A source term binning methodology for multi-unit consequence analyses

https://doi.org/10.1016/j.ress.2020.106989Get rights and content

Highlights

  • There are a lot of source term category (STC) combinations in multi-unit accidents.

  • A proper methodology for multi-unit level 3 PSA should be developed and verified.

  • A source term binning methodology to group many STCs into fewer groups was developed.

  • Qualitative and quantitative logic trees for grouping tasks were developed.

  • Several methods to designate representative inputs of each group were proposed.

  • Important assumptions for considering different reactor type or site were proposed.

Abstract

Public has concerned on a site risk after the Fukushima nuclear accidents. Therefore, multi-unit probabilistic safety assessment (MUPSA) has been researched actively because it is necessary to perform MUPSA for assessing a site risk. Most of the researches performed until now focus on multi-unit Level 1 PSA because it is the most important issue to model dependencies between units. However, multi-unit Level 3 PSA should be researched because many source term category (STC) combinations exist in multi-unit accidents. A binning methodology to group many STCs into fewer groups was developed in this research for considering this problem. First, a qualitative and quantitative logic tree to group similar STCs into a same group were developed, respectively. Second, five methods to designate representative MACCS inputs of each group were developed. Third, a verification procedure for the binning methodology was developed, and the most appropriate method for each logic tree was selected. The scheme of the binning methodology can be applied to any reactor type with several modifications such as the headings of the logic tree. Conclusively, the binning methodology will be an appropriate example of multi-unit Level 3 PSA and important element of a site risk estimation methodology.

Introduction

Researches on multi-unit probabilistic safety assessment (MUPSA) have been performed actively in the world. Single unit probabilistic safety assessment (SUPSA) was traditionally performed to assess a single unit risk. However, there are many efforts to develop the MUPSA methodology to assess a site risk after the Fukushima-Daiichi nuclear accidents in 2011. The MUPSA methodology is the most important technical element for assessing a site risk. Therefore, many countries are trying to expand the existing SUPSA methodology to the MUPSA methodology with several modifications. PSA generally consists of three levels which are Level 1, 2, and 3. Many different techniques can be applied for PSA such as the ‘large event tree (ET) and small fault tree (FT) method’, the ‘small event tree and large fault tree method’, etc. Among of them, the small ET and large FT method was considered in this research. In Level 1 PSA, core damage frequency is quantified by using ETs and FTs. All ETs considered in Level 1 PSA are expanded to plant damage state ETs (PDS ETs) by considering systems related to containment. Based on the results of the PDS ETs, all core damage accident sequences are grouped into PDSs by PDS logic diagram (PDS LD). All end points of the PDSs are expanded by containment event tree (CET) and decomposition event tree (DET). These results are used for assessing large early release frequency (LERF) or large release frequency (LRF) in Level 2 PSA. All end points of the CET are grouped into source term categories (STCs) by STC logic diagram (STC LD). Representative accident sequences of each STC are selected and utilized to obtain release characteristics by accident progression analyses. Based on these release characteristics of each STC, conditional consequences are assessed in Level 3 PSA. Finally, risks are assessed by multiplying frequencies and conditional consequences of each STC. This implementation process is shown in Fig. 1 [1], [2], [3].

There are several differences in multi-unit accidents compared to single unit accidents. One of the differences is a dependency between units. It is greatly related to common cause failure (CCF) models in multi-unit Level 1 PSA. Inter unit CCF models as well as intra unit CCF models should be considered in multi-unit Level 1 PSA [4,5]. Specifically, the dependencies and correlations between units are important in seismic multi-unit Level 1 PSA [6,7]. In addition to the CCF and the correlations, many important issues such as external hazards affecting multiple units, human reliability assessments, connected systems between units, cascading effects from damaged units to undamaged units are existed. Therefore, most of the researches performed until now focus on the multi-unit Level 1 PSA methodology [8], [9], [10]. Many STC combinations in multi-unit Level 3 PSA is the important problem, however, there are few researches on it [11,12]. The total number of the STC combinations increases exponentially as the number of units increases. This problem is shown in Fig. 2, and it is impossible and impractical to assess all the STC combinations. Multi-unit Level 3 PSA should be performed to assess a site risk, and it is necessary to develop a proper methodology considering many STC combinations. Therefore, an adequate methodology to group many STCs into fewer groups was developed and named as a binning methodology. The binning methodology was verified for multi-unit accidents in a reference site with two Optimized Power Reactor 1000 (OPR1000) units [13]. The reference site was selected among the four nuclear power plant sites in Republic of Korea.

Section snippets

Off-site consequence analysis model

Many computer codes for performing Level 3 PSA were developed in the world [14]. The most commonly utilized computer code is MELCOR Accident Consequence Code System (MACCS) developed in the Sandia National Laboratory (SNL). MACCS was originated from Calculation of Reactor Accident Consequence (CRAC) code utilized in the WASH-1400 [1]. MACCS was also utilized in the NUREG-1150 for assessing off-site consequences [2]. MACCS 3.11 version was utilized in this research, and the calculation flow in

Development procedure of the binning methodology

There are many STC combinations in multi-unit accidents, and this is the most important issue in multi-unit Level 3 PSA. Therefore, the binning methodology to group many STCs into fewer groups (GRPs) was developed in this research. The group is denoted by the GRP in the rest of this paper. A total number of consequence assessments can be greatly reduced by using the binning methodology, and this advantage is shown in Fig. 4. The binning methodology was developed and verified for the multi-unit

Verification procedure of the binning methodology

The binning methodology was verified for the multi-unit accidents in the reference site with two reference reactors [13]. The method 1, 2, 3, 4, and 5 were applied to the LT1, the LT2-EF, and the LT2-LF in the verification procedure. The conditional early and latent cancer fatality probability named as the population weighted risks in MACCS were selected for a consequence result type. The conditional early and latent cancer fatality probability were calculated for regions of 5 and 26 km,

Verification results of the binning methodology

The verification results of the step 4, 5, and 6 are presented in Section 5.1, 5.2, and 5.3, respectively. Five methods mentioned in Section 3.2 were used to designate the representative MACCS inputs of each GRP, and quantitative verification results of each method were calculated. Although this quantitative verification process to assess the appropriateness of each method should be needed, it is the most important task to qualitatively consider adequate factors affecting consequence results.

Limitations and future works

The binning methodology should be further developed with several adequate modification to be performed. The limitations and the future works related to those are suggested below.

The binning methodology basically has a large uncertainty. Many accident scenarios are grouped in Level 1 and 2 PSA because it is impossible and impractical to consider all the accident senarios individually. STCs resulted from these grouping tasks already have a large uncertainty. This uncertainty will increase by

Conclusions

The binning methodology was developed because it was impossible and impractical to calculate all STC combinations in multi-unit Level 3 PSA. The logic trees grouping many STCs into fewer GRPs were developed in the first part of the development procedure. Therefore, the qualitative logic tree (LT1) and the quantitative logic tree (LT2) were developed by considering the qualitative and quantitative headings, respectively. Two important considerations which were the amount of radioactive material

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgment

This work was supported by the Nuclear Safety Research Program through the Korea Foundation Of Nuclear Safety (KOFONS), granted financial resource from the Multi-Unit Risk Research Group (MURRG), Republic of Korea (No.1705001)

References (31)

  • D.W. Hudson et al.

    Multiunit Accident Contributions to quantitative health objectives: a safety goal policy analysis

    Nucl Technol

    (2017)
  • S.Y. Kim et al.

    Multi-unit level 3 probabilistic safety assessment: approaches and their application to a six-unit nuclear power plant site

    Nucl Eng Technol

    (2018)
  • W.J. Song

    Development of a multi-unit consequence analysis methodology, master thesis

    Hanyang Univ

    (2019)
  • Status of practice for level 3 probabilistic safety assessments

    NEA/CSNI/R(2018)1

    (2018)
  • D. Chanin et al.
    (1997)
  • Cited by (8)

    • An event sequence modeling method in multi-unit probabilistic risk assessment for high temperature gas-cooled reactor

      2023, Annals of Nuclear Energy
      Citation Excerpt :

      After the Fukushima Daiichi accident, the International Atomic Energy Agency (IAEA) conducted a series of studies on MUPRA (International Atomic Energy Agency (IAEA), 2012; International Atomic Energy Agency (IAEA), 2014; IAEA, 2014), and many MUPRA pieces of research have also been conducted (Liu et al., 2022; Jang et al., 2021; Kim et al., 2018; Cho et al., 2018; Kim et al., 2018; Song et al., 2020; Bixler and Kim, 2021; Le Duy et al., 2016; Korea Atomic Energy Research Institute, 2017; Kumar et al., 2015; Jung, 2018). These studies, by far, focus on different MUPRA aspects, such as the plant operating state analysis of multi-unit NPPs (Liu et al., 2022), the site risk assessment for internal events (Jang et al., 2021; Kim et al., 2018; Cho et al., 2018; Kim et al., 2018); and the source term analysis for multi-unit consequences (Song et al., 2020; Bixler and Kim, 2021), etc. Two main methods exist for multi-unit event sequence analysis—the single-top fault tree (SFT) and master event tree (MET) (Zhou et al., 2021).

    • Uncertainty analysis of source term and off-site consequence for WH600 using MELCOR and WinMACCS

      2022, Annals of Nuclear Energy
      Citation Excerpt :

      The consequences of the level 3 PSA and the accident frequencies from the level 2 PSA can be used to determine the risk of the accident at the nuclear power plant under evaluation. ( Kang and Jae, 2017; Song et al., 2020). Accident risk assessment of nuclear power plants involves uncertainties.

    • Exhaustive simulation approach for severe accident risk in nuclear power plants: OPR-1000 full-power internal events

      2022, Reliability Engineering and System Safety
      Citation Excerpt :

      However, one of the challenges for exhaustive simulation is how to make input data for a large number of accident scenarios with limited resources. For the simulation of NPP accident scenarios, integral severe accident computer codes such as MAAP5 (Modular Accident Analysis Program) [3,16–20] and MELCOR [4,21–24] are commonly used for various purposes. Such codes generally calculate mass, momentum, and energy conservation equations for water/steam behavior.

    • Multi-unit nuclear power plant probabilistic risk assessment: A comprehensive survey

      2021, Reliability Engineering and System Safety
      Citation Excerpt :

      Then one can approximate the consequence of each combination from the table. Song et al. [119] developed a logic-tree-based binning methodology that employs several headings to describe features of source terms, such as the start time of containment failure and the amount of radioactive materials released. Both qualitative and quantitative logical trees are developed depending on whether the early and latent cancer health effects need to be modeled separately.

    View all citing articles on Scopus
    View full text