Using dimensional-regression analysis to predict the mean particle size of fragmentation by blasting at the Sungun copper mine

A methodology was founded on the basis of a dimensional analysis procedure, together with multivariate nonlinear regression analysis which is used to predict mean particle size of fragmentation at the Sungun surface mine. A practical database was made through a number of blasting operations in various levels of the mine with geomechanical investigations and experimental tests. The mean particle size is first considered to be a function of various controllable and uncontrollable variables. Then by setting up a nonlinear correlation among the independent dimensionless products obtained from the dimensional analysis, a fundamental equation has been deduced. The equation can be practically used by mining engineers in all situations where the mean particle size of fragmentation by bench blasting should be predicted. Capability of the proposed method is determined by comparing its predictions with the real measurement (observation) of sieve analysis, a standardized image-processing technique, and the predictions by the Kuz-Ram model as the most applied model, together with the modified version of the Kuz-Ram model. The results obtained from the proposed method are closer to the real results of the sieve analysis than those of the other methods. Finally, the methodology was found to be strong and better in prediction than in image processing and also in both versions of the Kuz-Ram model.