矿井涌水量重标极差法等维预测模型

Equal dimension prediction model of mine water inflow based on rescaled range analysis method

目前,矿井涌水量预测方法较多,但预测效果并不十分理想。为了提高预测的精度和速度,鉴于重标极差法对复杂系统演变分析和预测具有的独特优势,建立基于MATLAB的重标极差法涌水量等维预测模型,采用试算法确定涌水量时间序列的后续最佳取值。利用平顶山天安煤业股份有限公司八矿的涌水量资料对模型进行验证,结果表明:该矿涌水量时间序列的Hurst指数H=0.968,涌水量序列均具有很强的持续性;平均循环周期为17个月;建立的模型预测精度达到了98.43%,说明此模型对该矿涌水量的预测十分有效。利用MATLAB软件编写的重标极差法预测涌水量的程序,避开了繁琐的计算,提高了预测速度和精度。

At present,there are many methods for predicting mine water inflow,but the prediction effect is not satisfactory.In order to improve the precision and speed of prediction,in view of the advantages of rescaled range analysis method for the analysis and prediction of complex system evolution,an equal dimension prediction model of mine water inflow based on rescaled range analysis method with MATLAB was established.The trial algorithm was used to determine the subsequent optimal value of the water inflow time series.The model was verified by using the water inflow data of No.8 Mine of Pingdingshan Tianan Coal Mining Co.,Ltd.The resnlts showed that:Hurst index of the mine water inflow time series was equal to 0.968,that was,the water inflow sequence had strong persistence;The average cycle time of the mine water inflow time series was 17 months;The prediction accuracy of the model reached 98.43%,indicating that the model was very effective for predicting the water inflow of the mine.The program of predicting water inflow written by MATLAB software based on rescaled range analysis method avoided the cumbersome calculation and improved the prediction speed and precision.