A case study on a real-time quality control system using data mining

The purpose of quality management is to produce a good quality product that satisfies the consumer. Computerized systems are more common in the production line, and playing an increasing role in quality management due to the development of IT technologies. Numerous quality management methodologies have been provided to integrate such systems. Quality control employing data mining has been recently introduced but has not yet been extensively applied in practice. A computerized quality control system can acquire the manufacturing process data in real time and report immediately the occurrence of defects. The data mining model based on values of production process features that affect the product quality predicts when and why defects occur in this system context. We collected the production process feature data in the painting processes of an automobile bumper production plant. We presented itemized decision-tree based data mining model by the type of defects. In addition, a new quality management methodology is presented through the real time quality control system development to predict defects and raise an alarm signal based on data mining model patterns on the occurrence of defects from real time data collection in manufacturing. The results of this study are expected to be a good alternative to prevent defects by stabilizing the production process quality beyond the traditional quality control methods.