Practical automation of statistical process control for injection moulding

A catastrophic process failure can be detected by local trips on the machine. The trip-limits should be set wide enough to allow for the cycle-to-cycle variation inherent in the process without tripping-out. However they cannot detect marginal process failure until it reaches catastrophic proportions. Statistical analysis of product and process is not only essential to detect marginal process failure, but can also go a long way to preventing a non-conforming product. The overall objective for automating statistical process control (SPC) is to identify and characterize quality problems earlier than a manual system. The author outlines the use of automated SPC in injection moulding and briefly describes the process monitoring techniques used and the role of expert systems.<
>