矿区地下水位的混沌时间序列局域法预测

矿区地下水系统是一个非线性演化系统,地下水位的变化也是一个不可积的过程,随着采矿的进行,其演化过程会通过一定的轨道进入混沌状态。对杨庄煤矿1号奥灰水位的水文地质长观孔观测时间序列数据进行了分析,根据Pacard和Takens提出的相空间重构技术,首先采用自相关函数法计算其时间延迟,并计算其最佳嵌入维数,然后利用Wolf提出的方法,从单变量中提取出了最大Lyapunov指数,由计算结果得出时间序列具有混沌特性。在此基础上,采用混沌时间序列的局域法对水位做了6步预测,结果表明,混沌时间序列方法对混沌序列的预测具有较高的精度。

Local Method Prediction of Chaotic Time Series for Groundwater Level in Mining Area

The underground water system in mining area is a nonlinear evolutionary system,and the variation of the groundwater level universally is also a non-integrable process.With the mineral extraction,the evolutionary process of the groundwater level will entry chaotic condition via particular path.This article analyses the time series data from the water level observation on the first long-term hydrogeologic observation drill of "Ordovician limestone" in Yangzhuang coal mine.Based on the phase space reconstruction technology provided by the Pacard and Takens,first the delay time is determined by the self-correlation function and the optimal inserted dimension is calculated,the maximal Lyapunov index is took out from the single variation by using method provided by Wolf.According to the results,it is found that the time series of the groundwater level is of chaotic feature.On the base of the above,the six-step prediction of the groundwater is done by the local method of chaotic time series.The consequence is that chaotic time series prediction is more accurate in the prediction of chaotic series.