Over the years, in the field of safety assessment of geological disposal system, system-level models have been widely employed, primarily due to considerations of computational efficiency and convenience. However, system-level models have their limitations when it comes to phenomenologically simulating the complex processes occurring within disposal systems, particularly when attempting to account for the coupled processes in the near-field. Therefore, this study investigates a machine learning-based methodology for incorporating phenomenological insights into system-level safety assessment models without compromising computational efficiency. The machine learning application targeted the calculation of waste degradation rates and the estimation of radionuclide flux around the deposition holes. To develop machine learning models for both degradation rates and radionuclide flux, key influencing factors or input parameters need to be identified. Subsequently, process models capable of computing degradation rates and radionuclide flux will be established. To facilitate the generation of machine learning data encompassing a wide range of input parameter combinations, Latin-hypercube sampling will be applied. Based on the predefined scenarios and input parameters, the machine learning models will generate time-series data for the degradation rates and radionuclide flux. The time-series data can subsequently be applied to the system-level safety assessment model as a time table format. The methodology presented in this study is expected to contribute to the enhancement of system-level safety assessment models when applied.

Conducting a TSPA (Total System Performance Assessment) of the entire spent nuclear fuel disposal system, which includes thousands of disposal holes and their geological surroundings over many thousands of years, is a challenging task. Typically, the TSPA relies on significant efforts involving numerous parts and finite elements, making it computationally demanding. To streamline this process and enhance efficiency, our study introduces a surrogate model built upon the widely recognized U-network machine learning framework. This surrogate model serves as a bridge, correcting the results from a detailed numerical model with a large number of small-sized elements into a simplified one with fewer and large-sized elements. This approach will significantly cut down on computation time while preserving accuracy comparable to those achieved through the detailed numerical model.

The most important thing in development of a process-based TSPA (Total System Performance Assessment) tool for large-scale disposal systems (like APro) is to use efficient numerical analysis methods for the large-scale problems. When analyzing the borehole in which the most diverse physical phenomena occur in connection with each other, the finest mesh in the system is applied to increase the analysis accuracy. Since thousands of such boreholes would be placed in the future disposal system, the numerical analysis for the system becomes significantly slower, or even impossible due to the memory problem in cases. In this study, we propose a tractable approach, so called global-local iterative analysis method, to solve the large-scale process-based TSPA problem numerically. The global-local iterative analysis method goes through the following process: 1) By applying a coarse mesh to the borehole area the size of the problem of global domain (entire disposal system) is reduced and the numerical analysis is performed for the global domain. 2) Solutions in previous step are used as a boundary condition of the problem of local domain (a unit space containing one borehole and little part of rock), the fine mesh is applied to the borehole area, and the numerical analysis is performed for each local domain. 3) Solutions in previous step are used as boundary conditions of boreholes in the problem of global domain and the numerical analysis is performed for the global domain. 4) steps 2) and 3) are repeated. The solution derived by the global-local iterative analysis method is expected to be closer to the solution derived by the numerical analysis of the global problem applying the fine mesh to boreholes. In addition, since local problems become independent problems the parallel computing can be introduced to increase calculation efficiency. This study analyzes the numerical error of the globallocal iterative analysis method and evaluates the number of iterations in which the solution satisfies the convergence criteria. And increasing computational efficiency from the parallel computing using HPC system is also analyzed.

To conduct numerical simulation of a disposal repository of the spent nuclear fuel, it is necessary to numerically simulate the entire domain, which is composed on numerous finite elements, for at least several tens of thousands of years. This approach presents a significant computational challenge, as obtaining solutions through the numerical simulation for entire domain is not a straightforward task. To overcome this challenge, this study presents the process of producing the training data set required for developing the machine learning based hybrid solver. The hybrid solver is designed to correct results of the numerical simulation composed of coarse elements to the finer elements which derive more accurate and precise results. When the machine learning based hybrid solver is used, it is expected to have a computational efficiency more than 10 times higher than the numerical simulation composed of fine elements with similar accuracy. This study aims to investigate the usefulness of generating the training data set required for the development of the hybrid solver for disposal repository. The development of the hybrid solver will provide a more efficient and effective approach for analyzing disposal repository, which will be of great importance for ensuring the safe and effective disposal of the spent nuclear fuel.

This paper intends to present considerations on the question of what is the “load standard” or “design load” for integrity evaluation under normal transportation conditions and what type of design load is good for users. This suggests a direction for subsequent research on producing design loads that transport business companies can utilize without difficulty. Several studies have been conducted to evaluate the integrity of spent nuclear fuel during normal transportation. A representative study recently conducted is the Multi-modal Transportation Test (MMTT) conducted using a commercial spent nuclear fuel cask by US DOE in 2017. In Korea, additional transport tests were planned to acquire sufficient test data under the conditions of road and sea transport considering the Korean situation. As a result, road transport tests were carried out in 2020 and sea transport tests were carried out in 2021. In the road transport test, a driving test that simulates various road conditions and a test that cycled a 4.5 km road eight times were performed. In most cases, the maximum acceleration of less than 1 g occurred, and the maximum strain was less than 48 με. For the sea transport test, the magnitude of both the maximum acceleration and the maximum strain were lower than those in the road transport test. We concluded tentatively that the integrity of spent fuel under normal conditions of transport was satisfactory with a large margin. However, when the storage business is realized and the transport of spent fuel becomes visible, the storage and transport business companies will have to prove the maintenance of the integrity of the spent fuel under normal transport conditions at the request of the regulatory agency. The transport business companies can transport the spent nuclear fuel by using different types of transport casks and different types of trucks and ships from those used in the tests mentioned above. However, it is absurd to have to prove the integrity of spent nuclear fuel by performing expensive tests again. Therefore, in this study, the design load that can be used by transport business companies is to be presented. The design load to be presented should satisfy the following requirements. The design load should be applicable including some differences in the transport cask or transport system, or different design loads should be presented according to the differences. The location where this design load is applied is to be specified (e.g. fuel rod, basket, internal structure). Requirements according to the operating speed of the transport system should be presented together. The type of design load is to be presented (e.g. PSD, SRS, FDS etc.). Other types of standards may be presented. For example, a speed limit for a vehicle carrying spent nuclear fuel may be suggested, or a speed limit for a vehicle passing through a speed bump may be suggested. In order to present such a reliable design load, a multi-axis vibration excitation shaker table test will be carried out. Though this shaker table test, the behavior of the nuclear fuel assembly is closely evaluated by applying the data obtained from the road and sea transport tests previously performed as an input load. In addition, FDS (Fatigue Damage Spectrum) will be produced and applied to experimentally evaluate the durability of fuel assemblies under normal transport conditions.

Currently, the development of evaluation technology for vibration and shock loads transmitted to spent nuclear fuel and structural integrity of spent nuclear fuel under normal conditions of transport is progressing in Korea by the present authors. Road transportation tests using surrogate spent nuclear fuel were performed in September, 2020 using a test model of KORAD-21 transportation cask and sea transportation tests were conducted from September 30 to October 4, 2021. In order to investigate amplification or attenuation characteristics, according to the load transfer path, a number of accelerometers were attached on a ship cargo hold, cradle, cask, canister, disk assembly, basket, and surrogate fuel assemblies and to investigate the durability of spent nuclear fuel rods, strain gages were attached on surrogate fuel assemblies. A ship named “JW STELLA” which has similar deadweight (5,000 ton) of existing spent nuclear fuel transportation ships was used for the sea transportation tests. The ship is propelled by 1,825 hp two main engines with two 4-bladed propellers. There are two major vibration sources in the ship. One is the vibration from waves and the other is the vibration from the engine and propeller system. The sensor locations on the ship were determined considering the vibration sources. The sea transportation test was performed for 5 days, the test data were measured successfully. The ship with the test model was departed from Changwon and sailed to Uljin, sailed west to Yeonggwang and then returned to Changwon. In addition to sailing on a designated test route, circulation test, braking/acceleration test, depth of water test, and rolling test were conducted. As a result of the preliminary data analysis of the sea test, power spectral densities and shock response spectrums were obtained according to the different test conditions. The vibratory loads caused by the wave mainly occurred in the frequency range of 0.1 to 0.3 Hz. The vibratory loads caused by the propeller occurred near the n/rev rotating frequencies, such as 5, 10, 20 Hz etc. However, those frequencies are far from the natural frequencies of local mode of the fuel rods, so it is considered that the vibratory loads from the wave and the propeller do not have a significant influence on the structural integrity of the fuel rods. Among all the test cases, maximum strain occurred at SG31 near the bottom nozzle on the test; the magnitude was 73.62 micro strain. Based on the analyzed road and sea transportation test data, a few input spectra for the shaker table test will be obtained and the shaker table test will be conducted in 2022. It is expected that the detailed vibration characteristics of the assembly which were difficult to identify from the test results can be investigated.

The amount of temporarily stored spent nuclear fuel in South Korea will be reaching saturation in a near future. Therefore, it is an urgent issue to construct a spent nuclear fuel storage system. In order to construct the storage system, some coastal environmental characteristics such as temperature, pH, and chemical composition of sea water in South Korea have to be evaluated and predicted because they can affect in deterioration of the storage system. However, in South Korea, the coastal environmental characteristics of area where the storage system is likely to be built are not well established until now. In this study, a time-series deep-learning algorithm is developed using the Long-Short Term Memory (LSTM) algorithm to predict and evaluate the coastal environmental characteristics based on the wellestablished data from Korea Meteorological Administration (KMA) and Ministry of Oceans and Fisheries (MOF). As a result, by developing the predictive model to evaluate the coastal environmental characteristics, we intend to apply it for site evaluation to construct the spent nuclear fuel storage system or many other applications related to the nuclear as well.