Value of information analysis for site investigation programs accounting for variability, uncertainty and scale effects with the Aspo HRL prototype repository as an example

摘要

An important feature of underground projects is the early site investigations, performed as a means to identify and quantify hazards. A methodology is presented for identifying the most cost-effective investigation program among a set of alternatives. Methodologies are presented for both investigation of thermal conductivity in hard rock and collection of rpck mechanic data for stress induced spalling problems. The cost-effectiveness of an investigation program is estimated by means of value of information analysis (VOIA). Each investigation program of thermal conductivity is associated with uncertainty due to natural variability and lack of knowledge. These uncertainties are taken into account in a simulation model with the aim to estimate the distribution of thermal conductivity values at different scales. The output is a set of thermal conductivity values from which a design parameter can be estimated. The simplest measure of the value of a site investigation is the expected reduction of uncertainty of the design parameter. The methodology is demonstrated with a case study for the prototype nuclear waste repository at Aspo Hard Rock Laboratory, Sweden. A set of four investigation programs for thermal conductivity were evaluated, and the most effective one identified. The application illustrates that an investigation program may supply very different value to a project, depending on how the objective of the investigation is defined. This is demonstrated by using two different objectives and comparing the results. Practical applications of the methodology on both thermal properties and rock mechanics are discussed, with emphasis on site investigations performed by the Swedish Nuclear Fuel and Waste Management (SKB).