基于混沌时序预测方法的冲击地压预测研究

STUDY ON ROCK BURST FORECAST WITH FORECAST METHOD BASED ON CHAOTIC TIME SERIES

冲击地压的预测研究通常是通过监测对冲击地压的发生、发展过程比较敏感的指标来进行的,监测指标值的大小和变化规律是进行预测的基础,监测数据在未来一定时期的峰值变化和走势变化规律对于预测过程具有重要意义。首先,通过对互信息和伪邻近点数的计算,确定观测序列的延迟时间和嵌入维数等相空间重构参数;然后,在对观测序列相空间重构的基础上,运用一阶局域近似法和基于最大Lyapunov指数法等混沌预测方法对冲击地压工作面的观测时间序列进行数学建模,并与传统的数理统计预测方法进行对比分析;最后,用实例对冲击危险区域的电磁辐射序列及顶板下沉速度序列等进行预测运算和分析,其结果表明,运用混沌理论的预测方法可达到较高的预测精度。

The forecast study on rock burst is carded out by monitoring some indices which are sensitive to the developing process of rock burst. The values and changing rules of the monitoring indices are the foundations to forecast rock burst; and the changing rules of peak values and trend in future of monitoring data have important meaning to the monitoring process. The forecast model considers that the monitoring series of rock burst is evoked by random factors; and the theory of random processes is used to simulate the movement rules of the system. The developing process of rock burst has its own rules and characteristics. The monitoring series is the exterior behavior of the systematic variables in the course of the development of rock burst; and the theory of random processes isn＇t adapt to forecasting the time series of rock burst. Based on the reconstruction of phase space of monitoring series, the monitoring data of electromagnetic emission on working face of rock burst are forecasted by the methods of one-order approximation and Lyapunov exponent. The results are compared with ones gained by mathematical statistics method and show that the prediction methods based on chaotic theory have higher forecast precision.