PSAM 16 - Paper Overview

PSAM 16 Conference Paper Overview

Welcome to the PSAM 16 Conference paper and speaker overview page.

Lead Author: James Knudsen Co-author(s): Curtis L. Smith (Curtis.Smith@inl.gov) Michael Calley (Michael.Calley@inl.gov)

Issues and Approaches Regarding Success Terms for Probabilistic Risk Assessment Models
Solving event tree accident sequences in probabilistic risk assessments (PRAs) involves assumptions about the success of systems (i.e., event tree top events). The primary assumption is that failure of the system is a rare event (i.e., a low probability of failure); therefore, the success probability is very close to 1.0. Under most conditions, this assumption is valid. However, such is not always the case. When event tree top events have higher failure probabilities—and thus success probabilities that are not close to 1.0—this assumption causes the sequences with success branches to be incorrect. To address this issue, it is necessary to properly account for the success probability of event tree top events in order to better quantify each sequence of the event tree. The current state of practice allows for some success event tree top events to be included in the accident sequence cut sets (i.e., only single basic events representing the success term and its value). In addition, sequence cut sets are typically quantified using the minimal cut set upper bound approximation versus the exact solution. Combining success terms in the sequence cut sets with the minimal cut set upper bound approximation tends to make the sequence’s quantification results appear close to the exact results. This paper presents two approaches to better quantify event tree sequences featuring cut sets that include success terms with low values (i.e., the respective failure value is higher). One approach quantifies each sequence, taking into account the success term at its low value by using the minimal cut set upper bound approximation. The second approach is similar, except that the success term is quantified via a binary decision diagram (BDD) approach, and the result is them multiplied back into each sequence cut set. These two approaches will be evaluated using a complex PRA model for two standard event trees. The evaluation will involve a loss of offsite power event and a seismic external event. Both of these event trees contain complex systems that can carry a high probability of failure The pros and cons of these approaches will be discussed along with the overall conclusions.

Issues and Approaches Regarding Success Terms for Probabilistic Risk Assessment Models

Solving event tree accident sequences in probabilistic risk assessments (PRAs) involves assumptions about the success of systems (i.e., event tree top events). The primary assumption is that failure of the system is a rare event (i.e., a low probability of failure); therefore, the success probability is very close to 1.0. Under most conditions, this assumption is valid. However, such is not always the case. When event tree top events have higher failure probabilities—and thus success probabilities that are not close to 1.0—this assumption causes the sequences with success branches to be incorrect. To address this issue, it is necessary to properly account for the success probability of event tree top events in order to better quantify each sequence of the event tree. The current state of practice allows for some success event tree top events to be included in the accident sequence cut sets (i.e., only single basic events representing the success term and its value). In addition, sequence cut sets are typically quantified using the minimal cut set upper bound approximation versus the exact solution. Combining success terms in the sequence cut sets with the minimal cut set upper bound approximation tends to make the sequence’s quantification results appear close to the exact results. This paper presents two approaches to better quantify event tree sequences featuring cut sets that include success terms with low values (i.e., the respective failure value is higher). One approach quantifies each sequence, taking into account the success term at its low value by using the minimal cut set upper bound approximation. The second approach is similar, except that the success term is quantified via a binary decision diagram (BDD) approach, and the result is them multiplied back into each sequence cut set. These two approaches will be evaluated using a complex PRA model for two standard event trees. The evaluation will involve a loss of offsite power event and a seismic external event. Both of these event trees contain complex systems that can carry a high probability of failure The pros and cons of these approaches will be discussed along with the overall conclusions.

Paper JA278 Preview

Author and Presentation Info

Lead Author Name: Jim Knudsen (james.knudsen@inl.gov)

Bio: Mr. Knudsen is Risk Analysis Engineer with over twenty years of probabilistic risk assessment (PRA) experience including all aspects of model development. He is the project lead for the different Standardized Plant Analysis Risk (SPAR) model development and enhancement projects and also the project lead for the Systems Analysis Programs for Hands-on Integrated Reliability Evaluations (SAPHIRE) software program developed for the NRC. He provides instruction and technical support to NRC and NASA staff on the use of the risk models. He also provides technical training in PRA concepts and PRA modeling along with the use of the SAPHIRE PRA code.

Country: United States of America
Company: Idaho National Laboratory
Job Title: Engineer

PSAM 16 Conference Paper Overview

Welcome to the PSAM 16 Conference paper and speaker overview page.

Paper JA278 Preview

Author and Presentation Info

Download paper JA278.

Download the presentation PowerPoint file.