Abstract
This thesis is concerned with the issue of rockburst damage in hard rock, mechanised underground mines. Specifically, given that a rockburst occurs, it quantifies the significance of the factors affecting the severity of damage, or how the variation in each contributing factor influences the likelihood of exceeding a certain damage severity.
A literature review was undertaken and it was shown that there was generally more than one factor controlling the stability of an excavation under dynamic loading. It was also shown that some factors are more significant than others at controlling the severity of damage. The lack of quantification on the weighting of factors controlling the stability of excavations was also seen in the Rockburst Damage Potential (RDP) method (Heal 2010). The Australian Centre for Geomechanics rockburst database was used to assess the weighting of factors controlling rockburst damage severity.
Factor analysis was first applied to explore the structure of the database, or how all variables associated with each damage case correlated with each other so that data reduction could be carried out. Stress factor, ground support, excavation span, geology factor, wall/back and Peak Particle Velocity (PPV) were determined to be potential factors controlling excavation stability. Rock Damage Scale (RDS) was selected as the most appropriate parameter to quantify damage severity.
Rock damage scale was regrouped into three categories based on its perceived operational implications: acceptable (R2), tolerable (R3) and intolerable (R4 and R5). Logistic regression was then used to model the probability of exceeding R2 and R3 severity damage. It was determined that ground support, geology factor and PPV are significant factors when predicting the probability of exceeding R2 severity damage. Stress factor and excavation span are additional significant factors predicting the probability of exceeding R3 severity damage.
The area under the receiver operating characteristic (ROC) curve was used in this study as a quantitative measure to assess the performance of the new regression models in comparison to the original RDP model. The inclusion of factor weighting for the new model resulted in a 6.3 and 8.3% improvement for predicting a pair of positive and negative cases for exceeding R2 and R3 severity damage respectively over the original RDP models.
Following the goodness of fit test, the adjusted RDP method was applied to a series of rockburst case studies to assess its performance in a range of mining conditions and to assess how generalisable this method is at predicting rockburst damage. The performance of the adjusted RDP method during the case study was deemed to be adequate, although several issues were encountered and discussed.
A literature review was undertaken and it was shown that there was generally more than one factor controlling the stability of an excavation under dynamic loading. It was also shown that some factors are more significant than others at controlling the severity of damage. The lack of quantification on the weighting of factors controlling the stability of excavations was also seen in the Rockburst Damage Potential (RDP) method (Heal 2010). The Australian Centre for Geomechanics rockburst database was used to assess the weighting of factors controlling rockburst damage severity.
Factor analysis was first applied to explore the structure of the database, or how all variables associated with each damage case correlated with each other so that data reduction could be carried out. Stress factor, ground support, excavation span, geology factor, wall/back and Peak Particle Velocity (PPV) were determined to be potential factors controlling excavation stability. Rock Damage Scale (RDS) was selected as the most appropriate parameter to quantify damage severity.
Rock damage scale was regrouped into three categories based on its perceived operational implications: acceptable (R2), tolerable (R3) and intolerable (R4 and R5). Logistic regression was then used to model the probability of exceeding R2 and R3 severity damage. It was determined that ground support, geology factor and PPV are significant factors when predicting the probability of exceeding R2 severity damage. Stress factor and excavation span are additional significant factors predicting the probability of exceeding R3 severity damage.
The area under the receiver operating characteristic (ROC) curve was used in this study as a quantitative measure to assess the performance of the new regression models in comparison to the original RDP model. The inclusion of factor weighting for the new model resulted in a 6.3 and 8.3% improvement for predicting a pair of positive and negative cases for exceeding R2 and R3 severity damage respectively over the original RDP models.
Following the goodness of fit test, the adjusted RDP method was applied to a series of rockburst case studies to assess its performance in a range of mining conditions and to assess how generalisable this method is at predicting rockburst damage. The performance of the adjusted RDP method during the case study was deemed to be adequate, although several issues were encountered and discussed.
Original language | English |
---|---|
Qualification | Masters |
Awarding Institution |
|
Supervisors/Advisors |
|
Award date | 27 Jun 2016 |
Publication status | Unpublished - 2016 |