Research on Grey Prediction of Deformation Laws in Backfill Based on Phase Space Reconstruction

李夕兵,刘志祥
DOI: https://doi.org/10.3969/j.issn.1009-6094.2004.06.015
2004-01-01
Abstract:The present paper intends to introduce the authors' research on the stress-strain laws of cemented tailings in different cement ratios to tailings through experimentation with monitoring the deformations of high backfill in Anqing Copper Mine production practice. The research is done to find a method, which aims to remove the noise from monitored data in reconstructive phase space by following the principle of adaptive filter. Cemented tailings backfill is by nature a nonlinear mechanical medium, and its deformation is a complex process of energy dissipation. Therefore, it is necessary to study the intrinsic laws of deformation, and predict the stability of backfill exactly during the mining practice. Analyzing the change laws of distance in phase space to deformation in backfill, a prediction model is built based on phase space reconstruction with gray theory. In order to reduce the prediction errors, the prediction model has to be improved with an amended model of error. Using the model proposed in this paper, we have analyzed the deformations in high backfill in the mine, to solve the reliable mining period. The research results also show that the deformations in backfill appear the random feature. However, in spite of the existence of different nonlinear dynamical characters in the backfill to different ratios of cement to tailings, the laws of deformation can still be well demonstrated by the method of reconstructive phase space. Comparing the predicted results with the monitoring data, they prove to be well coincided. The lower is the ratios of cement to tailings in backfill, the higher is the dimension of dynamic system in phase space to the time series of deformation. And in turn, the lower is strength of backfill, the more complex is the process of Phase Space Reconstruction.
What problem does this paper attempt to address?