Blast loading prediction in a typical urban environment based on Bayesian deep learning
Explosion events in urban environment, such as terrorist attacks, accidental industrial explosions and missile attacks in war, can be destructive to residents and properties, causing great casualties and structural damages. Rapid and accurate methods for blast loading prediction in urban environment...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2025-12-01
|
| Series: | Engineering Applications of Computational Fluid Mechanics |
| Subjects: | |
| Online Access: | https://www.tandfonline.com/doi/10.1080/19942060.2024.2445765 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841554308636606464 |
|---|---|
| author | Weiwen Peng Meilin Pan Chunjiang Leng Shufei Wang Wei Zhong |
| author_facet | Weiwen Peng Meilin Pan Chunjiang Leng Shufei Wang Wei Zhong |
| author_sort | Weiwen Peng |
| collection | DOAJ |
| description | Explosion events in urban environment, such as terrorist attacks, accidental industrial explosions and missile attacks in war, can be destructive to residents and properties, causing great casualties and structural damages. Rapid and accurate methods for blast loading prediction in urban environment are crucial for risk mitigation and emergency response planning. Commonly used numerical simulation methods can provide accurate blast loading prediction results, but suffer from high computation time and cost. In this paper, a novel and fast method for blast loading prediction in a typical urban environment is proposed. Deep learning-based prediction model is constructed by leveraging numerical simulations of urban explosions, for which a wide range of simulation data with diversified combinations of explosion and environment variables are obtained for assuring generalizability of the prediction model. This blast loading prediction model is further enhanced with Bayesian deep learning, which equips the prediction model with the capability of uncertainty quantification of blast loading prediction without sacrificing its rapidity or accuracy. Numerical study of blast loading prediction in a typical urban environment is presented to demonstrate the proposed method. The results show that the Bayesian deep learning-based prediction model can fulfil the blast loading prediction in this typical urban environment within 2.9 millisecond, and the mean absolute percentage error of blast loading prediction is controlled below 12.9%. Compared with existing data-driven methods, the proposed method can generalize better in this typical urban environment studied in this paper, and improved prediction accuracy and uncertainty quantification capability are acquired, while maintaining comparable prediction rapidity. This research provides a new method and prospect for blast loading prediction in a typical urban environment, enabling accurate and effective safety assurance, risk mitigation for residents and properties under explosion events in urban environment. |
| format | Article |
| id | doaj-art-250cbd8332b44ed0898c92cb3c2afdf5 |
| institution | Kabale University |
| issn | 1994-2060 1997-003X |
| language | English |
| publishDate | 2025-12-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | Engineering Applications of Computational Fluid Mechanics |
| spelling | doaj-art-250cbd8332b44ed0898c92cb3c2afdf52025-01-08T14:12:49ZengTaylor & Francis GroupEngineering Applications of Computational Fluid Mechanics1994-20601997-003X2025-12-0119110.1080/19942060.2024.2445765Blast loading prediction in a typical urban environment based on Bayesian deep learningWeiwen Peng0Meilin Pan1Chunjiang Leng2Shufei Wang3Wei Zhong4School of Intelligent Systems Engineering, Sun Yat-sen University, Guangzhou, People’s Republic of ChinaSchool of Intelligent Systems Engineering, Sun Yat-sen University, Guangzhou, People’s Republic of ChinaSchool of Intelligent Systems Engineering, Sun Yat-sen University, Guangzhou, People’s Republic of ChinaState Key Laboratory of Intense Pulsed Radiation Simulation and Effect, Northwest Institute of Nuclear Technology, Xi'an, People’s Republic of ChinaState Key Laboratory of Intense Pulsed Radiation Simulation and Effect, Northwest Institute of Nuclear Technology, Xi'an, People’s Republic of ChinaExplosion events in urban environment, such as terrorist attacks, accidental industrial explosions and missile attacks in war, can be destructive to residents and properties, causing great casualties and structural damages. Rapid and accurate methods for blast loading prediction in urban environment are crucial for risk mitigation and emergency response planning. Commonly used numerical simulation methods can provide accurate blast loading prediction results, but suffer from high computation time and cost. In this paper, a novel and fast method for blast loading prediction in a typical urban environment is proposed. Deep learning-based prediction model is constructed by leveraging numerical simulations of urban explosions, for which a wide range of simulation data with diversified combinations of explosion and environment variables are obtained for assuring generalizability of the prediction model. This blast loading prediction model is further enhanced with Bayesian deep learning, which equips the prediction model with the capability of uncertainty quantification of blast loading prediction without sacrificing its rapidity or accuracy. Numerical study of blast loading prediction in a typical urban environment is presented to demonstrate the proposed method. The results show that the Bayesian deep learning-based prediction model can fulfil the blast loading prediction in this typical urban environment within 2.9 millisecond, and the mean absolute percentage error of blast loading prediction is controlled below 12.9%. Compared with existing data-driven methods, the proposed method can generalize better in this typical urban environment studied in this paper, and improved prediction accuracy and uncertainty quantification capability are acquired, while maintaining comparable prediction rapidity. This research provides a new method and prospect for blast loading prediction in a typical urban environment, enabling accurate and effective safety assurance, risk mitigation for residents and properties under explosion events in urban environment.https://www.tandfonline.com/doi/10.1080/19942060.2024.2445765Urban blastsblast loadingBayesian deep learninguncertainty quantificationpeak overpressure |
| spellingShingle | Weiwen Peng Meilin Pan Chunjiang Leng Shufei Wang Wei Zhong Blast loading prediction in a typical urban environment based on Bayesian deep learning Engineering Applications of Computational Fluid Mechanics Urban blasts blast loading Bayesian deep learning uncertainty quantification peak overpressure |
| title | Blast loading prediction in a typical urban environment based on Bayesian deep learning |
| title_full | Blast loading prediction in a typical urban environment based on Bayesian deep learning |
| title_fullStr | Blast loading prediction in a typical urban environment based on Bayesian deep learning |
| title_full_unstemmed | Blast loading prediction in a typical urban environment based on Bayesian deep learning |
| title_short | Blast loading prediction in a typical urban environment based on Bayesian deep learning |
| title_sort | blast loading prediction in a typical urban environment based on bayesian deep learning |
| topic | Urban blasts blast loading Bayesian deep learning uncertainty quantification peak overpressure |
| url | https://www.tandfonline.com/doi/10.1080/19942060.2024.2445765 |
| work_keys_str_mv | AT weiwenpeng blastloadingpredictioninatypicalurbanenvironmentbasedonbayesiandeeplearning AT meilinpan blastloadingpredictioninatypicalurbanenvironmentbasedonbayesiandeeplearning AT chunjiangleng blastloadingpredictioninatypicalurbanenvironmentbasedonbayesiandeeplearning AT shufeiwang blastloadingpredictioninatypicalurbanenvironmentbasedonbayesiandeeplearning AT weizhong blastloadingpredictioninatypicalurbanenvironmentbasedonbayesiandeeplearning |