Automated structural defects diagnosis in underground transportation tunnels using semantic technologies

Detecting structural defects and finding the underlying causes accurately and timely is crucial for developing effective maintenance strategies to keep up tunnel safety and availability. Data-driven methods are gradually becoming the norm for tunnel health diagnosis but data involved are usually limited to a single system or a particular type of defect. These methods are not effective for cross-system or multi-stage detection because of the following difficulties: (1) high data heterogeneity, (2) complex spatiotemporal relationships, and (3) high expert knowledge involvement required. In order to overcome these challenges, this paper follows the constructive research approach to develop a system called Tunnel Defects Diagnosis System (TDDS) based on Industry Foundation Classes (IFC) and Semantic Web technologies. In TDDS, a meta-standard is introduced to set up mapping rules for the integration of heterogeneous data and a Tunnel Diagnosis Ontology (TDO) is established to formally define the complex spatiotemporal relationships among data. Predefined rules based on expert knowledge enable automatic reasoning to provide support for decision making with respect to cause detection. This system has been applied to the Dalian Road Tunnel in Shanghai to demonstrate its feasibility and effectiveness.