基于PSO-SVM模型的新建盾构隧道下穿既有地铁运营线路沉降动态预测方法

PSO-SVM Model-based Settlement Dynamic Prediction Method for New Shield Tunnel Undercrossing Existing Operational Subway Lines

  • 摘要:
    目的 新建盾构隧道下穿既有运营地铁隧道时会引发沉降,进而影响既有运营隧道的安全。目前研究大多集中在盾构下穿既有道路、隧道或建筑物的地面沉降预测及控制等方面,而对于双线盾构下穿既有运营隧道轨道沉降预测的研究相对较少,有必要对此进行深入研究。
    方法 以合肥市轨道交通2号线东延伸段新建隧道近距离下穿既有运营2号线隧道为工程背景,建立了PSO(粒子群优化)-SVM(支持向量机)预测模型,介绍了PSO-SVM模型动态预测流程。选取了9个典型的监测断面,基于监测数据训练模型,动态预测了新建隧道下穿既有运营隧道的沉降量。
    结果及结论 基于PSO-SVM预测模型,各监测断面的预测结果误差平均值为8.31%,进而验证了PSO-SVM模型动态预测的可行性。

     

    Abstract:
    Objective When new shield tunnels cross under existing operational subway tunnels, settlement will be induced, further affecting the safety of the existing operational tunnels. Current researches predominantly focus on predicting and controlling the ground settlement induced by shield tunneling underneath existing roads, tunnels, or buildings. In contrast, there are relative less studies on the settlement prediction of twin shield tunnels undercrossing an existing operational tunnel track. Therefore, it is necessary to conduct an in-depth research on it.
    Method Based on the engineering background of Hefei Metro Line 2 Eastern Extension’s new tunnel, which crosses beneath the existing operational Line 2 tunnel in close-range, a PSO (particle swarm optimization)-SVM (support vector machine) prediction model is established, and the PSO-SVM model dynamic prediction flow is introduced. The model is trained based on the monitoring data collected from nine typical monitoring sections, and the settlement value induced by the new tunnel undercrossing the existing operational tunnel is dynamically predicted.
    Result & Conclusion  Based on the PSO-SVM prediction model, the average value of prediction result errors across various monitoring sections is 8.31%, further verifying the feasibility of the PSO-SVM model for dynamic prediction.

     

/

返回文章
返回