This Is AuburnElectronic Theses and Dissertations

A Machine Vision and Sensing System for Braid Defect Detection, Diagnosis and Prevention During Manufacture

Date

2007-12-15

Author

Branscomb, David

Type of Degree

Thesis

Department

Mechanical Engineering

Abstract

The mechanical properties of braids are often quite convenient for achieving the high performance levels expected of composite materials. A favorable attribute of a braided structure is the ability for the design engineer to tailor reinforcing fiber position and orientation in such a way to achieve a desired geometry. Braiding is of particular interest to the modern manufacturer as new methodologies are required to meet the increasing performance demands of engineered materials technology. In this presentation, a computer controlled machine is designed, constructed, utilizing PC based motion control and machine vision technology to control the take-up process. The computer controlled system presented is capable of observing braid formation, monitoring braiding machine performance, and finally provide diagnostic measures to recognize and ultimately prevent the manifestation of common mechanical faults. Both visual and mechanical approaches to fault diagnostics are developed. Experiments are carried out in order to characterize the nature of the braiding dynamics and establish the optimum machine operating conditions as a baseline for comparison. The dynamics occurring at the braid formation point undergo three distinct regimes as the braid reaches steady state. It is observed that depending on the initial conditions, transients and migration toward the equilibrium point occur as the braid point reaches steady state. Extensive tests were performed in order to investigate the effects of common mechanical faults (high tension yarns) during the braiding process; particularly problems related to the proper function of yarn package tensioning mechanisms are investigated. The results of these experiments indicate the dominating influence that high tension yarns have on braid point formation. Knowledge of braid point dynamics is particularly useful if improving response time and reducing product waste is desired. Understanding the nature of braiding and common problems encountered during the braiding process provides a useful foundation necessary for developing control methods to mitigate the adverse effects of braiding faults.