Los puntos clave no están disponibles para este artículo en este momento.
• Proposed a ‘part-span approximation’ for in-duct tone noise predictions. • Characterised errors introduced by the approximation. • Applied to noise radiation from a realistic intake geometry. • Demonstrated a potentially significant computational saving. A part-span methodology is proposed for the propagation of linear tone noise in aeroengine intake ducts. For fan rotor-associated tone noise sources in particular, but also rotor-stator interaction tones, much of the acoustic energy flux is concentrated towards the duct outer wall, offering the opportunity to reduce the computational expense of numerical predictions by neglecting low-span regions of the duct. Conventional computational fluid dynamics (CFD) methods can predict the acoustic field with reasonable accuracy but are too computationally expensive for iterative application during the design process. The methodology is investigated in four stages, starting with the impact on mode shape representation and in-duct propagation of individual Fourier-Bessel eigenmodes in a cylindrical duct through to far-field radiation. The approach is then applied to cases involving an axisymmetric, yet otherwise realistic, intake geometry, including a test case that considers steady flow distortion effects. Negligible errors for in-duct and far-field predictions are observed in many applied cases. Higher error levels are introduced for modes at low azimuthal harmonics and with ducts of high hub-tip ratios. In-duct, the error is closely linked to the amount of acoustic energy truncated in an approximated duct with an artificial hub wall and is associated with the redistribution of high-order radial mode components. Potential sources of error are identified and categorised: source truncation error, in-duct propagation error, and far-field radiation error. Computationally, the approximation can significantly reduce the size of the mesh required to resolve the acoustic field.
Binns et al. (Mon,) studied this question.