An integrated streaming algorithm for real-time monitoring and prediction of open caisson sinking using multi-GNSS and earth pressure data
摘要
Open caissons are extensively employed in various underground engineering projects, necessitating real-time determination and prediction of their sinking states to ensure stable and secure sinking and prevent potential construction risks. Recognizing that the monitoring data of open caisson sinking is inherently streaming data updating in real-time, this study proposes an integrated RRCF-HMA-ERT streaming algorithm and establishes a practical engineering framework based on the algorithm. This framework gathers multi-constellation GNSS RTK and earth pressure monitoring data during open caisson sinking. The collected streaming data are then processed by the proposed algorithm in real-time, accurately determining the current sinking states and predicting future sinking states of the open caisson. Taking a large-scale open caisson project in Jiangsu Province, China as an example, a complete workflow for applying the constructed framework is illustrated, and the proposed algorithm and framework are validated. Results demonstrate that the algorithm and framework can accurately determine the sinking states (sinking speed, inclination, and horizontal deviation). When predicting various sinking states in real-time, the values of RMSE are all less than 2.1 and R2 values exceed 0.93, indicating high prediction accuracy. The average calculation time for each analysis round is 0.07 s, showcasing the streaming algorithm’s speed and efficiency.