Related Experiment Videos
Machine learning innovations for reliable gurney energy estimation in energetic materials
Mingyue Deng1,2, Raouf Hassan3, Alireza Baghban4
1School of Safety Engineering and Emergency Management, Nantong Institute of Technology, Nantong, Jiangsu 226002, China.
Iscience
|July 23, 2026
Summary
Predicting Gurney energy for energetic materials is now faster and more accurate using a machine learning framework. This data-driven approach bypasses traditional testing for rapid in silico screening and formulation design.
Area of Science:
- Materials Science
- Computational Chemistry
- Chemical Engineering
Background:
- Predicting Gurney energy (EG) for energetic materials is vital for industrial applications.
- Traditional methods involve costly empirical testing and complex thermochemical modeling.
- Existing methods are time-consuming and resource-intensive.
Purpose of the Study:
- To develop a robust, high-throughput machine learning framework for predicting EG.
- To enable direct prediction of EG from intrinsic molecular and thermodynamic descriptors.
- To provide an efficient alternative to traditional energetic materials characterization.
Main Methods:
- Data standardization and preprocessing.
- Optimization and evaluation of multiple machine learning algorithms.
- Artificial Neural Network (ANN) model development and validation.
Main Results:
- The ANN model achieved exceptional predictive accuracy (R2 = 0.995, MRD <1.5%).
- Key predictors identified include produced nitrogen gas (N2), oxygen balance, and solid carbon yield (C).
- Feature importance aligns with fundamental detonation physics principles.
Conclusions:
- The developed machine learning framework offers a highly accurate and efficient method for predicting EG.
- This data-driven approach facilitates rapid in silico screening of energetic materials.
- The findings support targeted formulation design for novel energetic materials.