Efficiency of applied supervised intelligent technique of extracting yarn faults in actual textile production in Ethiopia

Abstract

In fact, major textile and garment problems that have their origins in yarn manufacturing are still available as feedbacks from the weavers, knitters, traders and retailers show. Yarn faults are among the major problems which cause a direct effect on profit and quality of a yarn product. To solve the yarn faults, there should be a proper diagnosis and rectification measures taken by the industry workers. However, there are challenges to these industries including skill limitation of employees, maintenance delay, lack of training, shortage of manuals, mental workload, foreign language understanding problems and scarcity of human experts. The experts‟ interview has been used to develop the rule-based prototype called YFDRES (Yarn Fault Diagnosis and Rectification Expert ) using SWI-Prolog and Visual Geez tools. In order to tackle the yarn faults, domain knowledge is modeled using decision tree which represents concepts, parameters and procedures involved for yarn fault diagnoses and rectification. Knowledge of yarn faults at the prototype is coded by means of production rules. Therefore, the researchers concluded that implementing expert for yarn fault diagnosis is a useful technology and the results of this study could be applied for the development of full-fledged expert.

Keywords: Diagnosis, Yarn, Faults, Textile, Garment, Industry, Rule-Based, Expert