Research on control scheme of coordinate data based on multivariate SPC

Coordinate measuring instruments are used broadly in quality monitoring of automatic manufacturing, which usually generate three dimensional coordinate data (vector data). That traditional SPC focused on scalar data monitoring causes a great loss of data information and increases the second error probability of statistical process control easily. Thus it becomes more important to monitor coordinate data rather than scalar data in order to implement dimensional control. In this paper the coordinate monitoring method in manufacturing process is proposed. Multivariate statistical process control(MSPC) are used to build the monitoring scheme and the details of monitoring three basic geometrical elements, point, line and plane, are discussed. Performance of traditional SPC and the new proposed control scheme are compared. Research results show the second type of error probability of the presented control chart is smaller and it has more operating validity than the traditional control charts through a numerical example of vector dimension data.