Maximum detectable distance in mobile gamma spectrometry using physics-informed neural network: optimizing radiation detection

被引：0

作者：

Nancy A. Ibrahim ^{[1
]}

Amin Amirlatifi ^{[2
]}

Peixue Ma ^{[1
]}

Somayeh Bakhtiari Ramezani ^{[2
]}

机构：

[1] Mississippi State University,Dave C. Swalm School of Chemical Engineering

[2] Radioanalytical Services,New Jersey Department of Health

来源：

Journal of Radioanalytical and Nuclear Chemistry | 2025年 / 334卷 / 5期

关键词：

Physics-informed neural network; Machine learning; Mobile gamma spectrometry; Radiation detection;

D O I：

10.1007/s10967-025-10148-5

中图分类号：

学科分类号：

摘要：

Mobile gamma spectrometry is crucial for detecting gamma-ray sources in environmental monitoring and homeland security. The Maximum Detectable Distance (MDD) defines the farthest reliable detection range; however, conventional models fail to account for spatial variations in radiation intensity during movement. This study introduces a Physics-Informed Neural Network (PINN) to refine MDD predictions by incorporating detector speed, angular efficiency, and detection probability. For an unshielded 137Cs source with 95% detection probability, the optimal MDD of 5.4 m aligns with theoretical expectations. By dynamically adjusting for speed and angular effects, this framework enhances real-time predictions, improving field survey efficiency and remediation strategies.

引用

页码：3297 / 3309

页数：12