Characterization of gas turbine burner instabilities by wavelet analysis of infrared images

• A 2D analysis of combustion analysis in an industrial gas turbine burner was performed.
• A new image process to follow the time evolution of instabilities was developed.

In this paper, an optical approach is tested to spatially characterize combustion fluctuations in a single burner atmospheric gas turbine test rig. The target of present investigation was to test the optical technique during combustion tests campaign, in order to couple it to an acoustic one in a future activity to be developed on the same experimental setup. The analysis based on fast infrared imaging of flames, coupled with photomultiplier and microphones measurements, has been elaborated on a 3 MW gas turbine test rig equipped with full scale burner tested in atmospheric conditions. The rig has been purposely designed to be tuned on acoustic frequencies detected in the real gas turbine machine equipped with 24 burners and operating at 20 bars. The tests evidenced main oscillations at low frequencies around 82 Hz and 146 Hz. These frequencies have been recorded in real machine too. The IR technique allowed to identify these frequencies in the 2D dimensions under humming conditions. The results obtained by IR presented a good agreement with microphones and optical measurements. Moreover, further investigation based on wavelet analysis came out as an interesting tool to develop a methodology for fingerprinting different burners operating in different conditions from thermoacoustic point of view.