Statistical Monitoring of Injection Moulds
Proceedings of the 2008 conference on Artificial Intelligence Research and Development: Proceedings of the 11th International Conference of the Catalan Association for Artificial Intelligence
| Hi-index | 0.00 |
Short-run statistical process control (SPC) usually transfers the data of different types of parts through data coding and part-family formation. As plastic injection moulds are mostly made in one-off or very small batches and one mould part can be very different from another, constant machine set-up changes are needed. In this situation, part-family formation becomes very difficult. The part family either becomes too large, introducing variations that are unnecessary or more complex to handle, or there are too many part families to handle. An approach is proposed in the present study with the aim to overcome the aforementioned problems. This approach defines the SPC processes and part families based on data from manufacturing processes in the mould manufacturing workshop. For each mould part to be produced, its manufacturing processes are classified into particular SPC processes, and the part is classified into a specific part family according to prescribed memberships. Through this procedure, unwanted variations are effectively isolated to facilitate and simplify subsequent data processing and interpretation. Domain-specific methods and rules of SPC planning for mould manufacturing are also presented to simplify the application of SPC in this area.