Utilizing LSTM networks for LaBr 3 :Ce scintillators RUL prediction under extreme gamma irradiations

Cerium-doped lanthanum bromide (LaBr ₃ :Ce) scintillation detectors are widely deployed in harsh radiation environments, such as space exploration, radioactive monitoring in nuclear facilities, and high energy physics. Cumulative radiation damage introduces lattice defects and radiation-induced absorption (RIA), which results in progressive degradation of optical transmittance and light yield. Previous investigations have been largely confined to post-irradiation characterizations of light output and transmittance. Therefore, predictive methods for estimating the remaining useful life (RUL) of LaBr ₃ :Ce scintillators remain absent, especially under extreme gamma irradiation. This paper presents a prognostic method based on multiphysics simulation, utilizing a Long Short-Term Memory (LSTM) network to estimate the RUL of LaBr ₃ :Ce crystals. A kinetic multiphysics platform is constructed to simulate the spatiotemporal evolution of radiation damage across multiple working conditions with varying source activities and crystal geometries, generating multi-condition EWLT sequences. A geometry-dependent optical threshold is then established to define the condition-specific failure state. A single LSTM network is trained jointly on the pooled sequences and compared with GRU, vanilla RNN, and exponential fitting methods under strict chronological splits. Results demonstrate that the LSTM achieves the lowest forecasting errors across the evaluated conditions, with group-mean RMSE values of 1.47×10 ⁻⁴ and 2.71×10 ⁻⁴ for the activity and geometry conditions, respectively. This method provides a data-driven approach for the predictive maintenance of radiation detectors.

Authors

Institutions

Publication Details

Journal
Physica Scripta
Published
2026-11-18
DOI
https://doi.org/10.1088/1402-4896/aec3e5
Primary Topic
Radiation Detection and Scintillator Technologies
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Utilizing LSTM networks for LaBr 3 :Ce scintillators RUL prediction under extreme gamma irradiations

Xiang Li, Yanliang Tan, Yuxi Xie, Junwei Shi et al.
1 links
Physica Scripta
Radiation Detection and Scintillator Technologies
article

Utilizing LSTM networks for LaBr 3 :Ce scintillators RUL prediction under extreme gamma irradiations

Xiang Li, Yanliang Tan, Yuxi Xie, Junwei Shi, Rui Yan, Yuebing Xu, Wei Xu, Dongfan Gong, Heng Zheng, Wuning Ye
article en

Abstract

Cerium-doped lanthanum bromide (LaBr ₃ :Ce) scintillation detectors are widely deployed in harsh radiation environments, such as space exploration, radioactive monitoring in nuclear facilities, and high energy physics. Cumulative radiation damage introduces lattice defects and radiation-induced absorption (RIA), which results in progressive degradation of optical transmittance and light yield. Previous investigations have been largely confined to post-irradiation characterizations of light output and transmittance. Therefore, predictive methods for estimating the remaining useful life (RUL) of LaBr ₃ :Ce scintillators remain absent, especially under extreme gamma irradiation. This paper presents a prognostic method based on multiphysics simulation, utilizing a Long Short-Term Memory (LSTM) network to estimate the RUL of LaBr ₃ :Ce crystals. A kinetic multiphysics platform is constructed to simulate the spatiotemporal evolution of radiation damage across multiple working conditions with varying source activities and crystal geometries, generating multi-condition EWLT sequences. A geometry-dependent optical threshold is then established to define the condition-specific failure state. A single LSTM network is trained jointly on the pooled sequences and compared with GRU, vanilla RNN, and exponential fitting methods under strict chronological splits. Results demonstrate that the LSTM achieves the lowest forecasting errors across the evaluated conditions, with group-mean RMSE values of 1.47×10 ⁻⁴ and 2.71×10 ⁻⁴ for the activity and geometry conditions, respectively. This method provides a data-driven approach for the predictive maintenance of radiation detectors.

Physica Scripta
Hengyang Normal University (CN)
Natural Science Foundation of Hunan Province, State Key Laboratory of Particle Detection and Electronics, Scientific Research Foundation of Hunan Provincial Education Department
Radiation Detection and Scintillator Technologies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon