Dissertation > Excellent graduate degree dissertation topics show

Study of Fault Diagnosis Methods for Chemical Process Startups

Author: WangZhenHeng
Tutor: ZhaoJinSong
School: Beijing University of Chemical Technology
Course: Control Theory and Control Engineering
Keywords: fault diagnosis SDG Dynamic locus analysis chemical process startup safety
CLC: TQ062
Type: Master's thesis
Year: 2009
Downloads: 146
Quote: 0
Read: Download Dissertation

Abstract


Chemical plants operate in a variety of states, some of these are steady states while others such as startup, shutdown are transitions. During the startup of a chemical process, the automatic control system such as DCS system is turned off and its success mainly depends on the manual operations of operators. Therefore, the startup process is prone to abnormalities and/or disasters. According to statistics of reported accidents, about 40% of those accidents occurred during process startups and shutdowns. Therefore, there is a need to develop an intelligent system that detects the faults occurred during the startups and enables the operator to correct errors.There are mainly two types of methods applied to the fault diagnosis of chemical process startups. One is model-based methods, and the other is data-based methods. Due to the strong nonlinearity, dynamics with long time delay and large operating range of process parameters, it is difficult to obtain mathematical mechanism models of the startups. Even if the mechanism models are available, it is still hard to satisfactorily solve them in real time.A history-data based fault diagnosis method called dynamic locus analysis (DLA) is studied in this paper. It is modified to improve the efficiency of fault diagnosis, which is validated through a case study on penicillin fermentation. To apply the modified version of DLA algorithm to large scale chemical process startup, the principal component analysis (PCA) method is integrated with it for dimension deduction. The hybrid PCA-DLA algorithm has successfully been applied to fault diagnosis of a lab-scale distillation column startup process. In order to monitor unknown faults, sign directed graph (SDG) model that qualitatively describes the causal relationships between process variables is combined with the above hybrid algorithm. A prototype expert system for fault diagnosis of chemical process startups has been developed based on the above methods on the expert system platform of GenSym G2 Optegrity.

Related Dissertations

  1. Design and Development of Application Programme for Fault Analyzer Based on Wince Platform,TP311.52
  2. Research on the 6-Dof Fault Tolerant Control of the Vibration Isolation Platform with Eight Actuators,TB535.1
  3. Development of Fault Diagnosis System for Turbine Shaft Based on Neural Network,TK267
  4. Research of SIS Based on SDG-HAZOP,N941.4
  5. Studies on Safety Assessment Modeling Using Hierarchical SDG-HAZOP for Thermal Power Plant,TM08;F224
  6. Online Decision Support System Research on Fault Diagnosis and Restoration Control,TM76
  7. Study on the Fault Diagnosis of Asynchronous Induction Motor Base on the Wavelet Analysis,TH165.3
  8. Research on Fault Diganosis for Time-Delay Systems,TH165.3
  9. Research on Fault Diagnosis of Satellite Attitude Control System,V448.22
  10. The Study on Embedded Condition Detection and Fault Diagnosis Technology for Chassis of Self-propelled Gun,TJ818
  11. Research on Supporting System for Fault Diagnosis and Maintenance of Missile,E927
  12. A certain type of active radar -guided missiles Universal Automatic Test Equipment Research,TJ760.6
  13. Clutch comprehensive performance test bed control and fault diagnosis system,U467.5
  14. Research on Drilling Complex Incidents and Accident Diagnosis Based on Expert System,TE24
  15. Research on the Management System of Oilfield Automation Technology for Power Supply,TE43
  16. The Application of Fault Diagnosis in Mine Elevator Based on Wavelet Analysis,TH165.3
  17. Research on Non-Invasive Real-Time Diagnosis Method for PLC Control System of Minning Hoist,TD534
  18. Vibration Fault Diagnosis and Anti-Vibration Research of Mechanical Press,TG305
  19. AWS fault diagnosis and quality control system design,P415.12
  20. A Mobile Network Fault Detection System Base on Data Mining,TN929.5
  21. The Research for Remote Fault Diagnosis Key Technology Based on Virtual Instrument,TP277

CLC: > Industrial Technology > Chemical Industry > General issues > Chemical production process, final product handling and packaging > Production methods and processes
© 2012 www.DissertationTopic.Net  Mobile