The US Bureau of Reclamation has selected seven winners from the first phase of its Crack the Case Challenge, which is seeking new technologies to detect subsurface cracks in embankment dams.
The seven teams will receive prizes and advance to Phase 2 of the challenge, where they will refine and validate their proposed technologies and develop them into testable systems.
During the second phase, teams will address technical risks identified in Phase 1 and demonstrate the feasibility of their approaches through analysis, modelling, prototyping or testing. They will also define system architecture, performance targets, deployment plans and operational requirements for use in embankment dam environments.
The aim is to establish sufficient technical evidence and confidence for Phase 3 demonstrations.
The seven Phase 1 winners are:
Integrated DAS & EM-SP Crack Detection for Dams, submitted by Mahdi Haddad, combines electromagnetic-streaming-potential techniques with passive seismic interferometry using distributed acoustic sensing (DAS). The proposed system is designed to measure deformation and flow-related responses within a dam and detect millimetre-scale cracks at depths of up to 50m.
GAIA – Geophysical Anomaly Inference Algorithm, submitted by TAB Technologies, uses physics-informed, multimodal artificial intelligence to combine electrical resistivity tomography and seismic data. The system is trained using a synthetic dataset generated by DamForge, the team’s simulation engine, to address the lack of real-world ground-truth data.
Dam Monitoring from Space, submitted by the InfraSAR team, proposes using interferometric synthetic aperture radar (InSAR) to monitor changes in differential settlement that can create soil tension and subsurface cracking in embankment dams.
Nonlinear Dynamic Crack Signature Tomography, submitted by Ziga Gosar, uses active excitation and distributed fibre-optic sensing to identify hidden cracks through nonlinear dynamic signatures, including wave distortion and energy loss. The approach is intended to map crack location, depth and connectivity.
Surface Net Acoustic Computed Tomography (SNA-CT), submitted by HeroGuy’s team, uses a distributed network of active acoustic sensors to generate high-resolution 3D maps of subsurface materials. The system is designed to identify variations in density, material stiffness and water content using compact sensors that can be deployed by a single operator.
Surface Deformation Laser Array, submitted by Team Hydra, uses laser-linked poles to measure surface displacement and rotation associated with subsurface cracking. Machine learning is used to classify crack location and activity based on changes including static strain, soil softening and swelling.
CrackSense: Fiber-Optic Early Crack Detection, submitted by ESG’s team, uses a single fibre-optic cable as a continuous sensor along an embankment dam. Distributed strain sensing is used to identify localised strain changes, while DAS monitors broader changes in stiffness and wave speed. The data are combined to produce a crack-likelihood map for follow-up inspection and risk response.
The Bureau of Reclamation said the challenge is intended to support the development of improved methods for detecting subsurface cracks that can affect dam performance.
