考虑侧向变形的软土地基非线性沉降计算的简化法

软土地基侧向变形占总沉降的比例大,通常的计算方法采用有侧限的一维压缩状态下的压缩模量应用于分层总和法来计算,这样的方法不能反应侧向变形对总沉降的影响,而规范方法则在该方法计算结果的基础上乘以1.1~1.7的经验系数,以修正计算误差。但经验系数的选取具有较大的人为性。为此,在广义胡克定律的基础上推导出可考虑侧向变形的计算方法,把软土地基沉降分为有侧限的压缩沉降S_c和侧向变形产生的沉降S_d两部分,前者采用传统的e–p曲线分层总和法计算,后者采用基于e–p曲线和邓肯–张(Duncan–Chang)本构模型的概念求取非线性割线模量E_p应用于分层总和法计算,这样就可以由e–p曲线进行考虑软土侧向变形的非线性沉降计算。另外,鉴于工程中初始孔隙比e_0和压缩模量Es1-2(压力为100 kPa和200 kPa对应的压缩模量)是常用的参数,相对稳定且可较好反映软土的特性,建立了由e_0和E_(s1-2)求e–p曲线和e–lgp曲线的方法,从而可以求出不同应力水平下的压缩模量Esi,由Esi进行S_c和S_d的计算,实现可由初始孔隙比e_0和压缩模量E_(s1-2)进行考虑软土侧向变形的非线性沉降计算,为工程计算带来极大的方便。最后通过案例说明方法的可行性。

Simplified method for nonlinear settlement calculation in soft soils considering lateral deformation

The common method for computing settlement of soft soils is the layer-wise summation method using compression modulus under the confined condition. But the lateral deformation makes a large part of the whole settlement. To eliminate the calculation error, the method in the national specifications or codes is that the results are multiplied by an empirical coefficient from 1.1 to 1.7. Since the empirical coefficient is not sure, a new method considering the lateral deformation is proposed. This method divides the settlement of soft soil into two parts. One is the settlement S_c caused by 1 D confined compression using the layer-wise summation and e-p curve to compute, the other is the settlement S_d caused by lateral deformation using the same way but the parameter is secant modulus E_p from the e-p curve and Duncan-Chang model. So the e-p curve is the core for computing the settlement. In practical engineering, the initial void ratio e₀ and compression modulus E_s1-2 are frequently used and can reflect the characteristics of the soils well. The two parameters are combined with compression index C_c, thus we can use e-lgp curve to get the compression modulus E_si under different stress states. Since E_si, S_c, S_d are easy to get, it brings a big convenience to engineering computation. Finally, the applicability of the proposed method is validated through an engineering case.