Monitoring of Self-Tapping Screw Fastenings Using Artificial Neural Networks

Abstract

Screw fastenings account for a quarter of all assembly operations and automation of the process is highly desirable. This paper presents a novel strategy for monitoring this manufacturing process, focusing on the insertion of self-tapping screws. An artificial neural network (ANN), using “Torque-versus-Insertion-Depth” signature signals as input, is designed to distinguish between successful and failed insertions. The ANN is first tested using simulation data from an analytical model for screw insertions, and then validated using experimental torque signals obtained from an electric screwdriver. The results demonstrate that ANNs can effectively monitor the screw fastening process and cope with a wide range of insertion cases interpolating for unseen insertion signals.