Home / Current Issue / Paper 1719317
Nonlinear Prediction of Detonation Velocity in Aluminized Explosives Using Optimized Support Vector Regression
Subject area: Science,Engineering and Technology · Area of research: Machine Learning in Material Science
DOI: https://doi.org/10.64388/IREV9I12-1719317
Abstract
Accurate determination of the velocity of detonation (VoD) of aluminized explosives is essential for evaluating explosive performance in military, mining, and engineering applications. Conventional empirical and equation-of-state-based models often exhibit limited predictive capability due to the complex nonlinear interactions among explosive composition, density, oxygen balance, heat of explosion, and aluminum content. In this study, a Support Vector Regression (SVR) machine learning framework is developed for predicting the velocity of detonation of aluminized explosives. The proposed model utilizes key physicochemical descriptors of explosive formulations as inputs and establishes a nonlinear relationship between these descriptors and detonation velocity through kernel-based learning. The effects of important SVR hyper-parameters, including the regularization factor (C), epsilon-insensitive loss parameter (ε), and regularization parameter (λ), were systematically investigated to obtain optimal model performance. Predictive accuracy was evaluated using correlation coefficient (CC), root mean square error (RMSE), and mean absolute error (MAE). The developed SVR model was further compared with existing predictive approaches, including BHWS-EOS-full, BHWS-EOS-partial, and Keshavarz models. Results indicate that the optimized SVR model exhibits superior predictive accuracy, lower prediction errors, and stronger generalization capability than the conventional models. The demonstrated performance of the proposed framework highlights the potential of machine learning techniques as reliable and efficient tools for modeling detonation characteristics of complex explosive systems
Keywords
Velocity of Detonation, Aluminized Explosives, Support Vector Regression, Machine Learning, Energetic Materials, Predictive Modeling, Data-Driven Modeling, Nonlinear Regression, Explosive Performance, Artificial Intelligence
How to cite this paper
@article{1719317,
author = {James I. Agbi},
title = {Nonlinear Prediction of Detonation Velocity in Aluminized Explosives Using Optimized Support Vector Regression},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {12},
pages = {3307-3315},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1719317.pdf},
abstract = {Accurate determination of the velocity of detonation (VoD) of aluminized explosives is essential for evaluating explosive performance in military, mining, and engineering applications. Conventional empirical and equation-of-state-based models often exhibit limited predictive capability due to the complex nonlinear interactions among explosive composition, density, oxygen balance, heat of explosion, and aluminum content. In this study, a Support Vector Regression (SVR) machine learning framework is developed for predicting the velocity of detonation of aluminized explosives. The proposed model utilizes key physicochemical descriptors of explosive formulations as inputs and establishes a nonlinear relationship between these descriptors and detonation velocity through kernel-based learning. The effects of important SVR hyper-parameters, including the regularization factor (C), epsilon-insensitive loss parameter (ε), and regularization parameter (λ), were systematically investigated to obtain optimal model performance. Predictive accuracy was evaluated using correlation coefficient (CC), root mean square error (RMSE), and mean absolute error (MAE). The developed SVR model was further compared with existing predictive approaches, including BHWS-EOS-full, BHWS-EOS-partial, and Keshavarz models. Results indicate that the optimized SVR model exhibits superior predictive accuracy, lower prediction errors, and stronger generalization capability than the conventional models. The demonstrated performance of the proposed framework highlights the potential of machine learning techniques as reliable and efficient tools for modeling detonation characteristics of complex explosive systems},
keywords = {Velocity of Detonation, Aluminized Explosives, Support Vector Regression, Machine Learning, Energetic Materials, Predictive Modeling, Data-Driven Modeling, Nonlinear Regression, Explosive Performance, Artificial Intelligence},
month = {June},
doi = {https://doi.org/10.64388/IREV9I12-1719317}
}