Establishing a Bayesian Network model for predicting rockburst damage potential

K Maxutov, Amoussou Coffi Adoko

2021

Full text Open access: Gold Metadata (CSL-JSON)
DOI 10.1088/1755-1315/861/6/062094
Data confidence 70 medium
How this score is calculated

It is built from checkable properties of the record itself: whether the article has a DOI, how many independent sources confirmed its data, whether the required fields are filled in, and how recent the latest confirmation is. Disagreement between sources and a missing identifier lower it. A value of 80 or above means the record was published without further checking; 60 to 79 means it is published at reduced confidence, with a manual check advisable. A record verified by an editor by hand is never pushed below 90 by a recalculation. This scores our data about the article, not the article itself.

Abstract

In underground mines, rockburst can cause significant damages to underground excavations and equipment, which impacts negatively mine safety and productivity. Hence, the establishment of model capable of predicting the rockburst damage in a reliable manner is essential. Despite the existence of the extensive number of predictive models for rockburst, a reliable prediction of rockburst damage potential is challenging. One of the main difficulties to predict rockburst with existing models is their inability to account for missing or incomplete data, which is commonly encountered in practice. To overcome this issue, in this paper, a Bayesian network (BN) approach is employed to develop a short-term prediction model for rockburst on the basis of mine seismicity monitoring data. A rockburst database consisting of 254 case histories was compiled. The input parameters include the stress conditions, ground support capacity, excavation span, effect of geological structure, peak particle velocity, and the output defined as the rockburst damage scale (RDS) index. The Tree Augmented Naïve Bayes classifier structure was implemented. The proposed BN was tested using 5-fold cross-validation. The practical use of the BN was illustrated as well. Overall, the BN results indicated acceptable prediction accuracies ranging from 70 to 78%. It was concluded that BN could represent a reliable predictor to control excavation damages due to mine seismicity, especially when incomplete data prevail.