Process fault detection and diagnosis using a dynamic neural networks model

Process fault detection and diagnosis using a dynamic neural networks model

Recently, neural networks has generated considerable interest as an alternative non-linear modelling tool. The major attraction is the learning capabilities of neural networks, and the fact that multi-layer, feed forward networks can approximate any non-linear function with arbitrary accuracy. This...

Full description

Bibliographic Details
Main Authors: Abdul Rahman, Ribhan Zafira, Che Soh, Azura, Ahmad, Erny Arniza, Mohd Noor, Samsul Bahari, Gomm, J. B.
Format: Conference or Workshop Item
Language:English
Published: Universiti Putra Malaysia Press 2002
Online Access:http://psasir.upm.edu.my/id/eprint/33839/1/33839.pdf
Description
Summary:Recently, neural networks has generated considerable interest as an alternative non-linear modelling tool. The major attraction is the learning capabilities of neural networks, and the fact that multi-layer, feed forward networks can approximate any non-linear function with arbitrary accuracy. This study describes the application of the multi-layer perceptron (MLP) neural network, trained using back-error propagation, to obtain a representative model of a non-linear process over a wide operational region. The purpose of this study is mainly to investigate the use of dynamic neural networks model for fault detection and diagnosis of the process control. The MATLAB with SIMULINK process and Multi-Layer Perceptron Software Package is used as a method to procure the required result.

Similar Items