ABSTRACT A new procedure to assess in-situ rock stress useful for early stages of rock engineering projects is presented. Relationships between K (σ H/σ v) values obtained from instrumental measurements and TSI index (Tectonic Stress Index) have allowed to define empirical relationships from which K can be estimated. To account regional and local influencing factors in the far and near stress fields decision tree analysis are applied. The combined application of TSI and decision tree procedures can provide a reasonable estimation of the expected magnitude of horizontal stress. This methodology has been applied to a large number of cases, mainly from Europe.
1 INTRODUCTION A new procedure to assess magnitudes of horizontal stress in terms of K values (σ H/σ v) is presented. The procedure is based on application of decision tree analysis (DTA) and the empirical relationship between the TSI index (Tectonic Stress Index) and K. Both methods,DTA and TSI-K, are complementary and allow integration of the different factors affecting the state of stress. InDTAgeodynamic and geophysical factors are considered to account stress magnitudes at regional and local scale.The result allows an estimation of the horizontal stresses magnitudes in terms of very high, high, intermediate or low, as well as possible effects of local amplification. The index TSI considers geological history of the rock, elastic modulus of rocks and maximum lithostatic load supported. Correlations from a large database between TSI and K values have been obtained. K-TSI empirical relationships have provided estimations of expected magnitudes of horizontal stress.
2 DECISION TREE ANALYSIS APPLIED TO TECTONIC HORIZONTAL STRESS ASSESMENT Logic and decision tree methods are currently used in probabilistic analysis. The decision tree is comprised of branches and nodes. Each node matchs up with influential factors in the state of stress. Main regional and local factors influencing stress have been selected from Gonzalez deVallejo, et al., 2006. Two decision tree analysis (DTA) have been developed. The first one (DTA 1, Fig. 1) considers regional scale influencing factors and determineswhether foreseeable horizontal stresses are low, intermediate, high or very high. We defined stresses as low if their magnitude was less than 10MPa, intermediate if they were 10 to 25MPa, high if they were 25 to 40MPa and very high if they were more than 40MPa; in each case for depths less than 2000 m. The second decision tree (DTA 2, Fig. 2) represents local scale influencing factors and establishes whether stress amplification due to local effects could occur. The minimum degree of confidence adopted for this analysis was 60%. Table 1 shows the factors included in decision tree analysis 1 and 2 that allow estimation of the relativemagnitudes of the horizontal stresses that might be expected.