Probability of failure as a function of sampling method for compressive strength

ABSTRACT

By using a large data set of point load tests from a uniform rock type, several different sampling methods were simulated to develop a range of sample mean and standard deviation estimates that each should describe a similar distribution of values, but differed as a result of the structure of the sampling methodology. These estimates were, in turn, used in a simulated stability analysis to demonstrate the effect of sample methodology on probability of failure and expected failure volume. By comparing the results of the stability analysis performed with the complete data set (which serves as a proxy for population statistics) to the results obtained from the sampled data set (which serves as a proxy for sample statistics), the robustness of several sampling methods can be evaluated.

KEYWORDS:

• Sampling methodology
• sample bias
• probability of failure
• expected failure volume

Acknowledgements

The authors would like to thank the Josephine Mining Company for their permission to use the data that formed the basis of this study. They would also like to thank that reviewers for their insightful comments on early versions on the manuscript. Their suggestions significantly improved the value of the work.

Disclosure statement

No potential conflict of interest was reported by the author(s).