Skip to the main content

Original scientific paper

https://doi.org/10.17794/rgn.2026.2.9

PREDICTING PEAK PARTICLE VELOCITY INDUCED BY MINING BLASTING USING OPTIMIZED DEEP LEARNING WITH CORONAVIRUS HERD IMMUNITY ALGORITHM

Davood Mohammadi Sargini ; Faculty of Mining, Oil and Geophysics, Shahrood University of Technology
Mohammad Ataei ; Faculty of Mining, Oil and Geophysics, Shahrood University of Technology
Reza Mikaeil ; Faculty of Industries and Mining Technologies, Urmia University of Technology *
Akbar Esmaeilzadeh ; Faculty of Industries and Mining Technologies, Urmia University of Technology.

* Corresponding author.


Full text: english pdf 8.219 Kb

page 129-146

downloads: 67

cite


Abstract

One of the most important economic sectors is mining. Open-pit mines are the most common form, and blasting is the most efficient method. This method can cause significant damage due to vibrations. Thus, monitoring and controlling vibrations is important. This research aims to predict vibrations based on the distance and quantity of explosives. To study this, vibrations were estimated using Peak Particle Velocity from seismograph devices. Measurements were done in two large Iranian mines, providing approximately 1000 data points. Considering the data complexity, a deep learning method was selected for modelling. Deep learning has numerous hyperparameters, and their correct adjustment impacts efficiency. The Coronavirus Herd Immunity Optimizer (CHIO) was used to find optimal values. Various criteria were used to test the model efficiency. The results indicate the optimization method improved predictions by at least 4%. The coefficient of determination for random forest, basic deep learning, and optimized methods was 0.861, 0.932, and 0.972. The root means square error values are 3.687, 2.338, and 1.146. The absolute mean percentage error indices are 0.169, 0.128, and 0.068. It is concluded that deep learning has acceptable performance. Its efficiency can be increased by
optimization algorithms. The results show an excellent response to the integration of the CHIO method, demonstrating its good capability.

Keywords

blasting; peak particle velocity; deep learning; herd immunity; coronavirus

Hrčak ID:

345694

URI

https://hrcak.srce.hr/345694

Publication date:

13.3.2026.

Article data in other languages: croatian

Visits: 259 *