Evaluating the Inclusion of the Implementation Variability into Fragmentation Modelling

摘要

Fragmentation modelling has been in use in the mining industry for a long time; however, there are still some aspects to tackle with regard to its use for strategic/planning and day-to-day purposes.On one hand, an important amount of resources are allocated to the task of calibrating and validating a good fragmentation model. Nonetheless, errors during the implementation stage introduce uncertainty into the prediction.On the other hand, it is common practice in the mining industry to use only discrete values to represent the fragmentation from blasting (k_(20') k_(50') k_(80') which provides a biased interpretation of the situation as, for instance, the best and worst possible cases are not taken into account. This is particularly critical when a continuous improvement philosophy has been adopted and modelling is used to either modify the drilling and blasting designs or plan ahead.It is well known that every site has particular conditions (rock mass, equipment, drilling and loading practices, among others) that often define unique trends in implementation. These trends can be captured using statistic and probabilistic techniques and then used as proxies for further analysis.This paper explores the potential of including implementation data in the form of probabilistic distributions into the Swebrec model to improve the performance of this engineering tool during the decision-making process.The fragmentation results of four designs were modelled using three different degrees of accuracy in the implementation for the burden, spacing and explosive column length.The data analysis showed that the probabilistic approach helps to achieve a broader perspective regarding the fragmentation that it is possible to obtain and the implications for the subsequent processes. The analysis also showed that the likelihood of producing similar fragmentation with different designs (involving different powder factors and costs associated) increases with the variability of the implementation.