Twinfra: Digital Twinning Infrastructure From Damage to Repair Package in Minutes

Enhanced Inspection

Problem Statement: When damage is found on a ship, aircraft, vehicle, or facility, seeing the damage is often not the bottleneck. The delay comes from assembling the information engineering needs to define the repair.

Maintainers capture photos, take measurements, document locations, search manuals and parts catalogs, and relay findings to engineering. If key evidence is missing, personnel may need to return to the asset and repeat the process. These fragmented handoffs consume time and manpower during urgent maintenance.

Innovation Solution: twinfra helps the maintainer build an engineering-ready Repair Package while still at the damaged asset. A single capture creates an accurate, interactive Real-Time 3D record within minutes. From that record:

• AI-assisted Auto Inspect can locate and annotate damage.
• Quantity Measurements can quantify cracks, corrosion, damaged areas, clearances, and component dimensions.
• Comparison and Auto Change Detection can identify changes relative to earlier captures.

The Repair Package can consolidate:

• Damage location and annotated visual evidence
• Measurements and affected components
• Prior-condition comparisons
• Likely replacement parts
• Unresolved questions requiring engineering review

twinfra also checks whether enough information has been captured before the maintainer leaves the space. For a damaged or missing component, it can use imagery, dimensions, geometry, markings, and location to narrow a parts catalog to likely matches.

If multiple candidates remain, twinfra can identify the additional photo, marking, or measurement needed to distinguish them, reducing return trips and back-and-forth with engineering.

Benefits to the Department of War: twinfra unifies capture, assessment, measurement, repair definition, and part identification in one workspace.

• Reduces documentation and measurement effort
• Shortens the path from damage discovery to repair decision
• Reduces repeat visits caused by missing information
• Creates a persistent record for change tracking and post-repair verification
• Supports multiple asset classes without requiring a platform-specific digital model

Innovation Challenges: Priority challenges are integrating heterogeneous technical manuals and parts catalogs; validating AI assistance across platforms, environments, and damage types; and ensuring every recommendation is traceable to source evidence and suitable for human engineering review.

Technical Maturity / Demonstration Results: twinfra has been demonstrated with portable field capture, edge computing, remotely operated platforms, and underwater data. Interactive 3D scenes have been produced within approximately five minutes of completing capture.

In May 2026, twinfra was demonstrated during the Pacific Fleet Repair Training Exercise (REPTX) in Hawaii, providing an operationally relevant