Jian Xi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Jian Xi has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Kai Zong
Work Safety Risk Monitoring and Early Warning Center, China Academy of Safety Science and Technology, Beijing 100012, China
2 shared publications - Lei Guan
Work Safety Risk Monitoring and Early Warning Center, China Academy of Safety Science and Technology, Beijing, China
2 shared publications - Hongyu Jing
Work Safety Risk Monitoring and Early Warning Center, China Academy of Safety Science and Technology, China
1 shared publication - Jingcong Zhu
School of Emergency Management and Safety Engineering, China University of Mining and Technology-Beijing, Beijing, China
1 shared publication Lei Guan
1 shared publication- Lei Guan
Work Safety Risk Monitoring and Early Warning Center, China Academy of Safety Science and Technology, Beijing 100012, China
1 shared publication
Latest Publications
- 9 September 2026
- Published by Elsevier in Information Sciences
- Vol. 761
Toward interpretable prediction of coal mine gas leaks using a multi-channel temporal CapsNet driven by self-attention routing
- 9 September 2026
- Published by Elsevier in Information Sciences
- Vol. 761
Toward interpretable prediction of coal mine gas leaks using a multi-channel temporal CapsNet driven by self-attention routing
- Open Access
Pattern Recognition of Hazardous Gas Leak Monitoring Data Based on Field Sensors
Hazardous gas leaks are a major trigger of chemical incidents. If not handled in time, they can easily lead to secondary disasters such as fires and explosions. In recent years, with the construction of hazardous chemical monitoring and early-warning systems in China, large volumes of field operating data from flammable and toxic gas sensors have been accumulated, providing a data foundation for leak-pattern studies grounded in real-world scenarios. In this study, 56 leak samples verified by site feedback were selected. Time-aware interpolation and Z-score normalization were used for preprocessing, and time-series features—including standard deviation of first differences, autocorrelation coefficients, and frequency-domain energy—were extracted. Leak patterns were then identified using two unsupervised approaches: K-Means clustering and a 1D-CNN autoencoder. Results show that K-Means effectively distinguishes macro-patterns such as sustained leaks, instantaneous leaks, fluctuating leaks, and interrupted leaks, while the autoencoder demonstrates stronger capability in extracting temporal features, revealing leak evolution and transition characteristics. The two methods are complementary and together provide a viable route to developing an end-to-end model for leak scenario identification and risk discrimination. This work not only verifies the feasibility of conducting leak-pattern recognition using real GDS data but also offers technical guidance for the intelligent upgrading of hazardous chemical monitoring and early-warning systems.
Open AccessPattern Recognition of Hazardous Gas Leak Monitoring Data Based on Field Sensors
Hazardous gas leaks are a major trigger of chemical incidents. If not handled in time, they can easily lead to secondary disasters such as fires and explosions. In recent years, with the construction of hazardous chemical monitoring and early-warning systems in China, large volumes of field operating data from flammable and toxic gas sensors have been accumulated, providing a data foundation for leak-pattern studies grounded in real-world scenarios. In this study, 56 leak samples verified by site feedback were selected. Time-aware interpolation and Z-score normalization were used for preprocessing, and time-series features—including standard deviation of first differences, autocorrelation coefficients, and frequency-domain energy—were extracted. Leak patterns were then identified using two unsupervised approaches: K-Means clustering and a 1D-CNN autoencoder. Results show that K-Means effectively distinguishes macro-patterns such as sustained leaks, instantaneous leaks, fluctuating leaks, and interrupted leaks, while the autoencoder demonstrates stronger capability in extracting temporal features, revealing leak evolution and transition characteristics. The two methods are complementary and together provide a viable route to developing an end-to-end model for leak scenario identification and risk discrimination. This work not only verifies the feasibility of conducting leak-pattern recognition using real GDS data but also offers technical guidance for the intelligent upgrading of hazardous chemical monitoring and early-warning systems.
- Open Access
- 29 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 424-428
Research on Work Safety Intelligent Inspection System Based on Identifier Resolution and Dynamic Encryption Algorithm
- 29 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 424-428
Open AccessResearch on Work Safety Intelligent Inspection System Based on Identifier Resolution and Dynamic Encryption Algorithm
- Open Access
- 16 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 323-330
Research on the Dynamic Safety Supervision Model of Chemical Enterprise Based on Machine Learning
This paper first analyzes the accident data in China's chemical industry from 2018 to 2022, revealing that although the number of accidents and the mortality rate have decreased, the accident rate remains high, highlighting the urgency of strengthening safety supervision methods. To address this, a dynamic supervision indicator system covering multiple key dimensions such as production load, alarm conditions, operation quantity, and maintenance status has been constructed. This indicator system can capture key factors related to safety risks through multidimensional monitoring of enterprises' daily work safety status. The study uses the entropy weight method to objectively weight each indicator and employs K-means and DBSCAN clustering algorithms to classify enterprise data. While the entropy weight method provides an intuitional basis for feature weighting, further optimization is needed to improve sample discrimination. The K-means algorithm effectively distinguishes different risk states of enterprises, such as shutdown and full-load production, and reasonably classifies enterprises into three categories: special control, key attention, and general supervision. In contrast, the DBSCAN algorithm performs better in identifying high-risk enterprises but is limited in finely distinguishing between special control and key attention enterprises. The final results show that the K-means algorithm performs best when combined with the indicator system for dynamic supervision, particularly in accurately identifying low-risk and generally supervised enterprises. By establishing a comprehensive dynamic supervision indicator system and applying machine learning algorithms, this paper provides a new technological path for safety supervision of hazardous chemical enterprises, demonstrating the potential of intelligent supervision in enhancing accident prevention and risk control levels.
- 16 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 323-330
Open AccessResearch on the Dynamic Safety Supervision Model of Chemical Enterprise Based on Machine Learning
This paper first analyzes the accident data in China's chemical industry from 2018 to 2022, revealing that although the number of accidents and the mortality rate have decreased, the accident rate remains high, highlighting the urgency of strengthening safety supervision methods. To address this, a dynamic supervision indicator system covering multiple key dimensions such as production load, alarm conditions, operation quantity, and maintenance status has been constructed. This indicator system can capture key factors related to safety risks through multidimensional monitoring of enterprises' daily work safety status. The study uses the entropy weight method to objectively weight each indicator and employs K-means and DBSCAN clustering algorithms to classify enterprise data. While the entropy weight method provides an intuitional basis for feature weighting, further optimization is needed to improve sample discrimination. The K-means algorithm effectively distinguishes different risk states of enterprises, such as shutdown and full-load production, and reasonably classifies enterprises into three categories: special control, key attention, and general supervision. In contrast, the DBSCAN algorithm performs better in identifying high-risk enterprises but is limited in finely distinguishing between special control and key attention enterprises. The final results show that the K-means algorithm performs best when combined with the indicator system for dynamic supervision, particularly in accurately identifying low-risk and generally supervised enterprises. By establishing a comprehensive dynamic supervision indicator system and applying machine learning algorithms, this paper provides a new technological path for safety supervision of hazardous chemical enterprises, demonstrating the potential of intelligent supervision in enhancing accident prevention and risk control levels.