Multiple Wavepacket Decomposition of F404 Engine Military Jet Noise

Noise sources in heated, supersonic jets are challenging to measure directly due to the extreme environment. Inverse methods, such as acoustic beamforming, using data collected near these sources can be used to construct equivalent source models that accurately describe the acoustic radiation. However, generating these models for a complex source can require thousands of parameters for a complete description. To address this challenge, a multiple wavepacket decomposition is introduced, reducing the source into a set of analytic wavepackets that recreate the full-rank model. This paper provides an overview of the decomposition procedure and applies it to acoustic data collected near a T-7A-installed GE F404 engine. The decomposition is validated for a peak frequency at military power and physical implications are discussed. It is shown that the downstream radiation at MIL can be accurately reconstructed with as few as five wavepackets with minimal error. The number of wavepackets required to capture the primary radiation region tends to increase with frequency and engine power. Beyond the dominant frequencies, the number of wavepackets per wavelength increases drastically, indicating a rapid decrease in source coherence. Finally, a single wavepacket model is fit to the data at MIL, which captures primary radiation features.