Integrated Automatic Inspection in Robotic Composites Cells

Abstract

Real-time process control, quality verification and process enhancement are essential in automated composites manufacturing applications. Automatic inspection generates the data to optimize processes as well as ensuring that each part produced is fully within tolerance. New laser and imaging systems are being used in various robotic manufacturing cells on parts ranging in size from small complex parts (a half meter square) up to the full-size wing skins and spars for the Boeing 777X. The automatic inspection systems verify and document correct material location, fiber orientation and shear, flatness on surfaces; including detecting wrinkles, bridging and secondary bridging. In addition to ensuring that each part meets specification, realtime inspection systems can function as learning nodes on large, deep learning networks. Having immediate, comprehensive data on flaws, as they form, enables prediction and prevention of the conditions that resulted in flaw creation across all systems on the network that are performing similar tasks. The availability of detailed, as-built data enables analysis of real-world, rather than nominal characteristics when fed back to the design model. This feedback to the model, as well as process verification on a part by part basis, reduces uncertainty and overdesign. The implementation and operation of the automatic inspection systems performing these functions are described.