Computational Intelligence Approaches for Determination of Process Disturbance Patterns in an SPC-EPC System

In recent years, the integration of Statistical Process Control (SPC) and Engineering Process Control (EPC) has received considerable attentions due to the superiority for the process improvement. However, a drawback of the integration of SPC and EPC may be encountered for monitoring a process. Because EPC would compensate for the effects of underlying disturbances, the disturbance patterns could be embed and hard to be identified. However, the determination of Disturbance Patterns (DP) is crucial for process improvement due to the fact that DP would be associated with certain root causes for an unstable process. In this study, we propose three computational intelligence approaches to effectively determine the mixture patterns of process disturbances in a SPC-EPC system. Those three computational intelligence approaches include the artificial neural network, Time-Delay Neural Network (TDNN) and Rough Set (RS) techniques. Experimental results show that the proposed TDNN approach has the best performance of Accurate Identification rate (AIC) for determining the DP of an SPC/EPC system.