Track: Quality Engineering, Control and Management

Abstract

Aluminum extrusion is considered as one of the very challenging processes to produce good quality and low cost products. This paper presents an optimization design of the X-bar&R chart for effective statistical monitoring of output of aluminum extrusion process. The proposed optimal X-bar&R chart is compared with traditional 3-σ X-bar&R chart, in terms of the average extra quadratic loss (AEQL), for detecting a wide range of shifts in the mean and variance. The results reveal that the optimal X-bar&R chart overtakes the 3-σ X-bar&R chart by 94%, in terms of AEQL. In addition, the former has the smallest out-of-control average time to signal (ATS) over almost all shifts. Consequently, the replacement of traditional X-bar&R chart by the optimal one is recommended to detect the mean and variance shifts more efficiently and as a result, reduce the process variation and avoid economic loss.