A rock failure critical point discrimination method based on acoustic emission timing prediction residual
By using a method based on acoustic emission time-series prediction residuals, and employing a long short-term memory network model and dynamic threshold determination technology, the problems of weak anti-interference ability and data dependence in existing rock failure early warning methods are solved, and accurate identification and early warning of rock failure critical points are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 新疆葱岭能源有限公司
- Filing Date
- 2026-04-24
- Publication Date
- 2026-07-17
AI Technical Summary
Existing rock damage early warning methods based on acoustic emission have weak anti-interference capabilities in mining environments, struggle to distinguish between instantaneous noise spikes and actual pre-fracture signals, lack in-depth analysis of the temporal evolution of acoustic emission, and rely on scarce damage sample data.
A method based on acoustic emission time-series prediction residuals is adopted. The cumulative ringing count sequence is predicted through a long short-term memory network model. The mean and standard deviation of the residual sequence are calculated, a dynamic reference noise benchmark and adaptive adjustment coefficient are constructed, and a dynamic outlier determination threshold is generated to achieve accurate identification of the critical point of rock failure.
It enables accurate and timely identification of rock failure critical points, improves the anti-interference capability and generalization performance of the early warning system, and can adapt to signal fluctuations in different rock samples and monitoring environments.
Smart Images
Figure CN122409873A_ABST