پديد آورندگان :
صادقي ، عبدالحسين نويسنده Sadeghi, A.H. , اماني تهران ، محمد نويسنده دانشكده مهندسي نساجي- قطب علمي و پژوهشكده نساجي- دانشگاه صنعتي امير كبير Amani Tehran, M , همتي، نوشين نويسنده كارشناسي ارشد تكنولوژي نساجي, Hemati, N
كليدواژه :
درجه بندي , سامانه خبره , فرش , موتوراستنتاج , پايگاه دانش , عيب
چكيده لاتين :
Woven carpet is one of the most important products of the Iranian textile industries. The final product may include nonconformities, due to process faults. Therefore, carpet is graded at final inspection. Quality control of the carpet is still followed by observational actions. Normally, such human actions are affected by feelings, unavoidable mistakes, fatigue, slowness, etc. In addition, a quality controller must keep in his mind all carpet faults and their intensities simultaneously, and this is not an easy work.
According to these facts, in this research, an expert system was developed, which can grade the carpet after declaring the faults. Contrary to conventional evaluation system, in which only the greatest faults is considered, in the suggested system all faults are evaluated aggregately, and the suitable grade is devoted.
The knowledge of the system has been gathered from carpet manufacturing factories by interviewing their experts, and specified according to rule based representation. After entering the faults data by the user, the system processes the input data by choosing the proper rules from the defined knowledgebase, and at the end presents the resulted grade. Moreover, the system can show the inference used during carpet grading. To examine the validity of the designed system, 100 carpets evaluated by both suggested and conventional observational systems, and excellent agreement between them were noticed.