사무용 인쇄 기기 급지 시스템 설계를 위한 신경망응용 파라미터 분석

A printer is a device which produces a hard copy of documents stored in electronic form on physical print media. It feeds the media, develops the image on it, fuses it, and then exits it for users. Feeding unit has roles to feed single media from a cassette and to transfer it to a developing unit. Costumers demand fewer jam-rate and multi-feeding rate, which are important measurement for feeding unit. However, feeding unit is a complex system, which is consist with many components. They have factors which can not be controlled in manufacturing process. Each component has a various design parameters, relative with feeding quality and it is difficult to find the critical design parameters. In this paper, feeding unit is designed and its factors related with feeding quality are defined. With results of design of experiments, major design parameters are identified using back-propagation neural network based sensitivity.