Development of an Intelligent Fault Diagnosing System for Diagnosis of Multiple Concurrent Asynchronous Faults for Tennessee Eastman Process

Please enable javascript in your browser.

Akbari, Reza | 2018

474 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 50455 (06)
University: Sharif University of Technology
Department: Chemical and Petroleum Engineering
Advisor(s): Bozorgmehry Boozarjomehry, Ramin
Abstract:
Fault diagnosis in complex processes is very difficult due to their complexity and dimensionality, so it is of great importance. In previous works, concurrent fault diagnosing methods have not been able to diagnose faults with overlapping symptoms. On the other hand, there is no established and general method by which asynchronous concurrent faults with overlapping symptoms can be diagnosed. The purpose of this work is to come up with such a method in order to mitigate the shortcomings of the available methods which are not appropriate to be used for such a problem. In this problem, the main objective is to come up with course of events of faults according to their times occurrence. The proposed algorithm is based on process output signals, its performance has been assessed was applied to a case study in the TE process to evaluate the performance of the proposed algorithm. In this study, the algorithm was able to isolate faults and diagnose the sequence of occurrence and determine the time interval between two events with an acceptable accuracy
Keywords:
Frequency Domain Analysis ; Intelligent Algorithms ; Neural Network ; Fault Diagnosis ; Plantwide System ; Large-Sacle Process Systems

No TOC