Technical Briefs

Hybrid Model-Based Fault Detection and Diagnosis for the Axial Flow Compressor of a Combined-Cycle Power Plant

[+] Author and Article Information
Jesús A. García-Matos

Research Assistant
e-mail: Jesus.garcia@iit.upcomillas.es

Miguel A. Sanz-Bobi

e-mail: Masanz@upcomillas.es

Antonio Muñoz

e-mail: Antonio.Munoz@upcomillas.es
Institute for Research in Technology (IIT),
ICAI School of Engineering,
Comillas Pontifical University,
Santa Cruz de Marcenado 26,
28015 Madrid, Spain

Antonio Sola

Direction of Technical Services,
Iberdrola Generación S.A.,
Tomás Redondo 1,
28033 Madrid, Spain
e-mail: asrasrasr@telefonica.net

Contributed by the International Gas Turbine Institute (IGTI) of ASME for publication in the JOURNAL OF ENGINEERING FOR GAS TURBINES AND POWER. Manuscript received September 20, 2012; final manuscript received October 22, 2012; published online April 23, 2013. Editor: Dilip R. Ballal.

J. Eng. Gas Turbines Power 135(5), 054501 (Apr 23, 2013) (5 pages) Paper No: GTP-12-1368; doi: 10.1115/1.4007902 History: Received September 20, 2012; Revised October 22, 2012

This technical brief is focused on the research area of fault detection and diagnosis in a complex thermodynamical system: in this case, an axial flow compressor. Its main contribution is a new approach which combines a physical model and a multilayer perceptron (MLP) model using the best advantages of both types of modeling. Fault detection is carried out by an MLP model whose residuals against the real outputs of the system determine which observations could be considered abnormal. A physical model is used to generate different fault simulations by shifting physical parameters related to faults. After these simulations are performed, the different fault profiles obtained are collected within a fault dictionary. In order to identify and diagnose a fault, the anomalous residuals observed by the MLP model are compared with the fault profiles in the dictionary and a correlation that provides a hypothesis with respect to the causes of the fault is obtained. This methodology has been applied to axial compressor operational data obtained from a real power plant. A case study based on the successful diagnosis of compressor fouling is included in order to show the potential of the proposed method.

Copyright © 2013 by ASME
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Fig. 1

Hybrid model-based fault detection and diagnosis approach: fault dictionary construction

Grahic Jump Location
Fig. 2

Fault profiles obtained for several increases in stage blockage (only samples which are known by the MLP)

Grahic Jump Location
Fig. 3

Hybrid model-based fault detection and diagnosis approach: fault detection and diagnosis implementation

Grahic Jump Location
Fig. 4

Fault detection in the case study compressor

Grahic Jump Location
Fig. 5

Fault profile compared with the real fault profile




Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In