Download Prospects for Aero Gas-turbine Diagnostics a Review PDF

TitleProspects for Aero Gas-turbine Diagnostics a Review
File Size381.6 KB
Total Pages18
Table of Contents
                            Prospects for aero gas-turbine diagnostics: a review
	Introduction
	New role for the gas-turbine after-sales market
	Gas-path diagnostic methodologies
		Linear gas-path analysis with ICM inversion
		Non-linear gas-path analysis with ICM inversion
		Kalman-filter and weighted least-squares based GPA: linear approach
		Kalman-filter based GPA: non-linear approach
		Non-linear model-based optimal estimation by using genetic-algorithms
		Artificial neural-network based GPA
		Bayesian-belief network based GPA
		Expert systems
		Fuzzy-logic based diagnostics
	Comparison of the methodologies
	Conclusions
	References
                        
Document Text Contents
Page 1

APPLIED
www.elsevier.com/locate/apenergy

Applied Energy 79 (2004) 109–126
ENERGY
Prospects for aero gas-turbine diagnostics:
a review

Luca Marinai *, Douglas Probert, Riti Singh

Department of Power, Propulsion and Aerospace Engineering, Cranfield University,

Bedford MK43 0AL, UK

Accepted 21 October 2003

Available online 27 February 2004
Abstract

Despite inflating unit-fuel costs, the long-term prospects for the aircraft industry remain

buoyant. Nevertheless reducing direct operating-costs is crucial to ensure competitive ad-

vantages for airlines and manufacturers, and so effective advanced engine-condition moni-

toring methodologies are desirable. Hence gas-path diagnostic methods are reviewed and the

specifications for such effective tools deduced, together with pertinent future prospects.

� 2004 Elsevier Ltd. All rights reserved.

Keywords: Performance; Diagnostics; Gas-path analysis; Aftermarket

1. Introduction

At present, there are deep financial uncertainties in the civil aircraft market and

therefore intense competition among airlines. Hence the development of advanced

maintenance-techniques in order to reduce operating-costs [1,2]. Engine-related costs

contribute a large fraction of the direct operating-costs (DOCs) of an aircraft, be-

cause the propulsion system requires a significant part of the overall maintenance

effort that has to be expended for each aircraft – see Fig. 1.
The world market for transportation by air is expanding, despite the difficulties

and changes following the horrific terrorist attack on September 11th 2001 in New
*
Corresponding author. Tel.: +44-1234-750-111-526; fax: +44-1234-752-407.

E-mail address: [email protected] (L. Marinai).

0306-2619/$ - see front matter � 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.apenergy.2003.10.005

mail to: [email protected]

Page 9

Fig. 3. The diagnostics strategy uses an engine-performance model that is based on characteristics of the

gas-turbine components [18].

L. Marinai et al. / Applied Energy 79 (2004) 109–126 117
condition of the engine-components in the presence of measurement noise and bi-

ases. The measurement uncertainty is supposed to affect even the parameter setting

and the operating condition. Estimation is performed through an �ad hoc� GA. The
GA uses an accurate non-linear steady-state model of the engine�s behaviour. The
only statistical assumption required by the technique concerns the measurement

noise and the maximum allowed number of faulty sensors and engine-components.
The method suffers from the following limitations [18]:

• The methodology is more computationally burdensome than classic estimation-
techniques.

• Although multiple-faults can be detected, the technique is limited to four param-
eters experiencing simultaneous deteriorations.

• Care must be taken when using the GA in assigning the number of strings. Even
though the rule for the assignment of the number of strings for different fault clas-

ses can easily be established by trial-and-error and the achieved accuracy is not a
strong function of the rule of assignment itself, active awareness of these issues is

necessary for the correct utilization of the technique. This makes the method dif-

ficult to use, and requires a trained person for its worthwhile operation.

Some of these limitations have been overtaken by further developments of the

technique, as discussed later. In fact, the theory here described has been implemented

Page 10

118 L. Marinai et al. / Applied Energy 79 (2004) 109–126
for development engines where many measurements are available. It was applied to a

three-spool military turbofan engine RB199 and a two-spool low by-pass military

turbofan engine EJ200 [19,20]: it provided a high level of accuracy. Gulati et al.

[21,22] combined a multiple-point diagnostic approach [23] and a GA approach to

produce a GA-based multiple operating-point analysis (MOPA) method for gas-

turbine fault diagnostics. This approach is suitable for problems where only limited
instrumentation is available. It was applied to a RB199 engine and showed good

results. Further developments of the method have led to the study discussed by

Sampath [24] in which an hybrid approach has been adopted for the intercooled

recuperated WR21 engine, thereby gaining by reducing the computational burden.

The application of evolution strategy also has been investigated by Sampath et al.

[25]. Sampath [24] described a diagnostics model that operates in three distinct

stages. The first stage uses response surfaces for computing objective-functions to

increase the exploration potential of the search space, while reducing the computa-
tional load. The second stage uses the heuristics modification of genetics algorithm

parameters through a master-slave type configuration. The third stage uses the elitist

model concept applied to a GA to preserve the accuracy of the solution in the face of

randomness. This fault-diagnostics model has been integrated with a nested neural-

network to form a hybrid diagnostics model. The nested neural-network is employed

as a pre-processor or filter to reduce the number of fault classes to be explored by the

GA-based diagnostics model. The hybrid model improves the accuracy, reliability

and consistency of the results obtained.

3.6. Artificial neural-network based GPA

Artificial neural-networks (ANNs) have been investigated extensively for use in

fault diagnoses. An ANN consists of parallel distributed processors able to store

knowledge as experience and make it available for use. ANNs are trained to map

inputs to outputs via a non-linear relationship in a framework that loosely mimics

the learning process performed by the brain. Generally, the NN operates in two
phases – a learning phase and an operating phase. The purpose of the learning

phase is to determine the NN parameters, which will enable the network to

function properly in the operating phase. The multi-layer perceptron (MLP) with

back-propagation training is the most common architecture used for GPA pur-

poses. It is also called the feed-forward back-propagation neural-network

(FFBPNN). The use of ANNs in gas-path diagnostics experiences the following

limitations:

• Like other AI tools, ANNs are unable to perform creditably outside the range of
data to which they have been exposed: this implies that a massive amount of data

from encountered and foreseeable fault conditions of operation would be required

in each ANN development.

• Training times are long, though this depends on the network type, size and the
amount of training data. ANNs require retraining when machine operating

conditions change. This could mean, for instance, retraining after a machine

overhaul.

Page 17

L. Marinai et al. / Applied Energy 79 (2004) 109–126 125
[19] Zedda M, Singh R. Gas-turbine engine and sensor fault diagnosis. ISABE 99-7238, 13th ISABE,

Florence, Italy, 1999.

[20] Zedda M, Singh R. Gas-turbine engine and sensor fault diagnosis using optimization techniques. 35th

AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Los Angeles, CA, USA, 24 June

1999, AIAA 99-253020.

[21] Gulati A, Zedda M, Singh R. Gas-turbine engine and sensor multiple operating-point analysis using

optimisation techniques. 36th AIAA/ASME/SAE/ASEE Joint propulsion conference and exhibi-

tion,17-Huntsville, Alabama, USA, AIAA 2000-3716, 19 July 2000.

[22] Gulati A, Taylor D, Singh R. Multiple operating-point analysis using Genetic-Algorithm optimisa-

tion for gas-turbine diagnostics. ISABE-2001-1139, 2001.

[23] Stamatis A, Mathioudakis K, Berios G, Papailiou K. Jet engine-fault detection with discrete

operating points gas-path analysis. J Propulsion 1991;7(6):1043–8, or ISABE 89-7133.

[24] Sampath S. Fault diagnostics for an advanced cycle marine gas-turbine using Genetic-Algorithms,

PhD Thesis, School of Engineering, Cranfield University, 2003.

[25] Sampath S, Marinai L, Ogaji SOT, Singh R. Evolution strategy applied to aero-engine-fault

diagnosis. In: Seventh International Conference on Knowledge-Based Intelligent Information &

Engineering Systems – KES, 3–5 September 2003, Oxford, United Kingdom, 2003.

[26] Zedda M, Singh R. Fault diagnosis of a turbofan engine using neural-networks: a quantitative

approach. In: 34th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, AIAA 98-3602,

Cleveland, OH, USA, 13–15 July, 1998.

[27] Ogaji S. Advanced gas-path fault diagnostics for stationary gas-turbines, PhD Thesis, School of

Engineering, Cranfield University, 2003.

[28] Torella G, Lombardo G. Utilization of neural-networks for gas-turbine engines. XII ISABE 95-703,

1995.

[29] Kanelopoulos K, Stamatis A, Mathioudakis K. Incorporating neural-networks into gas-turbine

performance diagnostics. ASME 97-GT-35, International Gas-turbine & Aero-engine Congress &

Exhibition, Orlando, Florida, USA, June 8–13, 1997.

[30] Volponi A, DePold H, Ganguli R, Daguang C. The use of Kalman-filter and neural-network

methodologies in gas-turbine performance diagnostics: a comparative study. ASME 2000-GT-547,

2000.

[31] Green A. Artificial-intelligence for real-time diagnostics and prognostics of gas-turbine engines.

AIAA 97-2899 Conference Seattle, WA, USA, 6–9 July 1997.

[32] Eustace R, Merrington G. Fault diagnosis of fleet engines using neural-networks. XII ISABE 1993;95-

7085.

[33] Kadamb A. Bayesian belief network for aero gas-turbine module and system-fault isolation. MSc

Thesis, School of Engineering, Cranfield University; 2003.

[34] Romessis C, Stamatis A, Mathioudakis K. Setting-up a belief network for turbofan diagnosis with the

aid of an engine-performance model. ISABE-2001-1032, 2001.

[35] Mathioudakis K, Romessis C, Stamatis A. Probabilistic methods for gas-turbine fault diagnostics.

Von Karman Institute Lecture Series 2003-01, Gas-turbine Condition Monitoring & Fault Diagnosis,

Brussels, Belgium, 13–17 January 2003.

[36] Ali M, Gupta U. An Expert System for Fault Diagnosis in a Space-Shuttle Main Engine 26th AIAA/

SAE/ASME/ASEE Joint Propulsion Conference, Orlando, FL, USA, July 16–18, 1990, AIAA-90-

1890.

[37] DePold R, Gass D. Application of expert systems and neural-networks to gas-turbine prognostics and

diagnostics. J Eng Gas-Turbines Power 1999:121–612.

[38] Doel DL. The role for expert systems in commercial gas-turbine engine monitoring. The Gas-turbine

and Aero-engine Congress and Exposition, Brussels, Belgium, June 1990.

[39] Spina PR, Torella G, Venturini M. The use of expert systems for gas-turbine diagnostics and

maintenance. In: Proceedings of the ASME Turbo Expo, Amsterdam, The Netherlands, June 3–6,

2002.

[40] Torella G. Expert Systems for the Simulation of Turbofan Engines. ISABE 93-7133, Tokyo, Japan,

September 20–24, 1993.

Page 18

126 L. Marinai et al. / Applied Energy 79 (2004) 109–126
[41] Torella G. Expert systems and neural-networks for fault isolation in gas-turbines. ISABE 97-7148,

1997.

[42] Torella G, Torella R. Probabilistic expert systems for the diagnostics and trouble-shooting of gas-

turbine apparatuses. In: AIAA 99-2942, 35th AIAA/ASME/SAE/ASEE Joint Propulsion Conference,

LA, California, USA, June 20–24, 1999.

[43] Zadeh LA. Toward a theory of fuzzy systems. Washington, DC, USA: NASA CR-1432; 1969.

[44] Healy A, Kerr L, Larkin L. Model-based fuzzy-logic sensor fault accommodation. ASME 97-GT-222,

1997.

[45] Siu C, Shen Q, Milne R. TMDOCTOR: a fuzzy rule- and case-based expert system for

turbomachinery diagnosis. In: SAFEPROCESS�97, vol. 1, 1997. p. 556–63.
[46] Tang G. A practical intelligent system for condition monitoring and fault diagnosis of jet engines.

AIAA 99-2533, 1999.

[47] Ganguli R. Application of fuzzy-logic for fault isolation of jet engines. ASME 2001-GT-0013, 2001.

[48] Ganguli R. A fuzzy-logic intelligent system for gas-turbine module and system fault isolation. ISABE-

2001-1112, 2001.

[49] Ganguli R. Health monitoring of a helicopter rotor in forward flight using fuzzy-logic. AIAA J

2002;40(12):2373–81.

[50] Karvounis G, Frith P. Automated detection of engine heath using hybrid model-based and fuzzy-logic

approach. In: ISABE-2003-1232, XVI International Symposium on air-breathing engines, Cleveland,

OH, 31 August–5 September 2003.

[51] Ogaji SOT, Marinai L, Sampath S, Singh R. Comparison of artificial neural-network and fuzzy-logic

approaches to engine diagnostics. Int J Eng Intelligence Syst, submitted for publication.

[52] Volponi A. Extending gas-path analysis coverage for other fault conditions. Von Karman Institute

Lecture Series 2003-01, Gas-turbine Condition Monitoring & Fault Diagnosis, Brussels, Belgium, 13–

17 January, 2003.

[53] Kosko B. Fuzzy engineering. New Jersey: Prentice Hall; 1997.
Glossary

Bias: Measurement systematic error

Computational burden: Negative property of excessively time consuming software

Health parameters: Engine-components� efficiency and flow capacity as indicative of their health condition
Rules explosion: Number of rules in a fuzzy system increases according to the complexity of the process

that is being approximated

Smearing: Tendency to smear the faults over a large number of the engine�s components and sensors
String: A coded potential solution of the GA

String assignment: The process of assigning a string in a GA setting-up procedure

Similer Documents