A Defect Prevention Concept Using Artificial Intelligence

Abstract Defect prevention is particularly critical in operations such as aircraft assembly or service. Failure Modes and Effects Analysis (FMEA) procedures have been deployed manually for many years. However, the manual procedures fail to utilize capability to build intelligence into inspection processes that can facilitate elimination of human error. In this work, we introduce an artificial intelligence (AI)-based concept that can iteratively learn to assure zero defects from a given inspection process. This work introduces a schema that can serve as a knowledge management framework in a relational database for instantiation with inspection process information and data from a detection system. A companion algorithm is presented for the case of a wiring harness bracket installation in a fuselage. The schema and algorithm analyze and assess potential defects posed by Foreign Object Debris (FOD) in parallel to the assembly inspection. A closed loop of logic was introduced to enable anomaly detection by this algorithm to assure zero defects.