Automation of Argon unit using Neural Network Predictive Control

This paper presents a method of realization of a real time non linear multi input and multi output systems using mathematical modelling and generalise over a wide range of similar systems for pre-process, controlling and to achieve optimum productivity as per the supply and demands in the industry. In the present work real time input / output data of a large chemical process of Distillation Column of Argon unit in Air Separation Unit (ASU) is taken as sample space and through this data the plant is modelled and simulated over a wide range of real time data. Further to verify the accuracy of the non linear system which is mathematically modelled using State Variable Approach.