PSAM 16 - Session M15 Overview

PSAM 16 Conference Session M15 Overview

Session Chair: Albena Stoyanova (albena.stoyanova@kkl.ch)

Paper 1 ST94

Lead Author: Steve Prescott Co-author(s): James Knudsen (james.knudsen@inl.gov) Stephen T. Wood (stephen.wood@inl.gov)

Automated-fire PRA Scenario Modeling in SAPHIRE Using FRI3D
The current fire modeling practices require many manual steps and processes to transfer fire scenario information and data between software tools such CAFTA, CFAST, databases, etc. To help in minimizing these manual steps, the Fire Risk Investigation in 3D (FRI3D) software was developed. The primary goal of FRI3D is to automate as many steps as possible and link the 3D spatial information with the probabilistic risk assessment (PRA) information. Initially FRI3D was coupled directly with the existing FRANX fire input data and CAFTA to perform the fire analysis. The modular design allows for coupling with other PRA tools and the PRA software, Systems Analysis Programs for Hands-on Reliability Evaluations (SAPHIRE), was incorporated for a facility pilot project. SAPHIRE does not have designated tools that can be used to assign specific fire scenarios; therefore, the fire scenarios need to be added manually which is time-consuming and potentially error prone. To make FRI3D compatible with SAPHIRE and simplify the scenario modeling, a new module was created which automatically generates the fire scenario and inputs it into SAPHIRE for evaluation using SAPHIRE Macros. This module constructs the PRA components, following a standard currently used to evaluate fire scenarios within SAPHIRE. This paper outlines the SAPHIRE fire modeling standard and the methods used by FRI3D and provides an example of a fire scenario being added into SAPHIRE for evaluation based on a generic compartment from a generic facility.

Automated-fire PRA Scenario Modeling in SAPHIRE Using FRI3D

The current fire modeling practices require many manual steps and processes to transfer fire scenario information and data between software tools such CAFTA, CFAST, databases, etc. To help in minimizing these manual steps, the Fire Risk Investigation in 3D (FRI3D) software was developed. The primary goal of FRI3D is to automate as many steps as possible and link the 3D spatial information with the probabilistic risk assessment (PRA) information. Initially FRI3D was coupled directly with the existing FRANX fire input data and CAFTA to perform the fire analysis. The modular design allows for coupling with other PRA tools and the PRA software, Systems Analysis Programs for Hands-on Reliability Evaluations (SAPHIRE), was incorporated for a facility pilot project. SAPHIRE does not have designated tools that can be used to assign specific fire scenarios; therefore, the fire scenarios need to be added manually which is time-consuming and potentially error prone. To make FRI3D compatible with SAPHIRE and simplify the scenario modeling, a new module was created which automatically generates the fire scenario and inputs it into SAPHIRE for evaluation using SAPHIRE Macros. This module constructs the PRA components, following a standard currently used to evaluate fire scenarios within SAPHIRE. This paper outlines the SAPHIRE fire modeling standard and the methods used by FRI3D and provides an example of a fire scenario being added into SAPHIRE for evaluation based on a generic compartment from a generic facility.

Paper ST94 | Download the paper file. | Download the presentation PowerPoint file.

Name: Steve Prescott (Steven.Prescott@inl.gov)

Bio: Steve Prescott is a software engineer at the Idaho National Laboratory. He started working on the PRA software EMRALD as an intern and has never stopped working on risk analysis software since. Now his primary focus is on dynamic PRA and incorporating other tools such as fire, flood, and physical security. For fun and to have a place to live, over the last few years, he designed and built a net zero house that is solar powered and solar heated.

Country: ---
Company: Idaho National Labratory
Job Title: Software Engineer

Paper 2 DE55

Lead Author: Yoshikazu Deguchi Co-author(s): Keisuke Himoto, himoto-k92ta@mlit.go.jp

HIERARCHICAL BAYESIAN ESTIMATION OF FIRE GROWTH RATE FOR VARIOUS BUILDING USAGES BASED ON THE FIRE INCIDENTS REPORT
In the evacuation safety design of a building, it is necessary to set an appropriate design fire source, which can be represented by the fire growth rate α. In the basic Japanese procedure for evacuation safety design, the Verification method for evacuation safety, the fire growth rate α is defined by the sum of the αf, value calculated by the calorific value per unit floor area of loaded combustibles ql , and the αm, value correlated with the type of interior finish. However, it is not clear what levels of the fire growth rate α of the Verification method for evacuation safety are required for an actual fire. DEGUCHI et.al (2011) statistically derived the distributions of fire growth rate for rooms of selected usages by using the statistics on burnt floor area, extinction time, etc. of reported fires from 1995 to 2008. The authors concluded that the fire growth rates of an office, residence, restaurant and retail store can be approximated by lognormal distributions. However, the fire growth rates could have been obtained only for usages for which sufficient data had been available. In this study, distributions of fire growth rate α for rooms of various usages, including those with scarce data availability, have been estimated by using Hierarchical Bayesian Method. In addition, the levels of fire growth rates α of the Verification method for evacuation safety have been estimated by comparing them with the estimated fire growth rate α. First, fire data for 24 years from the Fire Incidents Report, which is issued annually by the National Fire Defense Agency, were categorized into 19 groups of 41 usages. Second, the distributions of the burnt floor area expansion rate Af⁄t^2 were estimated by the Hierarchical Bayesian Method, which can be used to estimate distributions where the number of data is insufficient. Finally, distributions of fire growth rates α were estimated by the Hierarchical Bayesian Method, from the product of the distributions of the burnt floor area expansion rate Af⁄t^2 and the heat release rate per unit floor area q". As a result, the following results were obtained: 1. Regression by the Hierarchical Bayesian Estimation was robust; the distributions of the fire growth rate α for various usages could be estimated in a plausible form by considering the overall characteristics. 2. In most usages, assuming the interior finish was a quasi-noncombustible material, the fire growth rate α of the method lay between the 75th and 95th percentiles of the estimated distributions of fire growth rate α.

HIERARCHICAL BAYESIAN ESTIMATION OF FIRE GROWTH RATE FOR VARIOUS BUILDING USAGES BASED ON THE FIRE INCIDENTS REPORT

In the evacuation safety design of a building, it is necessary to set an appropriate design fire source, which can be represented by the fire growth rate α. In the basic Japanese procedure for evacuation safety design, the Verification method for evacuation safety, the fire growth rate α is defined by the sum of the αf, value calculated by the calorific value per unit floor area of loaded combustibles ql , and the αm, value correlated with the type of interior finish. However, it is not clear what levels of the fire growth rate α of the Verification method for evacuation safety are required for an actual fire. DEGUCHI et.al (2011) statistically derived the distributions of fire growth rate for rooms of selected usages by using the statistics on burnt floor area, extinction time, etc. of reported fires from 1995 to 2008. The authors concluded that the fire growth rates of an office, residence, restaurant and retail store can be approximated by lognormal distributions. However, the fire growth rates could have been obtained only for usages for which sufficient data had been available. In this study, distributions of fire growth rate α for rooms of various usages, including those with scarce data availability, have been estimated by using Hierarchical Bayesian Method. In addition, the levels of fire growth rates α of the Verification method for evacuation safety have been estimated by comparing them with the estimated fire growth rate α. First, fire data for 24 years from the Fire Incidents Report, which is issued annually by the National Fire Defense Agency, were categorized into 19 groups of 41 usages. Second, the distributions of the burnt floor area expansion rate Af⁄t^2 were estimated by the Hierarchical Bayesian Method, which can be used to estimate distributions where the number of data is insufficient. Finally, distributions of fire growth rates α were estimated by the Hierarchical Bayesian Method, from the product of the distributions of the burnt floor area expansion rate Af⁄t^2 and the heat release rate per unit floor area q". As a result, the following results were obtained: 1. Regression by the Hierarchical Bayesian Estimation was robust; the distributions of the fire growth rate α for various usages could be estimated in a plausible form by considering the overall characteristics. 2. In most usages, assuming the interior finish was a quasi-noncombustible material, the fire growth rate α of the method lay between the 75th and 95th percentiles of the estimated distributions of fire growth rate α.

Paper DE55 | Download the paper file. | Download the presentation PowerPoint file.

Name: Yoshikazu Deguchi (deguchi-y92ta@mlit.go.jp)

Bio: Dr. Deguchi is a Senior Reseacher in NILIM(National Institute for Land and Infrastructure Management), Japan. He has worked for performance-based evacuation safety design and developments for fire safety technologes. Also, he is engaged in the revision of fire-protection-related regulations. His research area includes: burning behavior of combustibles, fire risk assesment and so on.

Country: JPN
Company: National Institute for Land and Infrastructure Management
Job Title: Senior Reseacher

Paper 3 A-50

Lead Author: Ayako Hirose Co-author(s): Daisuke Takeda (d-takeda@criepi.denken.or.jp) Kohei Nonose (nonose@criepi.denken.or.jp) Koji Tasaka (kotasaka@criepi.denken.or.jp)

An Exploratory Study on Decision to Main Control Room Abandonment due to Fire-Induced Loss of Habitability: Using a VR Nuclear Power Plant Main Control Room Simulator
In Fire PRA/HRA on nuclear power plant, the criteria related to concentration of smoke or room temperature for main control room (MCR) abandonment due to loss of habitability (LOH) are given by NUREG-6850. However, the criteria are so severe for human body that it is unclear at what point operators will decide to abandon the MCR if a fire actually breaks out. Also, because MCR fire events are very rare and impossible to be simulated in real MCR, it would be difficult for operators themself to answer clearly about their own timing to decide to abandon during MCR fires. Using virtual reality (VR) technology to allow operators to experience MCR fires may make it easier to collect data about their timing of MCR abandonment. Therefore, this study aims to develop a virtual environment of an MCR fire using a VR nuclear power plant MCR simulator and collect data on decision-makings on MCR abandonment to improve fire HRA. Based on NUREG-1934, content with event of an electrical cabinet fire in MCR was developed by using the VR nuclear power plant MCR simulator. To make the content more realistic, an interview about responses to fire events in MCR, with checking the content, was conducted with the sift managers of an electric power company, and malfunctions to be simulated were determined. An experiment with operators as subjects will be conducted using the contents revised with reference to the interview.

An Exploratory Study on Decision to Main Control Room Abandonment due to Fire-Induced Loss of Habitability: Using a VR Nuclear Power Plant Main Control Room Simulator

In Fire PRA/HRA on nuclear power plant, the criteria related to concentration of smoke or room temperature for main control room (MCR) abandonment due to loss of habitability (LOH) are given by NUREG-6850. However, the criteria are so severe for human body that it is unclear at what point operators will decide to abandon the MCR if a fire actually breaks out. Also, because MCR fire events are very rare and impossible to be simulated in real MCR, it would be difficult for operators themself to answer clearly about their own timing to decide to abandon during MCR fires. Using virtual reality (VR) technology to allow operators to experience MCR fires may make it easier to collect data about their timing of MCR abandonment. Therefore, this study aims to develop a virtual environment of an MCR fire using a VR nuclear power plant MCR simulator and collect data on decision-makings on MCR abandonment to improve fire HRA. Based on NUREG-1934, content with event of an electrical cabinet fire in MCR was developed by using the VR nuclear power plant MCR simulator. To make the content more realistic, an interview about responses to fire events in MCR, with checking the content, was conducted with the sift managers of an electric power company, and malfunctions to be simulated were determined. An experiment with operators as subjects will be conducted using the contents revised with reference to the interview.

Paper A-50 | Download the paper file. |

Name: Ayako Hirose (a-hirose@criepi.denken.or.jp)

Bio: Ayako Hirose is a researcher at Central Research Institute of Electric Power Industry (CRIEPI). She is currently engaged in research using virtual reality to construct stressful environments and measure human behavior in these environments. She plans to incorporate the obtained data into human reliability analysis. Her expertise is physiological psychology, and she has been conducting research on human factors since she joined the institute.

Country: JPN
Company: Central Research Institute of Electric Power Industry
Job Title: Senior Research Scientist