A new look at release event frequencies. (March 2021)

• Record Type: Journal Article
• Authors: Marx, Jeffrey D.; Ishii, Benjamin R.
• Is Part Of: Journal of loss prevention in the process industries. Volume 69(2021)
• Publication Date: 2021-03
• Subjects: Probability -- Frequency -- Quantitative risk analysis -- QRA; Chemical industries -- Safety measures -- Periodicals; 660.2804
• DOI: 10.1016/j.jlp.2020.104368
• Languages: English
• ISSNs: 0950-4230

Abstract

Abstract: Within the context of a quantitative risk analysis (QRA), the two main constituents used to describe petrochemical risks are, and have always been, consequence and probability. The consequences of hazardous material accidents are easy to apprehend – if a hazard is realized it can injure people or cause fatalities, damage equipment or other assets, or cause environmental damage. Frequencies for these consequences, on the other hand, are not as easy to understand. Process safety professionals develop event frequencies by evaluating historical data and calculating incident rates, which represent, in the QRA context, how often a release of a hazardous material has occurred. Incident rates are further modified by probabilities for various hole sizes, release orientations, weather conditions, ignition timing, and other factors, to arrive at unique event probabilities that are applied in the QRA. This paper describes the development of incident rates from historical database information for various equipment types, as well as defining a methodology for assigning hole size probabilities from the same data, such that a hole size distribution can be assigned within each QRA study. The combination of total incident rates and a hole size distribution relationship can then serve as a foundation within the frequency side of many QRA studies. Highlights: Simplified method for evaluating hydrocarbon process equipment failure frequencies. Data analysis to filter events not applicable to quantitative risk analysis (QRA). Fit to a log-logistic curve. Tabulation of total event frequencies and curve fit parameters for use in QRA studies. Flexible and adaptable methodology.