Local Acceleration Monitoring and its application in physical modelling of underground mining

A Micro-Electro-Mechanical Systems (MEMS) inertial sensor was introduced to perform Local Acceleration Monitoring (LAM) in a physical model to showcase mining-induced strata movements. To prove the feasibility of LAM, a physical model experiment was performed using four sensors mounted on the roof strata. Acquired acceleration signals were processed by a moving average and median filter, analysed with respective trend variation and fluctuation along with excavation sequence. To identify the movement of roof strata, a key reference was the mid-span section that characterized by adjacent sensors with opposite acceleration change direction. To assess the stability of strata, a clear evidence is the synchronicity loss of two adjacent sensors in terms of acceleration change. In the presented experiment, the first weighting and the fracture position can be predicted 1.5 h prior to final failure. This study represents a novel attempt to apply LAM in a physical modelling experiment and tentatively interpret signals in accordance with strata movement. The proposed LAM-based approach succeeded in identifying the movement and assessing the stability of the strata. Results presented in this paper verify the validity of LAM in terms of noise analysis, sensibility and rationality. This LAM technique is capable of serving as an easy-to-implement, high-precision and low-cost method in mining and geotechnical related physical modelling research.