| Περίληψη: | This study presents a complete analysis revolving around the near zero-defect manufacturing at process level. The computational modeling results of the rubber extrusion manufacturing process are presented, where the impact of the die swell phenomenon is quantified by running iterative models, differentiating the inflow rate. The extrudate’s dimensions are also identified by making use of image processing algorithms for detecting the edge. In addition, the velocity values at the outlet are calculated for various inflow rates and are presented graphically. The generated rules are correlating the manufacturing parameters with the Key Performance Indicators (KPIs) and can predict the extrudate’s dimensions towards zero defect manufacturing. Furthermore, an experimental study took place for classifying defected and non-defected profiles in terms of dimensional accuracy. By making use of a vision system, two different methods for classifying the profiles’ defects were tested. Both proved to have excellent accuracy and repeatability. The study’s results can be applied to real manufacturing environments, as the models can run fast and accurate enough.
|
|---|
