Hybrid Model-Driven Spectroscopic Network for Rapid Retrieval of Turbine Exhaust Temperature

Yalei Fu; Rui Zhang; Jiangnan Xia; Andrew Gough; Stuart Clark; Abhishek Upadhyay; Godwin Enemali; Ian Armstrong; Ihab Ahmed; Mohamed Pourkashanian; Paul Wright; Krikor Ozanyan; Michael Lengden; Walter Johnstone; Nick Polydorides; Hugh McCann; Chang Liu

doi:10.1109/TIM.2023.3328086

Hybrid Model-Driven Spectroscopic Network for Rapid Retrieval of Turbine Exhaust Temperature

Yalei Fu, Rui Zhang, Jiangnan Xia, Andrew Gough, Stuart Clark, Abhishek Upadhyay, Godwin Enemali, Ian Armstrong, Ihab Ahmed, Mohamed Pourkashanian, Paul Wright, Krikor Ozanyan, Michael Lengden, Walter Johnstone, Nick Polydorides, Hugh McCann, Chang Liu

School of Engineering

Research output: Contribution to journal › Article › peer-review

Abstract / Description of output

Exhaust gas temperature (EGT) is a key parameter in diagnosing the health of gas turbine engines (GTEs). In this article, we propose a model-driven spectroscopic network with strong generalizability to monitor the EGT rapidly and accurately. The proposed network relies on data obtained from a well-proven temperature measurement technique, i.e., wavelength modulation spectroscopy (WMS), with the novelty of introducing an underlying physical absorption model and building a hybrid dataset from simulation and experiment. This hybrid model-driven (HMD) network enables strong noise resistance of the neural network against real-world experimental data. The proposed network is assessed by in situ measurements of EGT on an aero-GTE at millisecond-level temporal response. Experimental results indicate that the proposed network substantially outperforms previous neural-network methods in terms of accuracy and precision of the measured EGT when the GTE is steadily loaded.

Original language	English
Article number	2531710
Pages (from-to)	1-10
Number of pages	1
Journal	IEEE Transactions on Instrumentation and Measurement
Volume	72
Early online date	27 Oct 2023
DOIs	https://doi.org/10.1109/TIM.2023.3328086
Publication status	E-pub ahead of print - 27 Oct 2023

Keywords / Materials (for Non-textual outputs)

Temperature measurement
Measurement by laser beam
Absorption
Gas lasers
Fitting
Data models
Spectroscopy
signal processing
wavelength modulation spectroscopy (WMS)
Deep neural network (DNN)
gas turbine engine (GTE)
exhaust gas temperature (EGT)

Access to Document

10.1109/TIM.2023.3328086

AFPRM_Hybrid Model-Driven Spectroscopic Network for Rapid Retrieval of Turbine Exhaust TemperatureAccepted author manuscript, 1.63 MBLicence: Creative Commons: Attribution (CC-BY)

https://ieeexplore.ieee.org/document/10298272/

Laser Imaging of Turbine Engine Combustion Species (LITECS)
Mccann, H., Liu, C. & Polydorides, N.
EPSRC
1/09/20 → 31/08/24
Project: Research
In-situ Chemical Measurement and Imaging Diagnostics for Energy Process Engineering
Mccann, H., Jia, J., Linne, M., Peterson, B. & Polydorides, N.
EPSRC
1/10/16 → 30/09/21
Project: Research

Cite this

Fu, Y., Zhang, R., Xia, J., Gough, A., Clark, S., Upadhyay, A., Enemali, G., Armstrong, I., Ahmed, I., Pourkashanian, M., Wright, P., Ozanyan, K., Lengden, M., Johnstone, W., Polydorides, N., McCann, H., & Liu, C. (2023). Hybrid Model-Driven Spectroscopic Network for Rapid Retrieval of Turbine Exhaust Temperature. IEEE Transactions on Instrumentation and Measurement, 72, 1-10. Article 2531710. Advance online publication. https://doi.org/10.1109/TIM.2023.3328086

@article{fc252518f2b44ba7bcfa34a552a9173c,

title = "Hybrid Model-Driven Spectroscopic Network for Rapid Retrieval of Turbine Exhaust Temperature",

abstract = "Exhaust gas temperature (EGT) is a key parameter in diagnosing the health of gas turbine engines (GTEs). In this article, we propose a model-driven spectroscopic network with strong generalizability to monitor the EGT rapidly and accurately. The proposed network relies on data obtained from a well-proven temperature measurement technique, i.e., wavelength modulation spectroscopy (WMS), with the novelty of introducing an underlying physical absorption model and building a hybrid dataset from simulation and experiment. This hybrid model-driven (HMD) network enables strong noise resistance of the neural network against real-world experimental data. The proposed network is assessed by in situ measurements of EGT on an aero-GTE at millisecond-level temporal response. Experimental results indicate that the proposed network substantially outperforms previous neural-network methods in terms of accuracy and precision of the measured EGT when the GTE is steadily loaded.",

keywords = "Temperature measurement, Measurement by laser beam, Absorption, Gas lasers, Fitting, Data models, Spectroscopy, signal processing, wavelength modulation spectroscopy (WMS), Deep neural network (DNN), gas turbine engine (GTE), exhaust gas temperature (EGT)",

author = "Yalei Fu and Rui Zhang and Jiangnan Xia and Andrew Gough and Stuart Clark and Abhishek Upadhyay and Godwin Enemali and Ian Armstrong and Ihab Ahmed and Mohamed Pourkashanian and Paul Wright and Krikor Ozanyan and Michael Lengden and Walter Johnstone and Nick Polydorides and Hugh McCann and Chang Liu",

note = "Funding Information: This work was supported by the U.K. Engineering and Physical Sciences Research Council (EPSRC) Programme Grant LITECS under Grant EP/T012595/1. Publisher Copyright: {\textcopyright} 1963-2012 IEEE.",

year = "2023",

month = oct,

day = "27",

doi = "10.1109/TIM.2023.3328086",

language = "English",

volume = "72",

pages = "1--10",

journal = "IEEE Transactions on Instrumentation and Measurement",

issn = "1557-9662",

publisher = "Institute of Electrical and Electronics Engineers",

}

Fu, Y, Zhang, R, Xia, J, Gough, A, Clark, S, Upadhyay, A, Enemali, G, Armstrong, I, Ahmed, I, Pourkashanian, M, Wright, P, Ozanyan, K, Lengden, M, Johnstone, W, Polydorides, N , McCann, H & Liu, C 2023, 'Hybrid Model-Driven Spectroscopic Network for Rapid Retrieval of Turbine Exhaust Temperature', IEEE Transactions on Instrumentation and Measurement, vol. 72, 2531710, pp. 1-10. https://doi.org/10.1109/TIM.2023.3328086

TY - JOUR

T1 - Hybrid Model-Driven Spectroscopic Network for Rapid Retrieval of Turbine Exhaust Temperature

AU - Fu, Yalei

AU - Zhang, Rui

AU - Xia, Jiangnan

AU - Gough, Andrew

AU - Clark, Stuart

AU - Upadhyay, Abhishek

AU - Enemali, Godwin

AU - Armstrong, Ian

AU - Ahmed, Ihab

AU - Pourkashanian, Mohamed

AU - Wright, Paul

AU - Ozanyan, Krikor

AU - Lengden, Michael

AU - Johnstone, Walter

AU - Polydorides, Nick

AU - McCann, Hugh

AU - Liu, Chang

N1 - Funding Information: This work was supported by the U.K. Engineering and Physical Sciences Research Council (EPSRC) Programme Grant LITECS under Grant EP/T012595/1. Publisher Copyright: © 1963-2012 IEEE.

PY - 2023/10/27

Y1 - 2023/10/27

N2 - Exhaust gas temperature (EGT) is a key parameter in diagnosing the health of gas turbine engines (GTEs). In this article, we propose a model-driven spectroscopic network with strong generalizability to monitor the EGT rapidly and accurately. The proposed network relies on data obtained from a well-proven temperature measurement technique, i.e., wavelength modulation spectroscopy (WMS), with the novelty of introducing an underlying physical absorption model and building a hybrid dataset from simulation and experiment. This hybrid model-driven (HMD) network enables strong noise resistance of the neural network against real-world experimental data. The proposed network is assessed by in situ measurements of EGT on an aero-GTE at millisecond-level temporal response. Experimental results indicate that the proposed network substantially outperforms previous neural-network methods in terms of accuracy and precision of the measured EGT when the GTE is steadily loaded.

AB - Exhaust gas temperature (EGT) is a key parameter in diagnosing the health of gas turbine engines (GTEs). In this article, we propose a model-driven spectroscopic network with strong generalizability to monitor the EGT rapidly and accurately. The proposed network relies on data obtained from a well-proven temperature measurement technique, i.e., wavelength modulation spectroscopy (WMS), with the novelty of introducing an underlying physical absorption model and building a hybrid dataset from simulation and experiment. This hybrid model-driven (HMD) network enables strong noise resistance of the neural network against real-world experimental data. The proposed network is assessed by in situ measurements of EGT on an aero-GTE at millisecond-level temporal response. Experimental results indicate that the proposed network substantially outperforms previous neural-network methods in terms of accuracy and precision of the measured EGT when the GTE is steadily loaded.

KW - Temperature measurement

KW - Measurement by laser beam

KW - Absorption

KW - Gas lasers

KW - Fitting

KW - Data models

KW - Spectroscopy

KW - signal processing

KW - wavelength modulation spectroscopy (WMS)

KW - Deep neural network (DNN)

KW - gas turbine engine (GTE)

KW - exhaust gas temperature (EGT)

U2 - 10.1109/TIM.2023.3328086

DO - 10.1109/TIM.2023.3328086

M3 - Article

SN - 1557-9662

VL - 72

SP - 1

EP - 10

JO - IEEE Transactions on Instrumentation and Measurement

JF - IEEE Transactions on Instrumentation and Measurement

M1 - 2531710

ER -

Hybrid Model-Driven Spectroscopic Network for Rapid Retrieval of Turbine Exhaust Temperature

Abstract / Description of output

Keywords / Materials (for Non-textual outputs)

Access to Document

Fingerprint

Projects

Laser Imaging of Turbine Engine Combustion Species (LITECS)

In-situ Chemical Measurement and Imaging Diagnostics for Energy Process Engineering

Cite this