Learning the relationship between operating condition and flame response from acoustic dataa).

Area of Science:

• Combustion dynamics
• Acoustic-thermoacoustic coupling
• Aerospace engineering

Background:

• Predicting thermoacoustic response in combustors is challenging due to flame response uncertainties.
• Accurate modeling is crucial for safe and efficient engine operation.

Purpose of the Study:

• To develop a robust method for predicting thermoacoustic response across operating conditions.
• To reduce experimental data requirements for flame transfer function identification.

Main Methods:

• Utilized Bayesian inference for flame parameter estimation.
• Employed Gaussian process regression to model flame parameter variations.
• Applied information-theoretic experiment design to select optimal forcing frequencies.

Main Results:

• Developed a Gaussian process model predicting flame response with quantified uncertainty.
• Created operating maps of the full thermoacoustic response.
• Identified three informative frequencies that reduce data needs by up to 90%.

Conclusions:

• The combined approach accurately predicts thermoacoustic response and significantly reduces experimental effort.
• This method offers a powerful tool for optimizing combustor design and operation.