Author(s)

Xiangsu Zeng

Corresponding Author

Xiangsu Zeng

Abstract

In order to cope with the multiple challenges faced by water supply and drainage pipe network construction in complex urban environment, such as complex geology and dense underground space, and to solve the shortcomings of traditional methods in data fusion, dynamic response and multi-objective coordination, this paper puts forward a risk prediction and dynamic regulation method for water supply and drainage pipe network construction in complex urban environment. This method constructs a multi-source heterogeneous data fusion layer, integrates geological and underground space, environment and society, real-time monitoring and construction process data to form a standardized data set; Building a dynamic risk knowledge graph (KG) based on fused data to achieve formal expression and dynamic updates of construction related entities and relationships; Adopting a hybrid intelligent prediction model that integrates finite element method physical model and physical information neural network (PINN) deep learning model to improve the accuracy of risk prediction. At the same time, a three-level hierarchical control system of "on-site edge cloud" is designed, combined with NSGA-III algorithm to achieve multi-objective collaborative optimization, and integrated to build an intelligent decision support system. The engineering application case shows that this method can significantly reduce the maximum settlement by 33%, avoid the alarm of settlement overrun and delay in construction period, increase the average daily driving speed by 20%, effectively realize the accurate prediction and dynamic regulation of construction risk, and provide intelligent solutions for similar projects.

Keywords

Risk prediction, dynamic control, water supply and drainage, pipe network, complex urban environment, knowledge graph