深基坑降水引起周边地面沉降量值计算修正系数M_s的确定

Determination of coefficient M_s during calculation of surrounding ground settlement due to foundation pit dewatering

从降落漏斗水位以上覆盖层、弱透水层的排水固结过程到多孔介质(岩土层)的压缩、水的压缩三方面分析了深井降水引起地面沉降的机理,得出降水引起地面沉降的压缩层厚度主要为降落漏斗水位以上降深范围内的覆盖层和弱透水层厚度的结论。同时收集了10个有代表性的深基坑降水工程的相关资料,通过实测值与理论值的比值,得出一批降水引起地面沉降修正系数Ms值(880组数据)。采用BP神经网络对这批数据进行处理,使它能在给定的输入(t,S,E s,h)和输出(M s)之间,建立一种网络映射关系。利用该网络任意输入一组(Es,h)值,即可得出不同t,S对应的修正系数Ms等值线图,为类似地层的基坑降水工程,在预测不同的时间、距离的实际沉降量上具有实用价值。

From overburden layers, semipervious layers above depression-cone level drainage consolidation, porous media （geo-formation） compression and water compression, the ground settlement mechanism is analysed during foundation pit dewatering. Compression layers thickness due to dewatering is chiefly overburden layers, semipervious layers thickness above depression-cone level. Collecting the related data of 10 foundation pits dewatering engineerings, the practical values and theoretical values are compared and a set of modified-coefficient Ms of surrounding ground settlement due to foundation pit dewatering （aggregately 880 groups of data ） are got. Using BP neural network method to process those data, we can find a sort of net mapped connection between given inputs （t, S, Es^-, h） and outputs （Ms）.We can use net connection to input a sort of （Es^-, h）, and get isoline of modified-coefficient Ms which is corresponding with different t, S. Accordingly we can find modified-coefficient Ms of arbitrary t, S, which has practical value to foundation pit dewatering engineering of similar layers to forecast practical settlement at different times and distances from foundation pit.