Research on coal pillar deformation prediction based on implantable distributed optical fiber

In order to effectively predict the deformation of coal roadway in rock burst mine and improve the monitoring accuracy and frequency of internal displacement deformation of section coal pillar. In this study, a distributed optical fiber monitoring method was used to monitor the internal deformation of overlying strata, and a new combined prediction model algorithm was used to intelligently predict and analyze the deformation of coal roadway under high stress conditions. Through the implantable distributed optical fiber sensing technology, the large-scale, high spatial resolution and high-precision real-time monitoring of coal roadway deformation was realized. In the study, it is found that there is a periodic mutation area in the roof and side wall of the coal roadway, and the average deformation mutation period is 5.6 days, which is consistent with the pressure period of the working face. At the same time, with the advancement of the working face, the deformation of the roof and both sides of the coal roadway continues to increase. Aiming at the denoising problem of optical fiber deformation signal, this paper proposes a hybrid prediction model of improved adaptive noise complete ensemble empirical mode decomposition ( ICEEMDAN ) and particle swarm optimization ( PSO ) extreme learning machine ( ELM ). The experimental results show that compared with the traditional ensemble empirical mode decomposition ( EEMD ) method, the average absolute error ( MAE ) of the ICEEMDAN-PSO-ELM model is reduced by 0.17, the mean square error ( MSE ) is reduced by 0.14, the root mean square error ( RMSE ) is reduced by 0.16, and the coefficient of determination ( R2 ) is increased by 0.10.