Process Monitoring for Integration of SPC and APC Based on BP Neural Network

Aimed at an integration system of SPC (statistical process control) and APC (automatic process control), the effect of APC on SPC detection capability is analyzed. In this paper, the process disturbance is assumed to be an ARMA (1, 1) process. BP Neural Network with good capability of mode identification is used to substitute traditional SPC technology. Then an integrated design methodology has been developed for APC and BP neural network monitoring for the purpose of process special cause. A simulation result reveals that BP Neural Networks can detect special cause quickly and ameliorate SPC monitoring capability effectively. The ARL performance is studied. Lots of simulation experiments reveal that the proposed design approaches outperform the traditional integrated scheme of SPC and APC.