BOOTSTRAP METHODS FOR INVESTIGATING THE ROCK MASS ZONAL FRACTURE AROUND UNDERGROUND WORKINGS

Abstract

Objective. Investigation of the problems of zonal disintegration of rocks around deep underground workings in extremely small sampling, arising in geomechanical phenomena and processes in rock mass during mining. Methods. The primary tool used is numerical resampling methods, namely randomization, bootstrap, and Monte Carlo methods, which allow increasing the sample size, according to the available field data, to the required size for conducting a statistically sound analysis.Results. The problem of zonal fracture of rocks around deep underground workings is solved, for which the analytical dependence of the periodicity parameter of the defect function on the rock strength limit is estimated. The primary indicator of the statistical significance of the constructed model in work is the determination factor, based on which the type of analytical dependence under study is selected. Its deviation in the final model does not exceed 0.5% for any bootstrap sample volume, while in the other models considered in this paper, this value is achieved at n>200. Conclusion. The obtained interval estimates using bootstrap methods have a significant advantage over traditional approaches, increasing the reliability of the result in extremely small data samples without losing the level of significance.

Keywords

• бутстреп методы
• алгоритм
• зональное разрушение массива
• интервальные оценки параметров