Related Experiment Video
Updated: May 31, 2026
Data Processing Methods for 3D Seismic Imaging of Subsurface Volcanoes: Applications to the Tarim Flood Basalt
Published on: August 7, 2017
A frame-difference-aware rockburst early warning method based on the spatiotemporal distribution of microseismicity
Jie Zhang1, Ke Yang1, Xin Lyu1
1School of Mining Engineering, Anhui University of Science and Technology, Huainan, 232001, Anhui, China; State Key Laboratory for Safe Mining of Deep Coal Resources and Environment Protection, Anhui University of Science and Technology, Huainan, 232001, Anhui, China.
A new Frame-Difference-aware Spatiotemporal Fusion Network (FDSF-Net) improves rockburst prediction in coal mines. This AI model analyzes microseismic data to provide intelligent early warnings, enhancing mine safety.
Area of Science:
- Geophysics
- Mining Engineering
- Artificial Intelligence
Background:
- Microseismic vibration fields in deep coal mines indicate stress changes, crucial for predicting dynamic rockbursts.
- Existing methods often use limited microseismic data, failing to capture complex spatiotemporal dynamics.
Purpose of the Study:
- To develop a novel deep learning framework for accurate microseismic vibration field prediction.
- To enhance intelligent early warning systems for rockbursts in underground coal mines.
Main Methods:
- Proposed the Frame-Difference-aware Spatiotemporal Fusion Network (FDSF-Net), a fully spatiotemporal deep learning model.
- Incorporated channel attention for dynamic disturbance information and multiscale spatial convolutions for pattern recognition.
- Utilized hierarchical fusion and temporal-channel mixing for comprehensive spatiotemporal feature extraction.
Main Results:
- FDSF-Net achieved Mean Absolute Error (MAE) of 109.75 and Root Mean Square Error (RMSE) of 3.53 on real-world microseismic data.
- Model demonstrated effective learning of multiscale spatial patterns and robust generalization across different mine sites.
- Feature visualization confirmed the model's ability to capture critical spatiotemporal dynamics.
Conclusions:
- FDSF-Net offers a significant advancement in microseismic vibration field prediction for rockburst early warning.
- The developed framework supports the creation of more accurate and interpretable intelligent warning systems.
- This approach enhances safety in deep underground coal mining operations.
