Research on structural health assessment methods for long‐distance underground culverts under multi‐source data collaboration

Abstract During long‐term operation, long‐distance underground culverts frequently undergo scouring and erosion from high‐velocity, sand‐laden water, causing issues like cracks, material degradation, and leakage. Traditional manual inspection methods are inefficient, risky, and have significant limitations for operating and maintaining such culverts. In recent years, intelligent sensing technologies and underwater robotic inspection methods have been increasingly adopted for these projects, though challenges persist. This paper introduces an underwater detection method using “acoustic‐optical” fusion diagnostics. This method employs image analysis to inspect sediment accumulation and surface defects in culvert concrete, simplifying the description and laying the foundation for constructing digital twin models of the structure. Additionally, a non‐destructive rebound testing technique is proposed to measure underwater concrete strength. Its validity is confirmed through underwater core sampling and compressive strength tests, verifying the rebound test results. Ultimately, a structural health assessment method for long‐distance water conveyance culverts is established by combining mechanism‐driven and data‐driven approaches. This method has been validated in a long‐distance underground culvert. The research methods provide technical support for the long‐term operation, maintenance, and health assessment of similar structures, showing promising application prospects.