case studies

AND PRESENTATIONS

Case Study: Offshore Infrastructure Imaging After Hurricane Ivan

Following Hurricane Ivan, approximately 25 oil-well conductor pipes were displaced and ultimately buried beneath approximately 35–45 m (115–150 ft) of subsea sediment, creating a difficult infrastructure-imaging and mitigation challenge.
EmPact AI combined high-resolution magnetic data acquired by an ROV with AI-driven 3-D inversion to reconstruct conductor geometry where acoustic imaging alone could not establish continuity toward the well pad; a significant extension of the ZeroDig Pro® technology.
The AI-derived reconstruction showed the main conductor body curving toward the well bay and achieved agreement of approximately 1 m (3 ft) with independent drilling contacts and acoustic imaging in the overlap region—providing additional evidence to support engineering mitigation decisions.

ZeroDig Pro® was applied to magnetic field data collected at multiple Texas sites containing buried metallic infrastructure.
Compared with conventional inversion, AI-driven imaging demonstrated 6X + improvement in vertical resolution and provided detailed 3-D visualization of subsurface structures.
At the Texas A&M RELLIS site, excavation results verified an 8-inch buried pipe at 36–42 inches depth, with the AI-driven image achieving an average calibrated uncertainty of approximately ±7 inches.

At Hollister Ranch near Santa Barbara, California, landowners needed better information about an abandoned buried pipeline to support decisions about potential intervention or removal.
ZeroDig Pro® analyzed drone-acquired magnetic data to generate a 3-D subsurface image of approximately 300 m of buried pipeline, revealing its location, geometry, jogs, bends, and variations in the magnetic response along the pipeline.
It demonstrated how AI-driven imaging can transform magnetic survey data into actionable subsurface information to support evaluation of difficult-to-characterize buried infrastructure. Read more here.

Presented at SEG/AAPG IMAGE 2021, this demonstration explores AI-driven 3-D geophysical inversion using synthetic datasets of increasing size and complexity.
The results demonstrate the ability of machine-learning inversion to recover the location, geometry, and spatial extent of subsurface targets from geophysical data.
The work helped establish the technical foundation for applying EmPact AI's approach to increasingly complex problems and, ultimately, real-world field datasets.