Summary: | 碩士 === 國立勤益科技大學 === 工業工程與管理系 === 105 === In recent years, Germany proposed Industry 4.0, combined with Internet of things (IoT), Cloud Computing, Big Data, intelligent manufacturing and other automated applications. Our government is also actively promoting the manufacturing machinery, metal processing, medical and other industries, to reform Industry 4.0 Enterprise Engineering. Give advices for enterprise transformation to increase the international competition of enterprises, the kernel of this study is to improve the production of medical equipment production equipment. Use medical equipment intelligent automation technology and process to provide case companies transform into small and medium-sized enterprises with advanced industry 4.0 advanced technology. To help the industry through the development of medical equipment to develop automated production technology and process the production yield is high, good quality and high production rate medical equipment.
First, according to interrelated document by Automatic optical detection(AOI) abnormal or misjudged influencing factors to search influencing factors. And then through the expert interviews please expert in connection with the factors for the selection of points. Filter to the key factors and Coordinate Tolerance Level. After through the data to explore a large number of data screening, use the Back-Propagation Network to compare the difference between the XY coordinates before and after screening and the data before and after screening. Using the Two-stage clustering method to verify the comparison. Verify the consistency of the results of each rating level. At last, use expert interviews screened five reasons for the impact of automatic optical detection anomalies to collect data. Through data, the detection of the number of abnormalities (Level A and Level E) is about 8%. Through explore with the experts to find out the reasons for the amendment, by improving the stability of AOI automatic optical detection, the number of abnormal detection significantly reduced to about 4%. This study effectively reduces the number of detection abnormalities occur, and then improves the detection quality and accuracy, and produce better quality products.
|