In our previous post in the series, Not Every HCA Analysis is Created Equal, we talked about the importance of verifying analyses provided through software or vendors, and the potential vulnerabilities to consider, as well as the basis of identifying HCAs and their boundaries. In this installment we’ll be deliberating how model settings and data can impact emergency response plans and generating release scenarios.
Common Concern: “Optimistic operator incident response time assumption is not adequately justified”
FAQ 3.3 How will an operator determine if a pipeline can affect an HCA?
Part 195 Appendix C of the rule provides guidance on factors an operator should consider in determining whether a pipeline can affect an HCA. An example is provided in the Appendix. The factors are: Response capability (time to respond, nature of response)
We should recall that HCA analysis software is a model, and it will simulate whatever we tell it to, and model settings/inputs must be consistent with the operator’s written IMP, be complete, accurate, and best reflect the reality of a worst case release. Otherwise the model will simulate a release, and provide us with results that could be suspect. Operators and vendors alike should have well defined processes that define specific data needs for HCA analyses. Furthermore, a data QC should be conducted prior to model execution, as the reality of data being handled by either electronic databases/exchanges and humans is imperfect. It is not beyond the realm of possibility for an operator to provide a modeling vendor with incomplete or incorrect data. For example, the use of incorrect units could have significant impacts to the analysis. This could include providing depth-of-cover in inches, while the documentation indicates feet or providing wall thickness in millimeters, while the documentation indicates inches. These issues, sometimes simply editorial, can significantly skew HCA analysis results and proper identification of consequences and could-affect segments. Thus, well-defined processes and QC steps are critical to ensuring correct and complete data, which is key for producing accurate results. A common area that yields incorrect modeling results is the time to detect a release, shutdown the pipeline, and isolate the segment. Prior to conducting an HCA analysis, the operator should closely review response and valve closure times and emergency response plans associated with their systems, as overly optimistic estimates will only serve to provide understated product release volumes and potentially inaccurate analysis results. If the system is automated and much of the leak detection is based on pressure anomalies, and valves close automatically, the detection and shutdown times can be fairly quick. Often, however, there is a dependence on personnel to recognize and shut down pumps and valves. In reality, even with automated systems, detection and shutdown and full segment isolation can take longer than operators expect, often on the order of an hour or more. Operators thoroughly vet these times with various stakeholders including operations staff, control room staff, emergency response staff, and integrity personnel. Being conservative in these estimates is prudent, as an operator does not want the liabilities associated with understated could-affect mileage.
FAQ 3.5 Do operators need to perform detailed consequence analysis to determine the specific impacts on population or USAs?
Yes. PHMSA Pipeline Safety expects that an operator will develop an understanding of the potential consequences of leaks and ruptures of its pipelines. The operator should be able to estimate the severity of releases in terms of volume of hazardous liquid that could be released, the physical pathways and dispersion mechanisms by which the commodity can be transported to an HCA, the amount of commodity that might actually reach the boundaries of the HCA, and the population and environmental resources that can be affected by such a release.
Protocol Question: IMP High Consequence Areas Direct Intersect Method and Direct Intersect Exceptions- Is the process adequate to determine all locations where pipeline systems "could affect" a high consequence area, including pipelines that are located in HCAs?
If there is a failure, what are the potential impacts and in what ways can those impacts occur? We should recall that there are different types of impacts, including both direct and indirect impacts. A common misconception is that if the pipeline does not go through an HCA, causing a direct impact, there is no potential for impact. In fact, there are a number of different ways that a pipeline rupture can affect HCAs and it is the responsibility of the operator to determine what those potential impact methods are (Figure 2, Direct and Indirect Liquid Impacts). For example, if a pipeline segment that is not located in an HCA ruptures, and the product flows via a creek into an adjacent ECO area, that would be considered an indirect impact.
Common Concern: “Using predefined or fixed-distance release locations that do not identify all release locations that could affect a HCA, resulting in missed segments that could affect HCAs”
“Operators that set a predefined spacing for release points could result in missing HCAs that could be affected by a release”
“If the nearest postulated release point does not result in spill migration analysis that predicts the spill reaching the stream, then a significant water transport mechanism could remain unanalyzed”
“Failure to adequately consider stream flow characteristics, including potential stream flow velocity and greater than average flow.”
FAQ 3.4 What is acceptable methodology and criteria for determining whether a segment could affect an HCA? (For example what spill volume should be considered - Worst-case discharge? Most likely discharge? Most likely worst-case discharge?) Can an arbitrary safe distance be applied or must location specific dispersion analyses be performed? Is air dispersion modeling expected or is spill trajectory adequate?
PHMSA Pipeline Safety expects each operator to develop a process for identifying what portions of its pipeline system could affect an HCA in the event of a failure. This process is a required Integrity Management program element per §195.452 (f). Operators are responsible for selecting a methodology and establishing any criteria needed to determine where pipeline failures could affect HCAs.
PHMSA Pipeline Safety will look for sound engineering judgment with a reasonable amount of conservatism to account for uncertainties in the assumptions and calculation methods used in the analysis. Operators should be able to justify the assumptions used in making these determinations.
Generating release scenarios in the correct locations is also important for generating accurate results. A key element is to provide enough release scenarios along a pipeline that if additional scenarios were added to the analysis, there would be no improvement in coverage or additional potential impacts found. For most hazardous liquids, a release interval of at least every 500 feet and every stream or river crossing is appropriate.
Properly identifying all potentially affected waterways and distances is crucial to understanding potential impacts and consequences of a release. Another common assumption, and input, is that all stream velocities are the same. However, common sense tells us this can lead to poor results, and to accurately calculate potential product transportation, realistic velocities for the specific waterway would provide more accurate results. This includes intermittent waterways that may run only during certain times of the year or after a storm. A velocity that represents the high end of potential velocities for a waterway is important. Furthermore, operators should review their analyses to confirm the use of the National Hydrography Dataset, a high resolution geometric network of stream and rivers, to provide a comprehensive analysis of stream and river impacts.
Understanding how model settings affect results, and inputting realistic parameters and data, are key components required to understanding potential impacts and identifying HCAs. Don’t miss the last part of the series tomorrow, where we’ll explore the effect of the type of product being transported and how to incorporate pipeline facilities into the analysis.