Computational AeroAcoustics of Realistic Co-Axial Engines

Abstract

This study, that is relevant from the turbofan engines noise prediction/reduction, aims at CAA-computing the aft fan noise propagation/radiation of a realistic full-3D exhaust (with pylon and internal bifurcations), the latter being affected of (i) typical in-flight (take-off) thermodynamic conditions and of (ii) a representative fan noise modal content. As for previous studies conducted over baseline geometries, this CAA computation is conducted following the usual hybrid process, where a preliminary aerodynamic calculation provides a heterogeneous steady mean flow on which an acoustic calculation is then conducted A RANS computation is first performed, delivering the stationary jet mean flow characterizing the 3D exhaust in its typical 'take-off flight' (M∞ = 0.25). A CAA grid (22 blocks, 28 millions cells) is then derived from the CFD one, before the RANS steady jet mean-flow is interpolated on it. After what the CAA computation is computed, a fan noise mode (26, 1) being emitted at a reduced frequency of kR = 30 (1 BPF) in the upstream of the engine's secondary exhaust, and numerically propagated along and outside the latter. Finally, a Kirchhoff post-treatment provides the far-field radiation characterizing these engine geometry, modal content and thermodynamic conditions.

Geometry optimization for reducing fan noise propagating from a bypass duct

“…The background mean flows for an axi-symmetric generic bypass duct case are solved first. The geometry and flight conditions are issued from an Airbus-France/ONERA collaboration in the framework of AMBIANCE Project 35 Fig. 10.…”

The non‐reflective interface: an innovative forcing technique for computational acoustic hybrid methods

“…Such an interface technique was not only straightforward to implement and to use, but it was also general enough to be applicable to many situations (at least, provided that the latter interface is defined with enough care ‡ [15,16]). As an illustration, regarding internal noise propagation problems, such an interface technique had been successfully applied to various cases (e.g., the aft fan noise emission by turbojet engines [17,18]), for which the acoustic datasets used to force the CAA simulation had been constructed via analytical means (based on the modal theory [5]). For external noise problems, the technique had also been applied with success to several airframe noise applications (e.g., the acoustic emission by a NACA0012 airfoil with a blunted trailing edge [15,19] or by a thick plate embedded within a flow [20]).…”

“…The first part of this section reviews the generic CAA governing equations used within the present context. This formulation was developed at ONERA a decade ago [13,14] and is now implemented in ONERA's time domain CAA solver sAbrinA [13,14,17,18]. This formulation is very general and solely relies on a small-perturbation hypothesis.…”

“…All CAA calculations were achieved by using ONERA's sAbrinA solver [13,14,17,18], which is a structured, time-accurate CAA code that solves either the full or the linear Euler equations, in a conservative or perturbed form. The solver employs high-order, finite-difference operators, involving sixth-order spatial derivatives and tenth-order filters, as well as a third-order, multi-stage, Runge-Kutta time-marching scheme.…”

“…More detailed information about the sAbrinA solver and its underlying methodology can be found in Refs. [13,14,17,18]. All CAA computations were conducted on one core of a laptop (64 bit, Intel® Pentium® 2.1GHz, 4GB RAM), requiring either a few (for case number 1) or 30 (for case number 2) minutes of wall clock time.…”