Value of information analysis of site investigation program for slope design

Site investigation is widely used for uncertainty reduction and design improvement in slope engineering. Before the site investigation is implemented, it is important to assess and compare the effectiveness of different site investigation programs. In this paper, an efficient method for analyzing the effectiveness of a site investigation program for slope stability assessment through value of information analysis is proposed. In this method, a logistic regression model is developed to approximate the relationship between failure probability of the slope and the site investigation data, through which the evaluation of posterior distribution and slope reliability analyses are both avoided. The suggested method is illustrated and verified with an example. For the example studied in this paper, the number of slope stability analyses required is only 1% of that reported in the literature. The suggested method can be used to study the effects of locations of the boreholes, the number of boreholes, and the depth of boreholes on the effectiveness of a site investigation program.

Automated summary

The paper proposes an efficient method for analyzing the effectiveness of a site investigation program for slope stability assessment through value of information analysis. A logistic regression model is developed to approximate the relationship between failure probability of the slope and the site investigation data, avoiding the need for posterior distribution evaluation and slope reliability analyses. The method is illustrated and verified with an example, showing significant reduction in the number of slope stability analyses required compared to traditional methods.