Classification of knitted defects and development of a defect diagnosing expert system for knitted fabrics

In textile products, several defects and non-conformities occur because of various uncontrolled factors in fabric production. The determination of the source of these faults are important for solving the problem and preventing the recurrence of defects. Every individual fault affects the fabric quality in negative way and also slows down the production. In order to solve the problems efficiently, an access to the knowledge of expert is vitally important. A chance of reaching expert people is difficult due to the limited number of experts and to relatively higher costs. Hence, this project is focused on the development of an expert system which can diagnose fabric defects and provide solution for remedy of the faults. A systematic classification study based on different criteria was made for knitted fabric faults, however, the source of the error is mainly classified on the basis of processes. In this study, a total of 108 errors have been defined and an expert system has been developed which includes the reasons for its emergence and removal options. Thanks to the development of such an expert system, fabric manufacturers shall be able to use a common terminology for identifying fabric defects and shall have a chance of reducing the manufacturing cost by utilising the hints for removal of faults. The photographs of fabric defects are also incorporated into the expert system to facilitate the matching of fabric defects.