GUIDELINES FOR APPLYING STATISTICAL PROCESS CONTROL METHOD TO MONITOR THE TEMPERATURE OF CERAMIC FURNACS

The temperature control of a firing process is critical to the quality of any ceramic products. One of the temperature monitoring techniques is statistical process control or the utilization of control charts. However, the structure of observed data, autocorrelation, significantly affects the capability of the monitored processes. The selection of inappropriate types of control chart in this scenario might lead to an excessive number of false alarms or the loss of the ability to detect a special cause. As a result, a characterization of widely used control charts is performed in order to understand the characteristics of each control chart under autocorrelated situations. The results indicate that EWMA charts are more robust to false alarms than MR charts except when φ is between 0.6 and 1.0. Moreover, when there is an assignable cause, EWMA charts have also outperformed MR charts because of their sensitivity to a shift. To validate the empirical results, both charts are utilized to monitor a set of actual autocorrelated processes in the observed temperature of a ceramic furnace. In conclusion, the characterization has given practitioners opportunities to choose and utilize a suitable control chart to monitor the temperature of ceramic furnaces.