Monitoring Data-Driven Dynamic Safety Assessment Framework for Deep Foundation Pit Construction Based on Grey Clustering and Moment Method

Pengliang Dang¹Zeliang LI^2*Dehai Zou¹Hangjun Li¹Zilong Cheng²Le Chang¹Yadong Lu¹

• ¹Sinohydro Bureau 14 Co., Ltd., Kunming, China
• ²Shenzhen University, Shenzhen, China

With the increasing scale and complexity of underground construction, ensuring the safety of deep foundation pit engineering has become a critical challenge. This paper introduces a data-driven dynamic safety assessment framework for deep foundation pit construction. By integrating grey clustering theory and the moment estimation composite weights, the framework address uncertainties and incompleteness in monitoring data. The study aims to enhance the accuracy and timeliness of risk assessments during excavation processes. The methodology constructs a three-level dynamic safety assessment index system, including 21 indicators. These indicators are categorized into three groups: structural deformation, structural internal forces, and environmental safety factors. Subjective weights are assigned using the order relationship analysis method, while objective weights are calculated via the entropy weight method. These weights are optimized into composite weights using the moment estimation method to ensure scientific rigor and objectivity. Grey clustering analysis classifies safety levels dynamically based on real-time monitoring data. Validation through a case study of a large-scale water diversion shaft project in Shenzhen demonstrates the framework's applicability and reliability. Results show consistent alignment between computed safety indices and observed field performance, effectively identifying critical risks at different excavation stages.