Value of Information Analysis in Rock Engineering Investigations

Doctoral thesis, 2014

A challenge in rock engineering is to strike a balance between a reasonable level of investigation and uncertainty in the geological conditions. Value of Information Analysis (VOIA) provides support for decisions about investigation strategies. The purpose of the method is to assess the economic value of an investigation before it has actually taken place. The overall aim of this study was to use VOIA to assess the value of prior investigations for underground construction projects. The thesis has focused on tunnelling projects in crystalline rock where the main rock engineering uncertainties are related to the anisotropy, heterogeneity and scale-dependent properties of the rock mass. The specific aims of the thesis were to develop and adopt VOIA as a basis for decisions regarding investigations in underground construction and to test the method under realistic conditions with real project costs. A further aim was to investigate how expert judgement can be incorporated transparently into VOIA. Three studies of VOIA with different levels of complexity have been conducted. The input probabilities were assessed by means of stochastic modelling of the rock mass or through expert knowledge. The method was tested in a case study of a tunnel at the Äspö Hard Rock Laboratory, Oskarshamn, Sweden, where the probabilities were assessed by experts as part of a structured elicitation procedure. VOIA models quickly become large and complex although the method provides useful input for uncertainty assessments and risk analysis of limited decision problems. One of the strengths of the method is that expert judgement can be incorporated stringently into the calculations. A recommended simplification of VOIA is to limit the analysis to calculation of the expected value of perfect information, which produces an upper boundary of the maximum value that can be obtained by any investigation. This value can be practically used as a benchmark when determining whether it is worth carrying out investigations or not.