深基坑施工变形预测与控制方法研究

    Deformation predictions and control methods for excavation of deep foundation pits

    • 摘要: 结合广州地铁某换乘车站深基坑工程,分别采用5种计算模型对深基坑施工引起的周边地表沉降进行了预测分析,与实测沉降值对比后发现:灰色GM(1,1)模型、灰色马尔科夫链模型和BP人工神经网络的短期预测结果比较可靠,但其长期预测结果精度不够,而经过残差修正后的灰色模型能够明显的提高预测精度,具有一定的工程实用价值;并结合具体工程实例提出了深基坑施工变形控制的基本方法。

       

      Abstract: The excavation deformation of deep foundation pits is predicted by 5 kinds of different models. Compared with the test data, the results show that the short-term prediction accuracy of the metabolic GM(1,1) model, grey Markov chain model and BP model can meet the requirements of the engineering. But their long-term prediction accuracy is unreliable. It is proved that the residual GM(1,1) model and the grey Markov chain model are suitable for the medium-term and long-term predictions. The excavation deformation control methods for deep foundation pits are summarized

       

    /

    返回文章
    返回