Open Access

ARTICLE

Coal Rock Condition Detection Model Using Acoustic Emission and Light Gradient Boosting Machine

Jing Li¹, Yong Yang^{2, *}, Hongmei Ge¹, Li Zhao³, Ruxue Guo^{3, 4}

1 School of Information Engineering, Nanjing Audit University, Nanjing, 211815, China.
2 School of Information and Electrical Engineering, Xuzhou College of Industrial Technology, Xuzhou, 221140, China.
3 School of Information Engineering, Southeast University, Si Pailou Campus, Nanjing, 210096, China.
4 Department of Psychiatry, Vagelos College of Physicians and Surgeons, New York, 10027, USA.

* Corresponding Author: Yong Yang. Email: .

Computers, Materials & Continua 2020, 63(1), 151-162. https://doi.org/10.32604/cmc.2020.05649

Received 07 January 2019; Accepted 04 June 2019; Issue published 30 March 2020

Abstract

Coal rock mass instability fracture may result in serious hazards to underground coal mining. Acoustic emissions (AE) stimulated by internal structure fracture should carry lots of favorable information about health condition of rock mass. AE as a sensitive non-destructive test method is gradually utilized to detect anomaly conditions of coal rock. This paper proposes an improved multi-resolution feature to extract AE waveform at different frequency resolutions using Coilflet Wavelet Transform method (CWT). It is further adopt an efficient Light Gradient Boosting Machine (LightGBM) by several cascaded sub weak classifier models to merge AE features at different views of frequency for coal rock anomaly damage recognition. The results denote that the proposed method achieves excellent recognition performance on anomaly damage levels of coal rock. It is an effective method to detect the critical stability further to predict the rock mass bursting in time.

---

Keywords

---

Citations