In preparation for actual repository implementation, designs that could be tailored to specific sites need to be considered. This requires a paradigm shift in design philosophy, moving away from a single reference design towards a flexible and systematic “catalogue” of engineered barrier components and associated repository infrastructure. To support this development, novel methodology for both developing and comparing repository design options for Japanese vitrified high-level radioactive waste (HLW) has been examined. This paper will outline the background of the project and, in particular, some novel approaches examined for facilitating optimisation of practical repository designs. It is complemented by a companion paper - Matsumoto et al: Application of Formal Knowledge Engineering Approaches to Develop A Design Catalogue for A Japanese HLW Repository -, which describes the repository design process in more detail and illustrates results of first dry runs of the methodology.

A sensitivity analysis was conducted to evaluate the impact of silicon migration through buffer material on the lifetime of vitrified waste. The results indicate that the lifetime depends on a combination of the dissolved glass fraction in the non-steady-state phase controlled by the silicon pore diffusion coefficient (Dp) and the silicon distribution coefficient (Kd) in the buffer material and the steady-state dissolution rate defined by Dp and the groundwater flow rate (Q) in the excavation disturbed zone. In the case where the glass dissolution rate reaches the steady-state dissolution rate, the sensitivity of the lifetime to Dp and Q varies according to the magnitude relationship between Dp and Q. We also discuss the impact on the lifetime of glass hydration, which proceeds simultaneously with glass matrix dissolution. The results show that glass hydration is less important for the lifetime than glass matrix dissolution in an open system and it can be concluded that silicon migration through the buffer material will be an important process for estimating the lifetime of the vitrified waste. A preliminary calculation of the long-term waste behavior with realistic assumptions indicates the importance of the silicon migration parameters Kd and Dp, which control the dissolution behavior of the vitrified waste in the non-steady-state phase, for evaluating radionuclide release.

The volunteer siting of disposal facilities for vitrified high-level radioactive waste (HLW) and TRU waste in Japan results in a particular need for flexibility to allow repositories to be tailored to specific sites. Instead of a single reference concept, a “catalogue” of designs for individual repository components is being developed, which can then be combined to form optimum solutions for particular site boundary conditions. As highlighted in a companion paper – Makino et al: Supporting Development of Practical Designs for a Japanese HLW Repository –, which puts the repository design process in the context of the Japanese HLW programme, the complexity of this work justifies use of advanced Knowledge Engineering approaches. In this paper three components are described – development of the knowledge base, assessing tacit system understanding and production of innovative solutions to resolve conflicts between requirements.

The H12 performance assessment (PA) provided a test for the robustness of a HLW repository system concept based on structured siting and design, taking account of a wide range of potentially suitable Japanese geological environments. The generic nature of the host rock in the H12 assessment means, however, that emphasis is placed verymuch on strong EBS performance. The assessment included a comprehensive evaluation of uncertainty and potentially detrimental factors, including perturbations due to external events and processes. Despite the considerable uncertainty at the current stage of the Japanese program, a safety case that is adequate for the aims of the assessment can be made by a strategy of employing conservatism where there is uncertainty and stressing the reliability and effectiveness of the performance of the near-field. The aim of this paper is to present the H12 PA in a way which makes the PA process clearer and the implications of the results more meaningful, both to workers within the PA field and to a wider technical audience.