Sonibel Instruments provides a torch‑mounted acoustic sensor that captures the weld‑arc’s acoustic signature and uses AI to detect defects such as porosity, lack of penetration, lack of fusion, and slag inclusions in real time.
Funding
Funding not disclosed
Founders
Product
Problem
Weld inspections often rely on post-process nondestructive testing, which requires stopping production, removing welds, and incurring additional labor and material costs. Delays caused by re‑inspection and rework extend project timelines and increase overall fabrication expenses.
Solution
Sonibel Instruments attaches a compact, rugged acoustic sensor directly to the welding torch, capturing the acoustic signature of the weld arc in real time. An AI‑driven model analyzes this data on‑site to identify common defects such as porosity, lack of penetration, lack of fusion, and slag inclusions as they occur. The system provides immediate feedback to welders, enabling them to correct issues on the spot and avoid downstream rework. By integrating seamlessly into existing FCAW semi‑automatic workflows without altering the welding process, the technology reduces reliance on separate NDT cycles, shortens production delays, and lowers overall fabrication costs.
Target Audience
Primary customers are welding shops and fabrication facilities in shipbuilding, structural steel construction, and oil & gas sectors that perform high‑volume FCAW welding and seek to reduce rework and inspection costs.
Features
- Torch‑mounted acoustic sensor that clamps onto the torch without requiring workflow changes
- Real‑time AI analysis of acoustic data to detect porosity, lack of penetration, lack of fusion, and slag inclusions
- Optimized machine‑learning models for FCAW semi‑automatic welding processes
- Immediate on‑site defect alerts delivered to the welder for instant correction
- Rugged design capable of operating in high‑noise shipyard and industrial environments
- Compatibility with existing welding equipment and procedures, preserving current operator skills