Part geometry for advanced quality control and process monitoring

Quality control is very important for industrial manufacturing. Shewhart-type control charts are a popular method of ensuring the quality in manufacturing. However, in various control charts, the employed direct measurements provide insufficient information for advanced quality control tasks. In this research, we develop new variables by using a modified Fourier Descriptor method to extract the shape characteristics of manufactured parts. Compared with various direct measurements, these new variables provide much more information about the shape of parts for advised quality control purposes.; By using the new variables, we develop two procedures, a batch procedure and a continuous procedure. For products not in produced sequence, the batch procedure can detect not only the non-conforming parts which are not in produced sequence but also provides the error patterns which can facilitate the process correction by indicating the potential causes. The continuous procedure is developed to monitor manufacturing process. By using the new variables, the procedure can early detect problems occurring in a process by identifying the error patterns and providing the reset point.; Frequently recurring geometric forms, such as segments of line, segments of circular curve and complete circles, are implemented in this research. The experiments show the proposed new variables as well as both procedures can well fulfill the requirements of modern quality control.