基于S型组合曲线的隧道洞口段施工地表沉降预测分析
Analysis of Surface Settlement Prediction of Tunnel Portal Section Construction Based on the S-Shaped Composite Curves

隧道洞口段施工易引起上覆地表大面积沉降，对洞口边坡稳定产生较大威胁。施工引起的地表累积沉降通常呈现为S型变化规律，选取5种S型曲线(Logistic、Gompertz、Richards、Weibull及MMF曲线)分别运用于黄泥岗头隧道洞口段施工诱发地表沉降预测，分析各单一模型的预测效果。利用加权平均算法获得5种曲线的组合权重值，构建组合S型曲线预测模型。结果表明：MMF模型在5种单一S型曲线中拟合效果最好；通过权重优化得到组合模型，其拟合评价R2及RMSE值均优于单一模型，提高了地表沉降预测精度，验证了该方法的地表沉降预测有效性。
The construction of tunnel portal section is easy to cause a large area of settlement on the overlying surface, which poses a great threat to the stability of tunnel entrance slope. The surface accumulated settlement caused by construction usually shows an S-shaped variation law. This paper selects five commonly used S-shape curves (Logistic, Gompertz, Richards, Weibull and MMF curves) and respectively applies them to the prediction of ground surface settlement caused by Huangnigangtao tunnel, and the prediction effects of each single model are analyzed. The weighted average algorithm is used to obtain the combined weight values of five curves and the S-shaped composite curve is built. Results show that MMF has the best fitting effect among five single S-shaped curves. The combined model was obtained by weight optimization, and its fitting evaluation R2 and RMSE values were better than that of the single models, which improved the prediction accuracy of surface subsidence and verified the validity of the proposed method.