Identification and modeling of acoustic bursts emitted from a supersonic turbulent jet

The acoustic radiation of turbulent jets occurs in form of intermittent bursts. This transient behavior complicates jet noise modeling endeavors, and explains the failure of linear jet noise models that do not take intermittency into account. In this study, we apply a conditional form of proper orthogonal decomposition (POD) to distill the waveform of the statistically most relevant, i.e. energetic, acoustic burst from a Large Eddy Simulation of a turbulent, hot, Mach 1.5 jet. The identified statistical burst event takes the form of spatially confined wavepacket that is emitted in the direction of the peak aft-angle noise. We identify the precursor of this loud event by defining a conditional space-time POD problem, which allows us to trace the time evolution of the statistical bust event—both forwards, and backwards in time. The precursor takes the form a compact wavepacket in the initial shear-layer of the jet. We demonstrate that this burst event and its temporal evolution can be modeled by means of mean-flow-based optimal linear growth theory. The identification and characterization of the precursor event is a first step towards model predictive control of loud events.