Implementation of neural control for neutralization process

Dissertant

Abd al-Karim, Dirar Isam

Thesis advisor

Butrus, Karimah M.

University

University of Technology

Faculty

-

Department

Department of Chemical Engineering

University Country

Iraq

Degree

Master

Degree Date

2009

English Abstract

The dynamic behavior of neutralization process in (CSTR) was studied theoretical and experimental and the process control was implemented using different control strategies, conventional feedback control, and neural network (NARMA-L2, NN Predictive) control.

For acetic acid / caustic soda.

The system transient responses to one type of perturbation in feed flow rate: step.

Step testing approximation is used for predicting a transfer function of the neutralization process, a good agreement is obtained between the simulated and experimental responses of the dynamic model for the pH neutralization system in open loop.

The optimum tuning of Control Parameters were found by two different methods ; Frequency curve method (Bode diagram) and Process Reaction Curve using the mean of Square Error (MSE) method.

PI and PID controls were implemented as control strategies in this work.

Artificial neural network is the best method to control on neutralization process and it better than the conventional method because the mean square error is lower.

And we use two different control strategies in neural network (NARMA-L2, NN Predictive) control.

For all cases, the NARMA-L2 controller can be a preferable method for control purposes because it has smaller value of mean square error (MSE).

MATLAB program is used as a tool of solution for all cases used in this work.

Main Subjects

Chemistry

Topics

• Neural networks(Computer science)

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-305299