TY - JOUR
T1 - Detection and classification of lean blow-out and thermoacoustic instability in turbulent combustors
AU - Bhattacharya, Chandrachur
AU - De, Somnath
AU - Mukhopadhyay, Achintya
AU - Sen, Swarnendu
AU - Ray, Asok
N1 - Publisher Copyright:
© 2020 Elsevier Ltd
PY - 2020/11/5
Y1 - 2020/11/5
N2 - Lean blow-out (LBO) and thermoacoustic instability (TAI) are common undesirable occurrences in modern lean-burn turbulent combustion systems, such as fossil-fuel furnaces for gas turbines and in both land-based power generation and marine or aviation propulsion applications. While LBO causes loss of power due to flame extinction, TAI leads to loud noise, vibrations and mechanical fatigue-failure. Timely detection and classification of the operating conditions are important for both open-loop and closed-loop control of combustion systems to ensure their long service life, high efficiency, and reliability & availability. Data-driven techniques already exist for detection of these phenomena; however, most of these techniques require high training and/or processing times. This paper presents a fast Fourier transform (FFT)-based method to generate, in real time, a single scalar-valued measure for detection and classification of operational regimes; this measure can also be used to identify precursors (i.e., for prediction of impending LBO and TAI). This FFT-based method utilizes prior knowledge of the combustion system acoustics; and the measure acts as a classifier to distinguish different operational regimes. The underlying algorithms have been validated on time series data, collected from a (commercially available) microphone sensor that is external to the laboratory-scale experimental apparatus.
AB - Lean blow-out (LBO) and thermoacoustic instability (TAI) are common undesirable occurrences in modern lean-burn turbulent combustion systems, such as fossil-fuel furnaces for gas turbines and in both land-based power generation and marine or aviation propulsion applications. While LBO causes loss of power due to flame extinction, TAI leads to loud noise, vibrations and mechanical fatigue-failure. Timely detection and classification of the operating conditions are important for both open-loop and closed-loop control of combustion systems to ensure their long service life, high efficiency, and reliability & availability. Data-driven techniques already exist for detection of these phenomena; however, most of these techniques require high training and/or processing times. This paper presents a fast Fourier transform (FFT)-based method to generate, in real time, a single scalar-valued measure for detection and classification of operational regimes; this measure can also be used to identify precursors (i.e., for prediction of impending LBO and TAI). This FFT-based method utilizes prior knowledge of the combustion system acoustics; and the measure acts as a classifier to distinguish different operational regimes. The underlying algorithms have been validated on time series data, collected from a (commercially available) microphone sensor that is external to the laboratory-scale experimental apparatus.
UR - http://www.scopus.com/inward/record.url?scp=85089227316&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85089227316&partnerID=8YFLogxK
U2 - 10.1016/j.applthermaleng.2020.115808
DO - 10.1016/j.applthermaleng.2020.115808
M3 - Article
AN - SCOPUS:85089227316
SN - 1359-4311
VL - 180
JO - Applied Thermal Engineering
JF - Applied Thermal Engineering
M1 - 115808
ER -