高边坡安全监测的改进GM模型预测研究

IMPROVED GM MODEL FOR SAFETY MONITORING OF HIGH ROCK SLOPES

由于一般的 GM(1,1)模型进行预测存在较大的局限性及产生系统误差,对一般 GM(1,1)模型进行了误差来源追踪分析并提出改进方法,得到改进后的 GM(1,1)预测模型,并将其用于高边坡安全监测。依据碧口水电站高边坡连续 8 a 的监测数据(第 6,7 a 的数据予以剔除),建立了碧口水电站高边坡灰色安全监控模型。把改进的 GM(1,1)预测模型与一般的 GM(1,1)模型、统计模型等预测模型进行了对比,同时还进行了平均误差、相关系数以及最大误差分析对比。研究表明,改进后 GM(1,1)模型监控精度较高,预测结果与实际吻合较好。

The conventional GM(1, 1) model is highly limited and prone to accumulate systematic error in prediction. The paper traces its error sources and proposes an improved method. This improved model is applied for analyzing the monitoring data from eight successive years of a high rock slope monitoring of the Bikou hydroelectric power station, where the sixth and the seventh years' data is omitted. A grey safety monitoring model of the high rock slope is established by using the improved GM(1, 1) prediction model. By comparing the prediction results of the improved GM(1, 1) model with those of the conventional GM(1, 1) model and the statistical model, and with those of the mean error, correlation coefficient, and maximum error, it is proved that the modified GM(1, 1) model is of high accuracy in safety monitoring and its prediction agrees well with the actual behavior of the high rock slope.