Study on the prediction and inverse prediction of detonation properties based on deep learning DOI Creative Commons

Zi-hang Yang,

Jili Rong, Zitong Zhao

и другие.

Defence Technology, Год журнала: 2022, Номер 24, С. 18 - 30

Опубликована: Дек. 1, 2022

The accurate and efficient prediction of explosive detonation properties has important engineering significance for weapon design. Traditional methods predicting performance include empirical formulas, equations state, quantum chemical calculation methods. In recent years, with the development computer deep learning methods, researchers have begun to apply performance. method advantage simple rapid properties. However, some problems remain in study based on learning. For example, there are few studies mixed explosives, parameters equation state application predict formulation explosives. Based an artificial neural network model a one-dimensional convolutional model, three improved models were established this work aim solving these problems. training data models, called JWL (EOS) inverse was obtained through KHT thermochemical code. After training, tested overfitting using validation-set test. Through model-accuracy test, accuracy real formulations by comparing predicted value reference value. results show that errors within 10% 3% pressure velocity, respectively. refers which consist TNT, RDX HMX. correlation coefficient between curves above 0.99. error 9%. This indicates utility engineering.

Язык: Английский

Towards the prediction of drug solubility in binary solvent mixtures at various temperatures using machine learning DOI Creative Commons
Zeqing Bao, Gary Tom,

Austin Cheng

и другие.

Journal of Cheminformatics, Год журнала: 2024, Номер 16(1)

Опубликована: Окт. 28, 2024

Abstract Drug solubility is an important parameter in the drug development process, yet it often tedious and challenging to measure, especially for expensive drugs or those available small quantities. To alleviate these challenges, machine learning (ML) has been applied predict as alternative approach. However, majority of existing ML research focused on predictions aqueous and/or at specific temperatures, which restricts model applicability pharmaceutical development. bridge this gap, we compiled a dataset 27,000 datapoints, including molecules measured range binary solvent mixtures under various temperatures. Next, panel models were trained with their hyperparameters tuned using Bayesian optimization. The resulting top-performing models, both gradient boosted decision trees (light boosting extreme boosting), achieved mean absolute errors (MAE) 0.33 LogS (S g/100 g) holdout set. These further validated through prospective study, wherein four predicted by then in-house experiments. This study demonstrated that accurately solutes different whose features closely align within (MAE < 0.5 LogS). support future facilitate advancements field, have made code openly available. Scientific contribution Our advances state-of-the-art predicting leveraging uniquely comprehensive dataset. Unlike studies predominantly focus solvents fixed our work enables prediction variety over broad temperature range, providing practical insights modeling realistic applications. along open access significant steps process new molecule discovery, analysis formulation. Graphical

Язык: Английский

Процитировано

6
Prediction and Construction of Energetic Materials Based on Machine Learning Methods DOI Creative Commons
Xiaowei Zang, Xiang Zhou, Haitao Bian

и другие.

Molecules, Год журнала: 2022, Номер 28(1), С. 322 - 322

Опубликована: Дек. 31, 2022

Energetic materials (EMs) are the core of weapons and equipment. Achieving precise molecular design efficient green synthesis EMs has long been one primary concerns researchers around world. Traditionally, advanced were discovered through a trial-and-error processes, which required research development (R&D) cycles high costs. In recent years, machine learning (ML) method matured into tool that compliments aids experimental studies for predicting designing EMs. This paper reviews critical process ML methods to discover predict EMs, including data preparation, feature extraction, model construction, performance evaluation. The main ideas basic steps applying analyzed outlined. state-of-the-art about applications in property prediction inverse material is further summarized. Finally, existing challenges strategies coping with proposed.

Язык: Английский

Процитировано

20
Machine learning-assisted quantitative prediction of thermal decomposition temperatures of energetic materials and their thermal stability analysis DOI Creative Commons
Zhixiang Zhang, Yilin Cao, Chao Chen

и другие.

Energetic Materials Frontiers, Год журнала: 2023, Номер unknown

Опубликована: Сен. 1, 2023

In this study, machine learning (ML)-assisted regression modeling was conducted to predict the thermal decomposition temperatures and explore factors that correlate with stability of energetic materials (EMs). The performed based on a dataset consisting 885 various compounds using linear nonlinear algorithms. tree-based models established demonstrated acceptable predictive abilities, yielding low mean absolute error (MAE) 31°C. By analyzing through hierarchical classification, study insightfully identified affecting EMs' temperatures, overall accuracy improved targeted modeling. SHapley Additive exPlanations (SHAP) analysis indicated descriptors such as BCUT2D, PEOE_VSA, MolLog_P, TPSA played significant role, demonstrating process is influenced by multiple relating composition, electron distribution, chemical bond properties, substituent type molecules. Additionally, Carbon_contents Oxygen_Balance proposed for characterizing EMs showed strong correlations temperatures. trends their SHAP values most suitable ranges were 0.2–0.35 −65 −55, respectively. Overall, shows potential ML temperature prediction provides insights into characteristics molecular descriptors.

Язык: Английский

Процитировано

12
Decoding hexanitrobenzene (HNB) and 1,3,5-triamino-2,4,6-trinitrobenzene (TATB) as two distinctive energetic nitrobenzene compounds by machine learning DOI
Rong Wang, Jian Liu, Xudong He

и другие.

Physical Chemistry Chemical Physics, Год журнала: 2022, Номер 24(17), С. 9875 - 9884

Опубликована: Янв. 1, 2022

Energetic materials (EMs) are a group of special energy materials, and it is generally full safety risks costs much to create new EMs. Thus, machine learning (ML)-aided discovery becomes highly desired for EMs, as ML good at risk cost reduction. This work decodes hexanitrobenzene (HNB) 1,3,5-triamino-2,4,6-trinitrobenzene (TATB) two distinctive energetic nitrobenzene compounds by ML, in combination with theoretical calculations. Based on series accurate models density, heat formation, bond dissociation molecular flatness, the predictions show that HNB most among ∼370 000 single benzene ring-containing compounds, while TATB possesses moderate content very high safety, determined experimentally. exhibits significant power presents an instructive procedure using field The ML-aided design efficient synthesis fabrication combined strategy expected accelerate

Язык: Английский

Процитировано

18
Transfer learning for a foundational chemistry model DOI Creative Commons
Emma King‐Smith

Chemical Science, Год журнала: 2023, Номер 15(14), С. 5143 - 5151

Опубликована: Ноя. 24, 2023

Harnessing knowledge from crystal structures yields a model that can predict variety of chemistry-relevant outcomes.

Язык: Английский

Процитировано

11
Force field-inspired transformer network assisted crystal density prediction for energetic materials DOI Creative Commons
Jun-Xuan Jin,

Gao‐Peng Ren,

Jianjian Hu

и другие.

Journal of Cheminformatics, Год журнала: 2023, Номер 15(1)

Опубликована: Июль 19, 2023

Abstract Machine learning has great potential in predicting chemical information with greater precision than traditional methods. Graph neural networks (GNNs) have become increasingly popular recent years, as they can automatically learn the features of molecule from graph, significantly reducing time needed to find and build molecular descriptors. However, application machine energetic materials property prediction is still initial stage due insufficient data. In this work, we first curated a dataset 12,072 compounds containing CHON elements, which are traditionally regarded main composition elements materials, Cambridge Structural Database, then implemented refinement our force field-inspired network (FFiNet), through adoption Transformer encoder, resulting (FFiTrNet). After improvement, model outperforms other learning-based GNNs-based models shows its powerful predictive capabilities especially for high-density materials. Our also capability crystal density (i.e. Huang & Massa dataset), will be helpful practical high-throughput screening

Язык: Английский

Процитировано

10
Screening heat-resistant energetic molecules via deep learning and high-throughput computation DOI
Jian Liu, Jie Tian, Rui Liu

и другие.

Chemical Engineering Journal, Год журнала: 2025, Номер unknown, С. 160218 - 160218

Опубликована: Фев. 1, 2025

Язык: Английский

Процитировано

0
Applications of Predictive Modeling for Energetic Materials DOI
Nasser Sheibani

Challenges and advances in computational chemistry and physics, Год журнала: 2025, Номер unknown, С. 339 - 364

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
Predictive Modeling for Energetic Materials DOI
Didier Mathieu

Challenges and advances in computational chemistry and physics, Год журнала: 2025, Номер unknown, С. 265 - 310

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
Harnessing machine learning enabled quickly predicting density of CHON molecules for discovering new energetic materials DOI Creative Commons

Ruoxu Zong,

Zi Li, Ziyu Hu

и другие.

AIP Advances, Год журнала: 2025, Номер 15(4)

Опубликована: Апрель 1, 2025

The application of machine learning in the research and development energetic materials is becoming increasingly widespread for performance prediction inverse design. Many advances have been achieved, especially discovery various new materials. However, main properties such as data acquisition, molecular characterization, limitations objects insufficient. Density, a critical factor influencing detonation materials, difficult to predict with high precision speed at large scale. In this study, techniques are employed density CHNO result explore simultaneously possessing stability. By screening dataset 16 548 candidate molecules, 175 potential high-performance molecules were identified. Among candidates, it noted that molecule velocity 7.328 Km/s pressure 24.48 GPa was which comparable TNT. study shows transformative accelerating novel vital diverse applications optimized expected accelerate next-generation

Язык: Английский

Процитировано

0