A&P Technology advances U.S. Air Force defense composites with AI-inspected overbraiding process

A&P Technology, a Cincinnati-based braiding specialist, has advanced its overbraiding process for the U.S. Air Force under a Small Business Innovation Research (SBIR) contract, integrating AI-driven inspection to improve quality assurance for defense composite structures. The company demonstrated automated defect detection during the overbraiding of complex geometries, targeting applications such as aircraft structural components and missile casings. The AI inspection system, developed in partnership with an undisclosed machine vision provider, identifies tow misalignment, gaps, and fiber damage in real time, reducing post-process NDT requirements. This work aligns with Air Force Research Laboratory (AFRL) priorities for scalable, repeatable composite manufacturing for下一代 platforms.

This development matters because it addresses a persistent bottleneck in braided composite production: manual inspection of dry-fiber preforms. Overbraiding - where a braided sleeve is formed directly over a mandrel or core - offers near-net-shape preforms with continuous fiber architecture, but defect detection has historically relied on post-consolidation ultrasonic or CT scanning. A&P’s AI-inspection layer shifts quality control upstream, potentially reducing scrap and rework costs that can reach 15-20% of total preform cost in defense programs. The work fits the broader pattern of digitizing composite manufacturing, where sensor fusion and machine learning are being applied to processes like automated fiber placement (AFP) and filament winding. Competitors such as Electroimpact and Coriolis Composites have pursued similar in-process monitoring for AFP, but A&P’s focus on braiding - a less automated but geometrically flexible process - opens a distinct niche. The U.S. defense market for advanced composites, estimated at $4-5 billion annually, is increasingly demanding domestic, auditable production lines, making this SBIR work strategically relevant beyond its modest contract value.

From a practical standpoint, A&P Technology must now transition from lab-scale demonstration to production-rate qualification - specifically, proving that the AI model generalizes across different fiber types (carbon, glass, aramid) and braid architectures (diamond, regular, Hercules). The Air Force will likely require statistical process control data across at least 100 consecutive defect-free parts before approving the system for critical-structure use. For buyers in defense prime supply chains, this signals that overbraiding is moving toward the same quality-governance maturity as tape laying, but the qualification timeline remains measured in years, not quarters.