Search

International audience; This paper is aimed to present a computational chain developed in the framework of the CleanSky 2 European CONCERTO project. This computational chain is devoted to the prediction of the noise emitted by the lateral rotors of the Airbus Helicopters' RACER demonstrator using both fast prediction and high fidelity approaches. One of the main challenges addressed in this paper is the capacity to predict rapidly the aerodynamic and acoustic interaction effects due to the vicinity of the wings and the fuselage. From an aerodynamic point of view, it is proven by comparison with high-fidelity results that it can be achieved thanks to a free-wake code using velocity perturbation fields to properly take into account the wings wake viscosity effect and to solve issues linked to the rotational speed difference between the main rotor and the propellers. Comparisons with a Boundary Element Method code show also that a Kirchhoff approximation appears to be a good option to rapidly assess, at least in terms of tendencies, the impact of near solid surfaces on the noise radiation.

In the usual approach to multi-disciplinary optimization (MDO), a few experts in optimization select and integrate disciplinary tools into a working MDO process, run it on target cases, and analyze the results. This approach is expedient, but it tends to oversimplify employment of disciplines. German Aerospace Center (DLR) is currently pursuing a more cooperative approach, where disciplinary experts from different DLR institutes are involved throughout in the definition, implementation, and evaluation of MDO processes. The goal is to establish a modular MDO platform, with well-balanced fidelity levels of employed disciplines, that can be used both for overall aircraft design, as well as for studying the influence of particular subsystem improvements on the overall design objective. Two MDO processes resulting from this effort so far are presented. The objective in both process is to minimize the mission fuel burn, as evaluated through Breguet range equation. One is a derivative-free process employing multi-level analysis, comprising the conceptual design level for determining the dependent top-level parameters, dynamic level for computing the sizing loads, and detailed level for sizing the structure and computing the performance indicators. The other is a gradient-based process on the detailed level alone, employing a split of design parameters between CAD-based and CAD-free, and treating all structural constraints independently. The set of sizing-relevant load cases is automatically determined based either on the actual configuration (in the derivative-free process) or on the baseline configuration (in the gradient-based process), rather than applying a set of generic load cases. The contribution concludes with several example studies carried out using these processes.

This report presents an organizational approach towards assembling and maintaining high-fidelity multi-disciplinary optimization (MDO) processes, in the environment of multiple institutes of the German Aerospace Center (DLR). Elements of the approach include both the technical conventions and the specialized software support. The approach is demonstrated on an early MDO process, that minimizes the mission fuel burn of an aeroelastic transport aircraft configuration, by modifying the wing planform, twist, and structural element thicknesses. Two optimizations are performed, one where the fuel burn is approximately evaluated through Breguet range equation, and the other where the ordinary differential equation (ODE) for the step-climb cruise is formally integrated. This is done in order to determine if the Breguet equation is still sufficient in face of high-fidelity aeroelastic simulations. The two optimized designs ended up practically the same, thus confirming the applicability of the Breguet equation, at least for the limited number of design parameters that were open for modification.

In this paper the FlowSimulator framework for multi- disciplinary computational fluid dynamics simulations on high performance computer platforms is described. An overview of the provided functionality is given and possible benefits for tool developers as well as design engineers are presented.