A Comparative Analysis on the Performance of Modified Kuz–Ram and Kuznetsov–Cunningham–Ouchterlony Models on Small and Large Diameter Drill-Hole Blasts.

A comparative analysis on the performance of the modified Kuz–Ram model (MKM) and the Kuznetsov–Cunningham–Ouchterlony (KCO) model on a small diameter drill-hole dolomite quarry and large diameter granite quarry blasting operation was carried out. Blast data set for the quarries was collected and analysed using MKM and KCO models to ascertain the pre-blast particle-size distribution. The blast images were analysed and interpreted using the WipFrag software. The mean fragmentation size and percentage-passing sizes were calculated and the prediction accuracy of the two models' blast was compared with the measured results from WipFrag analysis to determine the prediction accuracy using four model performance indicators. Performance indices, such as variance account for (VAF), root-mean-square error (RMSE), mean absolute error (MAE), and coefficient of determination (R2), were used to check the prediction capacity of the models. The percentage-passing results showed that for both fines and the materials of desired range in Q1 and for fines in Q2, the MKM predicted values are closer to the WipFrag than the KCO model. For the materials of desired range in Q2, the KCO model predicted values are closer to the WipFrag measured values. There were no boulder predictions by both models for both Q1 and Q2. The error checkers of the particle mean size (X50) revealed that the KCO model is more efficient in predicting Q1 blast particle size as compared to Q2, while the MKM predicts more accurately than KCO model for Q2 blast rounds. A new modified Kuz–Ram model was developed for both Q1 and Q2 and the results are more closer to the results from the WipFrag image analysis and it also has higher coefficient of correlation as compared to Kuz–Ram model and existing MKM model. Highlights: Comparative analysis on the performance of modified Kuz–Ram model and Kuznetsov–Cunningham–Ouchterlony model on small and large diameter drill-hole blasts was carried out. The predicted values of both fines and materials of desired range from the percentage-passing distribution by the modified Kuz–Ram model are closer to the WipFrag values for small diameter drill holes. For fines, the modified Kuz–Ram model predicted values are closer to WipFrag values, and for the materials of desired range, the Kuznetsov–Cunningham–Ouchterlony model predicted values are closer to WipFrag values for the large diameter drill holes. The error values for means size predictions show that Kuznetsov–Cunningham–Ouchterlony model has a lower error rate in predicting mean size distribution for small diameter, while modified Kuz–Ram model is better for large diameter. The new modified Kuz–Ram model values for both the small and large diameter drill hole are more closer to the results from the WipFrag image analysis and have higher coefficient of correlation with lower error analysis values. [ABSTRACT FROM AUTHOR]