A systematic analysis of Temperature and ph in a continuous Stirred tank reactor using system Identification control and Optimization

Abstract

The research is centred on designing novel controller structures and exploiting novel algorithms to optimize the controller parameters. The new controller structures and optimization techniques were applied to a chemical reactor. Chemical engineering is a field that has a specific application to control systems. Hence, a Continuous Stirred Tank Reactor (CSTR) was considered for the study. The critical variables for the flawless operation of the reactor were the flow rate, temperature, and pH. The identified variable for temperature control was the coolant flow rate, and the manipulated variable for the pH control was the base flow rate. Once the input and output variables have been decided, the chemical reactor modeling is necessary. The models are structures that mimic the performance of the real-time system. Hence, modeling is a vital part of control system design. Data was acquired from the open loop system to identify the model of the reactor. The obtained data samples were used to identify models for usage in the closed loop system. Various models such as Nonlinear Auto Regressive eXogenous (NARX), Nonlinear Hammerstein Wiener (NLHW), and Neural Network-NARX (NN-NARX) were considered from which the best model that predicts the output similar to the measured data was retained for the implementation of the closed loop system. On comparing the models in terms of error indices and fit percentage, the NN-NARX model had the best performance. newline