The U.S. Navy awarded Senvol funding for a project titled Additive Manufacturing Sensor Fusion Technology for Process Monitoring and Control. The project is a multi-year initiative starting in July 2025 and continuing until July 2027. This project applies Senvol ML to multi-sensor field data collected from wire directed energy deposition (DED) to predict mechanical performance and help standardize data-driven part acceptance protocols.
Senvol fuses signals from multiple sensor streams collected during DED construction to produce a model that correlates process signatures with part-level characteristics. The primary goal is to reduce expensive and time-consuming qualification testing by demonstrating that sensor fusion and machine learning can provide sufficient evidence of a part’s performance to make installation decisions. Project results are intended to inform NAVSEA policy and expand the Navy’s ability to accept qualified additively manufactured parts from a broader supplier base.
For the printing industry, this is important in a concrete way. A successful demonstration could shift some of the acceptance efforts from destructive testing and lengthy qualification campaigns to validated field monitoring records and predictive analytics. This impacts lead times, scrap rates, and operating economics for metal DED shops wishing to supply defense contracts. Shops that have already deployed melt pool cameras, pyrometers, acoustic sensors, or fringe-based monitoring should prioritize synchronized timestamps, robust metadata, and retention policies so that field records can support traceability and downstream ML workflows.
The technical challenges are not easy. Sensor fusion requires high-quality, time-aligned data, consistent calibration, and repeatable build parameters across machines and suppliers. Senvol ML’s role is to ingest multi-sensor streams and generate performance predictions that NAVSEA and inspectors can trust. Standardization of data formats, reporting methods, and acceptance thresholds is central to turning sensor evidence into policy-ready evidence.
For community labs and small shops, this project is a signal to get started. Be prepared by documenting current monitoring capabilities, validating sensor calibration, and archiving build logs in machine-readable format. Suppliers who cannot readily adapt a complete sensor suite may still benefit as the project investigates what minimum sensor combination and analysis will yield acceptable reliability.
Sembol’s research currently has a two-year lead-in period for concrete results. If this initiative is successful, we hope to see new guidance from NAVSEA and widespread acceptance of data-driven parts. This lowers the barrier for qualified suppliers and potentially accelerates the path from prototype to assembled part.
