露天矿边坡监测中的小波滤噪与BPANN预测

针对边坡变形量预测难的问题, 将小波分析与BP神经网络预测相结合, 采用小波变换对边坡变形监测数据进行信噪分离, 进而消除观测误差, 通过BP神经网络预测模型BPANN对处理后数据进行再处理, 对边坡变形量以及变形趋势进行预测。进而提出了一种基于小波变换和BPANN模型对露天矿边坡变形监测数据进行处理分析的方法, 并在鞍山某露天矿进行了实际应用。实例结果表明: 利用小波去噪与BPANN模型预测的监测点精度达到3 mm, 满足二等变形监测的要求, 数据处理简便, 在露天矿边坡变形监测数据的消噪与预测中具有实际应用价值。

Wavelet Denoising and BPANN Forecast for Monitoring Slope Deformation in Open Pit

In view of difficulty in slope deformation forecast, a data analysis method was put forward based on the combination of wavelet analysis and BP artificial neural network forecast. Wavelet transform was firstly adopted to do signal-noise separation for slope deformation monitoring data so as to eliminate observation error. Then, BP artificial neural network model (BPANN) was used in post-processing. The slope deformation in open-pit mine and its trend were therefore forecasted. The practical application of this method in an Anshan open-pit mine indicates that wavelet denoising combined with BPANN model can predicts the monitoring point with the precision of 3 mm, up to the requirement for the second-class deformation monitoring. It is of practical value to apply such method into denoising of monitoring data and predication of open-pit slope deformation, due to simplicity of data processing.