An experiment in automation of statistical process control using fuzzy logic

This paper describes the development of a system to aid in maintaining a successful and reliable quality control program in a manufacturing environment. Such a system besides acquiring process related data will have the basic functions for the handling of massive statistical data which are necessary in order to monitor the performance of a production line. Starting with raw production data, the system outlined besides flagging for out-of-control case, has the capability to analyze unnatural statistical trends to determine the possible assignable cause(s) for the out-of-control situation. Based on knowledge acquired from a local expert, the system may suggest preventive or corrective actions. To achieve this, the technique used in the analysis of statistical trends or patterns is based on a fuzzy logic approach which has proven to be more suitable compared to conventional expert system.