Models Development for Prediction of Blast Efficiency and Total Charge in a Typical Quarry

The prediction of blast efficiency is usually achieved by using models; this in turn, gives better and more efficient rock fragmentation. However, the accuracy of the prediction often times relies on the model development validation. In this study, models were developed and compared upon validation for predicting the blast efficiency and total charge required for efficient fragmentation using artificial neural network (ANN). Rock samples were gathered from the study are, and the uniaxial compressive strength (UCS) test was carried out on all the samples based on international standard. The average UCS obtained from the rock samples at the Eminent quarry (EQ) is 153.61 MPa. The dimension of in-situ rock mass considered in the study area is 60 m x 40 m, and the in-situ block sizes obtained vary from 2.02 m2 to 3.20 m2. The average percentage value of F50 obtained from the Split-Desktop image analyses is approximately 72.44 cm. The various results obtained from the UCS, in-situ block size distribution, image analysis of the blasted rocks and the total charge were used to develop the models for the prediction of blast efficiency. The key issue of concern about these models is that they are mostly site specific and the fact that if they perform well in a location does not guarantee the other. Hence, the validation and suitability of these models on the mine site. The blast efficiency prediction using ANN is compared with measured efficiency and the value of coefficient of determination, R2 obtained is 0.9733. The value of the coefficient of determination, R2 obtained from ANN by comparing the prediction of the total charge and the measured total charge is 0.9773. The findings showed that, the proposed ANN based mathematical models are suitable and thus, give better prediction to blasting efficiency and the possible total charge.