An application of gray theory on process parameter for molding press

• Main Authors: Wen-Chieh Chen, 陳文潔
• Other Authors: 陳智勇
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/61237345186011299531

碩士 === 樹德科技大學 === 電腦與通訊系碩士班 === 104 === This thesis aims to apply the gray theory to predict the trends of molding press parameter in the semiconductor manufacturing process. The proposed approach can help engineers to analyze molding press parameter and determine the status of the machine cavities, before exception occurs, early preventive maintenance machine cavity to avoid creating a defective product and improve process yield. In this study, based on the characteristic, only four data records is required for forecasting by using gray theory GM (1, 1) model, the raw data of mold clamping pressure actual voltage output are cumulated to establish a first-order differential equation and prediction model. Then, coming records from the fifth one can be obtained by the white response and inverse accumulated generating operation (IAGO). Six actual datasets with 22-days data are used to evaluate the performance and several common prediction methods including linear regression, exponential smoothing, moving average method are compared with each other in the experiments. Finally, many various types of error test methods experiment on the effectiveness of prediction methods applied to compression molding machine parameters predictive problem. From the experimental results, Grey theory shows the best effectiveness in all error test methods. The characteristic without the necessity of history training data, which is suitable for the production parameters quite dramatic changes in the semiconductor factory machine.