Infill sampling design for tunnel rock classification

Tunneling companies reduce their uncertainty in predicting the geologic conditions through further investigation, usually with a fixed predefined budget. In gathering information through further exploration, there are several alternatives for spending exploration dollars. Each alternative requires a different cost and yields a different type of information, each with a different degree of certainty. In tunneling projects, situations lacking direct measurements are often encountered due to economic constraints. Instead, an appreciable amount of qualitative data generally exist. A statistical method for classification of a rock mass and infill sampling design was developed based on the indicator kriging formalism, which is a non parametric approach, and the cost of errors criterion developed for the binary classification by Aspie and Barnes. Math. Geol. 22 , 915–932 (1990) . For the multiple classification of a rock mass, the cost of error criterion was generalized and reformulated. As a case study, the application of this method is given with real tunnel data.