Sizes range from a first-order stream all the way to the largest, 12th order stream (e.g., the Amazon River). A first to third-order stream corresponds to small, headwater streams, while a higher order stream indicates a larger channel with a

Table 1. Strahler order summary. Figure 2 shows an example of the stream classification employed for a small section of the study area. The headwater streams join together into a few medium size rivers which, in turn merge into a large river in the north (top) end of Figure 2. A majority of the streams shown in Figure 2 (and within the study) are classified as headwater streams.

Figure 2. Geobase NHN hydrology dataset with stream network organized by stream category size.

Copyright 2012 by ASME

An assumption of the watercourse sensitivity index is that larger streams have a higher potential environmental sensitivity if impacted by a pipeline rupture. Large streams are more likely to be fish bearing and the negative publicity resulting from a spill in these streams is likely to be substantial. Also, while this assumption is not necessarily valid for all cases, the amount of biological data required for a more accurate assessment is prohibitive for a screening-level analysis like the one presented here. Thus, watercourse size, in this study is considered a suitable proxy for environmental sensitivity. The downstream proximity to a larger stream or river was developed by overlaying the pipeline stream crossings with publicly available digital hydrographic data and calculating the distance to the next watercourse size category. Three distances were measured: Downstream distance from a pipeline crossing of a headwater stream to a medium river; Downstream distance from a pipeline crossing of a headwater stream to a large river; and Downstream distance from a pipeline crossing of a medium river to a large river. As large rivers are considered the final watercourse category, no downstream distance measurements are required for pipelines crossing a large river. Watercourse size and crossing proximity were then combined for each crossing into one of five classes of watercourse sensitivity (Table 2). Watercourses within 5 km of a higher order stream (e.g., medium or large river) were given an elevated sensitivity rating due to the larger streams increased carrying capacity compared to watercourses that are further than 5 km from higher order streams. For example, if a pipeline ruptured in a reach with low stream order (e.g., headwater stream) and the distance to downstream reaches with a higher watercourse category was greater than 5 km, then the sensitivity would be classed as very low. The 5 km distance was based on observations from the distribution of all proximities. This was arbitrarily chosen and may vary in other studies. Table 2. Watercourse sensitivity index classes. Class Very Low Low Medium High Very High Description Headwater Stream with greater than 5 km distance to a Medium River Headwater Stream within 5 km distance to a Medium River Medium River with greater than 5 km distance to a Large River Headwater Stream or Medium River within 5 km to a Large River Large River

Figure 3. Distribution of 15,767 pipeline crossings identified in the study area. PIPELINE FLOW RATES Liquid flow rates for each of the pipeline crossings were provided by the pipeline operator. These rates provide a relative indication of the volume of product that could be released in a potential rupture. Liquid products for this pipeline operator typically include fuel gas, oil emulsion, and salt water. The consequence of gas leaks would partially depend on wind and topography, thus requiring a different consequence model as to what has been developed here therefore only liquid content pipelines have been considered. Crossings were categorized into one of three vulnerability index classes (low, medium or high volumes) as summarized in Table 3. The three classes were selected based on observations of the entire pipeline flow rate data distribution. Table 3. Vulnerability index classes. Class Low Medium High Pipeline Flow Rate < 10 m3/day 10 to 100 m3/day > 100 m3/day

Figure 3 shows the frequency distribution of 15,757 crossings within the operators system categorized into the five sensitivity index classes.

Figure 4 shows the frequency distribution of the flow rates for the crossings used in this study. Since not all pipelines in the study had liquid flow rate information, either because the pipeline had no liquid contents or such information was not available at the time of the study, the final number of crossings which could be assessed was reduced from 15,767 to 1,883 as shown in Figure 4. This reduction in the dataset indicates the challenges faced with acquiring consequence related data. As more additional pipeline liquid flow rate data becomes available from the operator it will be added into the model database. For simplicity in this model no attempt was made to estimate pipeline liquid flow rate based on pipeline diameters or other attributes.

Copyright 2012 by ASME

Figure 4. Pipeline flow rates for 1,883 pipeline crossings in the study area. Only pipelines with operator provided liquid flow rates were used in this study. CONSEQUENCE INDEX The watercourse sensitivity was then combined with the 1,800+ categorized pipeline flow rates (low, medium or high volume) to derive the environmental consequence rating. Table 4 shows the matrix employed for combining the elements together into an overall consequence rating. Table 4. Environmental consequence matrix.

Figure 5. Example map showing consequence ratings for various pipeline crossings. Applying the environmental sensitivity consequence matrix in Table 4 to the 1,883 crossings in the study shows that 95% of all crossings are placed in the low consequence category. The medium consequence category contains an additional 3% of the crossings. The higher consequence crossings represent 2% of all crossings, as shown in Figure 6.
2000 1800 1600 1400 1200 1784

Frequency

1000 800 600 400 200 61 0 Low Medium Consequence Ratings High 38

Figure 5 shows a subset of crossings ranked by environmental consequence. The majority of crossings are on headwater streams and are more than five km upstream from larger order streams and therefore ranking as low or very low. However there are six crossings ranked as medium consequence that are either due to a combination of the pipeline liquid flow rate or proximity to larger order streams. One crossing is ranked as high consequence because it occurs on a larger watercourse and has a medium to high pipeline flow rate.

Figure 6. Environmental consequence ratings for 1,883 pipeline crossings.

Copyright 2012 by ASME

CAMBIOTM DATABASE AND GEOHAZARD MANAGEMENT PROGRAM Results of the pipeline consequence model analysis are stored within the pipeline operators CambioTM geohazards database. (Ripley et al 2010). Users logging into CambioTM are presented with a list of all crossings found along the pipeline right-of-way (RoW). Users can sort the crossing list by location, category, or consequence risk ranking among other methods. Subsets of the list can be selected by filling out a selection criteria form (Figure 7).

Figure 8. Detailed information is available for each crossing including the crossings geographic location, distances and names of downstream rivers, and the site ranking on different indices. SUMMARY This paper has discussed the design and implementation of a simple system for ranking pipeline stream crossing sites by environmental consequence. The system can be used as a screening tool for identifying higher consequence elements of a pipeline system. The operator may use this information, in conjunction with additional criteria outside the scope of this study, to establish corporate standards for operational contingency plans. Operator resources can therefore be allocated towards the stream crossings and other elements identified as higher consequence in a systematic fashion, such as identifying sites or areas to be prioritized for installation of block valves or increased patrols. Figure 7. Defining criteria for selecting crossing sites. Detailed information for individual crossings is available by selecting an individual crossing from the list. An example is shown in Figure 8. Details available for each crossing site include the pipeline license and line number, the geographic location of the crossing, the names and distances to downstream rivers and information about the pipeline and the ranking of the site on different indices. ACKNOWLEDGEMENTS The authors would like to recognize the support of the member operators of the Pipeline Geohazards Working Group for commissioning this study. This system is intended to fit within the pipeline operators current geohazard integrity management program and direct resources for a multi-year baseline field inspection program. REFERENCES Buckley, R.C. 1982. Environmental sensitivity mapping what, why and how. Environmental Geochemistry and Health 4: 151-155. Fell, R., Ho, K.K.S., Lacasse, S., and Leroi, E. 2004. A framework for landslide risk assessment and management. Landslide Risk Management. O. Hungr, R. Fell, R. Couture, and E. Eberhardt, eds. A.A. Balkema, London Gundlach, E.R. and Hayes, M. 1978. Classification of coastal environments in terms of potential vulnerability to oil spill damage. Marine Technology Society Journal, vol.12(4), p.18-27.

Copyright 2012 by ASME

Holloway, R. 2009. Using RivEX and National Hydro Network Data to Classify Water Quality Stations by Strahler Stream Order. Government of Newfoundland and Labrador. Jafari, J., Khorasani, N. and Danehkar, A. 2010. Using Environmental Sensitivity Index (ESI) to assess and manage environmental risks of pipelines in GIS environment: A case study of a near coastline and fragile ecosystem located pipeline. World Academy of Science, Engineering and Technology 68: 1092-1101. Leir, M.. 2004. Bridging the gap between field operations and risk management. Proceedings, Terrain and geohazard challenges facing onshore oil and gas pipelines. Thomas Telford, London. Leir, M., Reed, M., Yaremko, E. Field Inspection Module for Hydrotechnical Hazards. Proceedings, IPC 2004, International Pipeline Conference, ASME, New York. National Oceanic and Atmospheric Administration (NOAA). 2008. Environmental Sensitivity Index Mapping. NOAAs National Ocean Service, Office of Response and Restoration. Odegard, L. and J. Humber. 2005. Inventorying directly and indirectly affected high consequence areas. Geotech Event, February 13-16, 2005. Paez, J. and A. Roy. 2010. Developing a pipeline risk assessment tool for the upstream oil and gas industry. Proceedings, Northern Area Western Conference, NACE International, Calgary, Alberta.

Porter, M., Logue, C., Savigny, K.W., Esford, F., and Bruce, I. 2004. Estimating the influence of natural hazards on a pipeline risk and system reliability. Proceedings, IPC 2004, 5th International Pipeline Conference, ASME, New York. Porter, M., Marcuz, G., Reale, R., and Savigny, K.W. 2006. Geohazard risk management for the Nor Andino gas pipeline. Proceedings, IPC 2006, 6th International Pipeline Conference, ASME, New York. Ripley, N., Simpson, T., and Leir, M. 2010. Lessons learned from supporting a geohazard management program. Proceedings, IPC 2010, 8th International Pipeline Conference, ASME, New York. Strahler, A.N. 1952. Hypsometric (area-altitude) analysis of erosional topography. Bulletin Geological Society of America. 63:1117-1142 US Department of Transportation. 2002. Pipeline Safety: Pipeline Integrity Management in High Consequence Areas (Gas Transmission Pipelines). 49 CFR Part 192, Docket No. RSPA-00-7666, Amendment 192-95, RIN 2137-AD54. Wise, M.P., Moore, G.D. and VanDine, D.F. (editors). 2004. Landslide risk case studies in forest development planning and operations. B.C. Ministry of Forestry, Resource Branch, Victoria, B.C. Land Management Handbook No. 56. Zaleski, M., Greaves, T., and Bracic, J. 2010. Meeting the geohazards management guidelines of Annex N. Proceedings, IPC 2010, 8th International Pipeline Conference, ASME, New York.

Copyright 2012 by ASME

Вам также может понравиться