Fault detection in braiding utilizing low-cost USB machine vision

This paper investigates the effect of yarn tension on braid formation point (braid point) motion. A computer-controlled take-up machine is developed to facilitate braiding experiments. The results of several experiments are used to recognize tension aberrations that lead to poor quality and wasted product. Optimal braid performance is observed, which serves as the baseline for comparing the behavior of faults. By studying the effects of common faults, a diagnostic tool is developed to recognize the onset of defects and provide some insight into what might be causing the fault. Radial fluctuation of braid point position is a good indicator of mechanical faults of the tensioning mechanisms. Observations based on mechanical and visual methods are also presented as diagnostic tools. From visual observations, it can be concluded that as the tension in one yarn increases with each revolution, the radial fluctuation increases until a yarn breaks. From mechanical observations, it can be concluded that the fluctuation in motor speed increases until the point when the yarn breaks. Through this study, braid point motion can be used as a diagnostic for shutting down the machine before irreversible damage occurs.