基于最大Lyapunov指数的冲击地压预测模型

Predicting Model of Rock Burst Based on Lyapunov Index

冲击地压的拟合预测方法是对监测数据在未来一定时期的变化和走势规律进行预测,传统的数理统计预测模型把冲击地压监测序列认为是由于外在随机因素引起的,而冲击地压观测时序大多是一个貌似随机的非线性混沌序列,随机过程理论并不完全适合冲击地压时序的预测.最大Lyapunov指数作为量化动力系统对初始轨道的指数发散和估计系统的混沌量,是系统的一个很好的预报参数,本文在对观测序列相空间重构的基础上,基于最大Lyapunov指数对冲击地压工作面观测时序建立了预测模型,并与传统的数理统计预测方法进行了对比分析.通过对冲击危险区域的电磁辐射日平均值序列及顶板下沉速度序列等实例的预测运算和分析,得到冲击地压的最大Lyapunov指数预测模型达到了较高的预测精度,是完全可行的.

The fitting predicting of rock burst is to predict changes of observation data during a certain future period. In traditional model of mathematic statistics, rock burst observation series are regarded as being caused by external random factors. Actually, it mostly is a seemingly random nonlinear chaotic series, and stochastic process theory is not suitable for prediction of rock burst time series. As a value ensuring index discrete extent to initial orbit of dynamical system and estimating chaotic extent of dynamical system, Lyapunov index is a very perfect predicting parameter in the system. Based on the phase space reconstruction of observation time series in workface of rock burst, a prediction model was proposed according to the maximal Lyapunov index, and the prediction results were compared with those from mathematic statistics model. By prediction analysis to mean series of daily observation data of electromagnetic emission and roof subsidence velocity series, the results show that the prediction model based on the maximal Lyapunov index has very high prediction precision and is feasible.