1 PAPER NUMBER 32

Recent Safety Assessment of a Reference Geological Disposal System for Radioactive Waste from Pyro-processing in Korea

J.-W. Kim1, D.-K. Cho1, J. Jeong1, M.-H. Baik1, K. Kim1

1 Korea Atomic Energy Research Institute (KAERI), Daejeon, Korea

E-mail contact of main author: jw_kim@kaeri.re.kr

Abstract. For a long-term safety assessment to be comprehensive, complex scenarios should be assessed systematically by combining various scenarios with aleatory uncertainty. A methodology for a risk-based safety assessment of complex scenarios considering the long-term complementary impacts on the disposal system has been newly suggested by KAERI. This new methodology was recently implemented in an upgraded version of KAERI’s TSPA model (K-PAM). KAERI’s current TSPA model contains many necessary abstractions and a limit in associating the key physical processes. As a further study, the TSPA model will be moved to the process model level by utilizing a high-performance computing system.

Key Words: Complex scenario; Risk-based safety assessment; K-PAM.

1. Introduction

Since 2007, the Korea Atomic Energy Research Institute (KAERI) has studied the geological disposal of radioactive waste generated from the pyro-processing of PWR spent nuclear fuel [1]. The study mainly includes the characterization of geological media, the design of a reference disposal system, and the overall safety assessment of the disposal system. The characterization of geological media at different scales has been mainly conducted at the KAERI Underground Research Tunnel (KURT) area, the host rock of which is granite. The conceptual design of the reference disposal system is basically based on the Swedish KBS-3 concept. For the safety assessment of a hypothetical disposal system, a total system performance assessment (TSPA) model was developed using GoldSim. For a long-term safety assessment to be comprehensive, complex scenarios should be assessed systematically by combining various scenarios with aleatory uncertainty. In this study, a methodology for a risk-based safety assessment of complex scenarios considering the long-term complementary impacts on the disposal system is presented and implemented in an upgraded version of KAERI’s TSPA model (K-PAM). For an illustration, a statistical analysis of historical seismic events and well exploitation in Korea was utilized to generate a complex scenario for a risk-based safety assessment.

2. Reference Disposal System

KAERI presented a preliminary conceptual design of a geological disposal system for the radioactive wastes generated from the pyro-processing of PWR spent nuclear fuel. The radioactive wastes were classified into two groups: (1) low & intermediate-level metal waste which consists of hull materials and support frames and (2) high-level ceramic waste which is vitrified molten salt from the electrowinning process. The metal wastes are emplaced in a

2 PAPER NUMBER 32

storage canister (stainless steel) and then the storage canisters are packaged by polymer concrete, so-called metal waste disposal package (MWDP). MWDPs are supposed to be stacked up with buffer materials in the tunnel at 200 m depth. The ceramic wastes are emplaced in a storage canister (stainless steel) and then the storage canisters are packaged in a disposal canister which consists of an inner container for the structural stability and an outer shell for corrosion resistance. The disposal canisters are supposed to be emplaced with buffer materials in the borehole at 500 m depth (FIG. 1).

< Reference Disposal System >

< Metal Waste >

< Ceramic Waste >FIG. 1. Conceptual design of a geological disposal system presented by KAERI.

3. Risk-based Safety Assessment

3.1. K-PAM Methodology

The risk-based safety assessment methodology consists of 5 steps as shown in FIG. 2.

The external events include natural disruptive events, such as earthquake, etc., and human intrusion. In the 1st step, the properties of those events related to the performance of the disposal system are digitized and represented by probability density functions (PDFs). In the case of an earthquake, for example, the properties can be the event occurrence rate, magnitude, distance from the hypocenter, etc. The PDFs of each property have to be carefully determined based on the historical records, a statistical analysis, expert judgments, etc. The PDFs of each property are converted into cumulative density functions (CDFs) for a scenario combination.

In the 2nd step, how the external events will affect the disposal system is defined and the complex scenario generation criteria are determined. The external events will discriminatorily affect each part of the disposal system, such as an engineered barrier system (EBS), natural barrier system (NBS), and the biosphere. The impacts on the disposal system are also dependent on the properties of the external events. Some impacts can be irreversible so that the influence continues during the period of assessment, and some impacts can be reversible so that the disrupted parts are recovered after some time, or the influence of some repeating impacts can be increased gradually. This process also has to be carefully conducted based on the analogical interpretations of the experimental results and the relevant field data.

3 PAPER NUMBER 32

In the 3rd step, a complex scenario is generated based on the criteria defined in the previous step. Monte-Carlo sampling method is utilized as random numbers are independently generated and converted into the occurrence times and/or the values of the properties using the predefined CDFs for each property of the external events. The types of impacts by the external events are then determined based on the criteria. As all impacts on the disposal system are arranged in the process of time, a complex scenario is finally completed. For every iteration, a new complex scenario is preliminarily generated through this step.

In the 4th step, each complex scenario developed in the 3rd step is simulated using the user-defined TSPA model. As the results of the scenario assessments, the exposure dose rates to the representative person are computed for each scenario. Because the complex scenario was randomly generated based on the criteria and their probabilities, the resulting exposure dose rates already involve the probability of the scenario. In other words, if an exposure dose rate is obtained often from the iterations of scenario assessments, it implies that the scenario related to the exposure dose rate has a high occurrence probability. After each iteration, the exposure dose rates are cumulatively averaged and converted into the total risk using a dose-to-risk conversion factor. As the number of iterations increases, the results will be statistically stabilized. If the difference between the risks calculated in each iteration is less than the user-defined convergence criteria, it is assumed that the number of iterations is sufficient to consider exhaustively all possible scenarios.

In the final step which is a post-process step, the final risk, the occurrence probabilities of each scenario, and the complementary safety indicators are computed as ordered. Additionally, sensitivity analysis can also be conducted in this step.

3.2. K-PAM Modeling System

The methodology above was numerically implemented in an upgraded version of KAERI’s TSPA model (K-PAM). The overall computing steps in FIG. 2 are conducted using Matlab except the scenario assessment (4th step) which is conducted using GoldSim. That is, a Matlab-based overall computing system is equipped with a scenario assessment module which was developed using GoldSim (FIG. 3). The GoldSim-based safety assessment model explains the source term, radionuclide transport in the EBS and far-field host rock (NBS), and radionuclide transfers in the biosphere. The radionuclide transport in the EBS includes the radionuclide release from a MWDP (metal waste) or disposal canister (ceramic waste), diffusive transport through buffer material, sorption, precipitation, and radioactive decay in

FIG. 2. Flowchart of risk-based safety assessment.

4 PAPER NUMBER 32

the EBS. In the far-field host rock, radionuclide transports through the fractured rock undergoing sorption, precipitation, matrix diffusion, and radioactive decay.

102 103 104 105 10610-10

10-9

10-8

10-7

10-6

10-5

10-4

10-3

10-2

10-1

100

Dos

e (m

Sv/

yr) /

Ris

k (/y

r)

Time (yr)

FIG. 3. GoldSim-based TSPA model and an illustrative result of the risk-based safety assessment.

3.3. Illustration

An illustrative result of the risk-based safety assessment is depicted in FIG. 3. In the illustration, two external events, earthquake and well intrusion, were considered in the complex scenario generation. From the results, the computation was successfully converged into less than 1% risk-change after about 400 iterations. The time-series of dose for each iteration are depicted with gray line, and the median dose and the risk are depicted with black and red lines, respectively.

4. Concluding Remarks

A methodology and a modeling system (K-PAM) for a risk-based safety assessment of complex scenarios considering the long-term complementary impacts on the disposal system were developed in this study. The results reasonably confirm the efficiency and stability of the modeling system. From the risk-based safety assessment of complex scenarios, the reliability, safety and public confidence of the disposal system is expected to be convinced more efficiently.

KAERI’s current TSPA model contains many necessary abstractions and a limit in associating the key physical processes. As a further study, the TSPA model will be moved to the process model level by utilizing a high-performance computing system.

5. References

[1] KOREA ATOMIC ENERGY RESEARCH INSTITUTE, Geological Disposal of Pyroprocessed Waste from PWR Spent Fuel in Korea, KAERI/TR-4525/2011, KAERI, Korea (2011).