Prediction models have yet to reach levels of full comprehensive understanding, where they can predict any given situation with a high level of accuracy and reliability. Models are used to simplify the complex nature of blasting, but at this stage more research into the behaviour of blasting is still required to justify how a prediction model processes parameters, and determines fragmentation results.
The research completed makes use of the results collected from previously fired stopes at Gympie Eldorado Gold Mine (GEGM), in order to test the accuracy of Onederra’s (Mr Italo Onederra) fragmentation prediction model that has been incorporated into the software package SIMBLAST. The first step involved setting up each blast according to
specifications, and comparing the simulated results against the actual results. Secondly, the research tests the prediction model against what is theoretically expected in blasting. This was done by altering one parameter at a time such as, explosive density, hole diameter, and rock strength.
The results indicate that the behaviour of the Onederra prediction model is accurate and consistent with theory. A comparison of the results at GEGM versus simulated results shows a reasonable level of accuracy, and the model responds extremely well to theoretical results.
However the prediction model has been created from a limited amount of data, that was used as the basis in creating the modelling framework, and the model also includes some parameters which effect fragmentation, but the modelling theory behind these parameters cannot be justified by Onederra.