基于LSTM深度学习模型的大跨度多孔框架桥顶进预测技术研究

Research on Jacking Prediction Technology of Long-Span Multiple Box Culvert Based on LSTM Model

在进行大跨度多孔框架桥顶进施工时,在顶进过程中箱体所受顶力往往左右不均,导致轴线偏移,工程质量以及施工安全因此受到影响。为解决这一问题,论文提出了基于LSTM深度学习模型的大跨度多孔框架桥智能顶进技术,该技术通过对框架桥实时顶进姿态数据的分析与学习,实现对其轴线偏差值的实时预测,从而为框架桥顶进姿态的纠偏动作提供参考。

During jacking construction of a long-span box culvert,due to uneven force on the box body,axis deviation is likely to occur,which affects the quality and safety of the project.To solve this problem,the technology of intelligent jacking of long-span multiple box culvert based on LSTM model was established,which could analyze and learn the axis deviation value in real time,then judge the future trend of the axis through the predicted value,forming a reference for the intelligent control of the box culvert jacking attitude.