Neural Modelling Approach to Integrated Process Control

Abstract

Control system engineers have contributed to the development of PID (proportional, integral, differential) controllers for use in the process control industry. These developments are based on control theory principles such as poles-zeros cancellation, root-locus method and disturbance (white noise) rejection, etc. There is also wide literature using artificial neural networks (ANN) applications to integrate statistical process control (SPC) and automatic/engineering process control (APC/EPC). These literature publications are heavily mathematics based of higher order and may not be within the grasp and comprehension of most of the researchers with mathematical knowledge of college or university graduate standards. In order to overcome these difficulties in wide knowledge gap, this technical note discusses and suggests an ANN approach into the design of quality regulator that uses integral regulator algorithm, (given below developed from first principles), the mathematical and stochastic modelling and feedback control algorithm are developed and based on research publications of Box and Jenkins, technical literature available in Technometrics journals, IEEE Transactions, Journals of Royal Statistical Society and Quality Technology, etc. It is expected that the quality integral regulator developed using neuron models and ANN principles will also optimize variance of product quality variable through predictive and adaptive process control regulator algorithm. The quality integral regulator will help minimize output quality variations due to process operating conditions. The regulator algorithm developed on the basis of ANN theory and principles will answer challenging demands of modern computer remotely controlled complex dynamic processes. It may not be possible to understand well the technical capacity of the existing PID and industrial controllers to extend their process control capabilities up to certain level and extent only. The quality integral regulator that can be designed and developed on the basis of ANN models will provide and meet the needs of the modern, process control industry for efficient quality control.

Keywords: SPC, APC, ANN, PID controller, stochastic, integral, regulator, algorithm

Cite this Article: G. Venkatesan. Neural Modelling Approach to Integrated Process Control. International Journal of Algorithms Design and Analysis. 2019; 5(2): 46–51p.