Tuning of a combustion controller by extremum seeking: a simulation study

Attenuation of thermoacoustic instabilities by active control has received considerable attention. Many experimentally successful endeavors have used phase-shifting controllers feeding back a phase-shifted/time-delayed version of the dominant harmonic of acoustic oscillations through appropriate actuation (loudspeakers or fuel flow modulation). The phase-shifters are tuned by trial and error to produce a minimum oscillation amplitude. Hence, for widespread practical application of these controllers, there exists a need for adaptive tuning of the phase-shift to produce a minimum oscillation amplitude in a dynamic environment. The present simulation study demonstrates the efficacy of the extremum seeking algorithm and a modified version in optimally tuning the phase-shift in a model of a controlled combustion process identified from data. Step and ramp changes in the optimal phase-shift are successfully tracked, and oscillation amplitude kept at its minimum value.