The simulation of pH control by artificial neural networks using MATLAB

• Main Authors: Lim, Zervin G., Espiritu, Mark Erwin M., Barcelona, Richard Jefferson B., Olaño, Servillano SB., Jr.
• Format: text
• Published: Animo Repository 2003
• Subjects: Hydrogen-ion concentration; Neutralization (Chemistry); Neural networks (Computer science); Chemical Engineering
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/7834
• Institution: De La Salle University

The control of a neutralization process between a strong acid and a strong base has been a challenging problem due to the asymptotic behavior of the titration curve near the pH 7. Many techniques have been studied bur acceptable method has yet to be developed. The use of Artificial Neural Networks (ANN) to control non-linear systems has been found to be promising. This study attempts to simulate the ability of ANN to control the pH of the neutralization process by using the toolbox included in MATLAB®. A SIMULINK® model of the control system, which is similar to the form of a simple feedback control system, was developed. A PIO controller is added to the control system to act as the controller for the entire process. The objective is for the ANN to generate the signal to regulate the flow of the base. The base reacts with the acid introduced at constant flow rate as to neutralize the acid to pH 7. The PID controller, which was tuned using the Ziegler-Nichols method, was used to generate the training data for the neural network. The effectiveness of the neural network was tested both by introducing step changes in the acid concentration and by changing the pH set point. From the results of the simulation, it was observed that the neural networks can effectively control the pH as it responds and adapts well to the changes introduced in the system.