[MODELING OF PH NEUTRALIZATION PROCESS PILOT PLANT]

HASSAN, FARIDAH (2004) [MODELING OF PH NEUTRALIZATION PROCESS PILOT PLANT]. [Final Year Project] (Unpublished)

[thumbnail of 2004 Bachelor - MOdeling Of pH Neutralization Process Pilot Plant.pdf] PDF
2004 Bachelor - MOdeling Of pH Neutralization Process Pilot Plant.pdf

Download (1MB)

Abstract

System Identification is an art of constructing a mathematical model for a dynamic
response system. The modeling process is based on the observed input and output
data for a system. To start a modeling process, a good understanding of process
behavior is required as it will determine the important parameters and characteristics
to be analyzed.
pH neutralization is a very nonlinear process. It is not easy to get an accurate model
as compared to the actual model. Modeling using conventional methods does not
seem to give a reliable model for this process. Thus, for wide a range of
neutralization pH values, conventional modeling methods are not sufficient.
Therefore, for this project, intelligent approaches are considered.
The conventional methods that are used by the Author are mathematical modeling,
empirical modeling and statistical modeling. Mathematical modeling is done to see
the relation of inputs and output. Empirical modeling is the common method used
for plant modeling. Statistical modeling is more a to computerized modeling where it
requires a good computer configuration basic in order to achieve the desired output.
Neural Network is used for the intelligent method. Neural network is an intelligent
approach that has the capability to predict future plant performance by training
several datasets.
These conventional and intelligent methods are compared between each other in
term of the model accuracy, model reliability and flexibility. Modeling using
mathematical modeling is tedious and requires more effort on the block diagram
configuration in order to get an accurate result. Empirical modeling is basically good
enough for plant identification, unfortunately for a highly nonlinear system, the
method does not seem reliable. Statistical modeling has the ability to predict an
acceptable higher order model. On top of that, neural network could give a more
reliable and accurate result.

Item Type: Final Year Project
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering
Departments / MOR / COE: Engineering > Electrical and Electronic
Depositing User: Users 2053 not found.
Date Deposited: 30 Sep 2013 16:55
Last Modified: 25 Jan 2017 09:46
URI: http://utpedia.utp.edu.my/id/eprint/7937

Actions (login required)

View Item
View Item