Aeroacoustic optimization of unsteady transonic cavity flow using a Bayesian approach

Abstract The embedded weapon bay in military aircraft forms a cavity when opened to release stores. The complex unsteady flow dynamics and high levels of acoustic noise associated with these storage bays present significant design challenges for fighter aircraft. The primary objective of this study is to optimize the geometry of the cavity based on aerodynamic and aeroacoustic criteria using a constrained Bayesian approach. To represent the store bay, we use a generic M219 cavity geometry with a length-to-depth ratio of 5, along with experimental data from a wind tunnel test, to validate the methodology and model. During the optimization process, a computationally efficient aeroacoustic solution is generated using the Spalart–Allmaras turbulence model on a coarse mesh. In other words, solutions for each geometry can be obtained with approximately 80 $$\%$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mo>%</mml:mo> </mml:math> (1/6) shorter analysis run duration. The improved delayed detached Eddy simulation (IDDES) method is employed on a fine mesh to validate the optimized geometries with a higher-fidelity solution. Flow simulations are performed using the OpenFOAM HISA open-source solver. The acoustic signal, in other words the acoustic quantities (SPL, PSD and OASPL), is obtained by performing FFT transformation from probe-based, unsteady (time-dependent) pressure data obtained via the flow simulations. The results indicate that both low-fidelity and high-fidelity solutions accurately represent the aeroacoustic behavior of the cavity flow compared to the experimental data. Furthermore, the Bayesian optimization process achieved a reduction of approximately 5 $$\%$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mo>%</mml:mo> </mml:math> (7–10 dB) in OASPL and SPL values.