充填体变形的混沌时序重构与神经网络预测

Reconstruction of Chaotic Time Series for Backfill Deformation and Prediction with Neutral Network

通过对充填体变形时间序列重构相空间 ,研究了充填体变形在相空间中相点距离的演变规律 ,建立了充填体变形的神经网络预测模型。研究结果表明 ,充填体变形具有非线性混沌特性 ,不同配比的充填体表现出不同的非线性动力学行为 ,重构相空间能充分展示充填体变形的内在规律。应用所建立的模型 ,对安庆铜矿高阶段充填体变形进行了预测与分析 。

Phase space reconstruction method was used for time series of backfill deformation. After the changing laws of distance between two phase points in the phase space have been studied for backfill deformation, a prediction model of neural network has been established for deformation of backfill. Research results show that deformation of backfill is characterized by nonlinear chaos. Different nonlinear dynamical behaviors exist in backfill with different ratios of cement to tailing, and the intrinsic laws of backfill deformation can be well demonstrated by the phase space reconstruction method. So deformations of high backfill are predicted with the model established for Anqing Copper Mine, and a reasonable stopping cycle for high-level mining was also discussed.