Empirical and data-driven blast fragmentation prediction considering discontinuity orientation: A case study at Elkem Tana mine
Accurate prediction of rock fragmentation is important for optimizing blast design and downstream mining operations. This study compares empirical fragmentation models with a site-specific data-driven approach for predicting mean fragment size ( X 50 ) at the Elkem Tana quartzite mine in Norway. Ten production blasts were analyzed using three-dimensional drone photogrammetry. Four empirical approaches were evaluated: KuzRam, Kuznetsov–Cunningham–Ouchterlony (KCO), Chung and Katsabanis (C&K), and Modified C&K. The rock factor ( A ) was estimated using the Lilly Blastability Index and calibrated using measured fragmentation data. An ElasticNet model was developed using blast geometry, powder factor, characteristic in-situ block size, and an angular-difference (AD) parameter describing the angular contrast between the two dominant discontinuity sets relative to the bench face. Uncalibrated empirical models produced mean absolute percentage error (MAPE) values of 70.9% for KuzRam/KCO and 167% for C&K/Modified C&K. Calibration improved performance, but the best calibrated empirical model resulted in an Root Mean Square Error (RMSE) of 27.1 mm. ElasticNet achieved a leave-one-out cross-validation R 2 of 0.934, RMSE of 7.93 mm, and MAPE of 2.7%. Bench height-to-burden ratio produced the largest variation in predicted X 50 , followed by AD, demonstrating the potential of combining site-specific blast parameters and rock-mass structural descriptors for improved fragmentation prediction.
Authors
- Hakan Başarır (ORCID: https://orcid.org/0000-0003-4556-4953)
- Zahir Ahmed (ORCID: https://orcid.org/0009-0008-1015-590X)
Institutions
- Norwegian University of Science and Technology (NO)
Publication Details
- Journal
- Mining Technology Transactions of the Institutions of Mining and Metallurgy
- Published
- 2026-11-18
- DOI
- https://doi.org/10.1177/25726668261506307
- Primary Topic
- Blasting Impact and Analysis
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Elkem