基于Richards-BP模型的地表沉降特征预测模型研究

A Combined Model of Richards Model and BP Neural Network Model to Predict the Surface Subsidence Characteristics

地下开采引起的地表沉降大致呈S形发展,最终趋于稳定状态,这能够运用生长曲线Richards函数进行预测分析。同时,又因为测量和外界存在不确定性和随机性,使得地表沉降也具有动态的特征。针对地下岩矿开采矿区地表沉降曲线与Richards预测模型曲线的相似性,分析了Richards模型在地表沉降预测中的适用性,提出通过Richards生长曲线模型预测矿区地表沉降的趋势性变形部分,利用BP神经网络模型降低地表沉降的随机性影响部分,提高模型预测效果。计算结果证明了其在地表沉降预测中的适用性和可行性。

Ground subsidence caused by mining is almost growth in an s-shaped curve,which tends to be stable at last,and it can be predicted and analyzed by applying Richards function curve.Because there will be uncertainty and randomness along with the measurement and external environment,which makes the characteristics of surface subsidence also be dynamic.According to the similarity of two curves which caused by underground rock ore mining and Richards prediction model curve,the applicability of the Richards model is analyzed to predict the surface subsidence in this paper.It is suggested that Richards model can be used to predict the trend parts deformation of the surface in the mining area,and BP neural network model can be used to re-correct random effect parts of it.The empirical results demonstrate that the proposed method has high prediction accuracy,and its feasibility and applicability in surface subsidence prediction is proved.