Essais & Simulations n°134

dossier mesures tures of

dossier mesures tures of the adjoint Navier–Stokes equations and the primal equations are similar. But the adjoint flow field has difficulty to converge. The main difference between those equations is the minus sign in front of the convective term of adjoint momentum equation. In fact, informations are convected upstream for adjoint solution rather than down-stream for the primal flow. As boundary conditions for adjoint velocity and pressure, we have : > Adjoint boundary conditions for the wall and inlet: > Adjoint boundary conditions for the outlet: The optimisation process is decomposed in three steps. In a first time, the physical field is solved with taking into account the turbulence equations. Then, the physical velocity and turbulent viscosity are used for the solution of the adjoint equations. Else, the calculation of the topological sensitivities with respect to the design variable is realised and the update of porosity field is performed using the conjugate gradients method. 2. Results This method has been applied to improve and optimize the performance of a charge air cooler, in figure 1. The pressure drop of inlet box obtained by the topology optimization is around 10 mbar. This final design (figure 2) has been compared to the original design computed by an iterative and classic simulation process. In this previous iterative study, the best value of pressure drop of inlet box performed was around 15 mbar. Thanks to the topology optimization, the pressure drop of inlet tank has been reduced by 32% and the global pressure drop of charge air cooler has been decreased by 12% compared to classic and iterative simulation process. In the meantime uniformity supply through the heat exchanger to maximize its thermal efficiency has been slightly improved with a reduction of velocity gradient by 2%. The complete simulation process is performed in a few days including optimisation shape definition, validation and CAD format geometry reconstruction. 3. Conclusion A fast and robust topology optimization has been implemented into a Valeo CFD simulation process from Open- Foam [3] based on continuous adjoint solver and a porosity field as term source. The result is very useful and relevant. This method offers a great advantage for improving and optimizing quickly a design in industrial context. It proposes significant improvement way of design in order to reduce the pressure drop and homogenize flow distribution through the cooler in a controlled development time. ● Olivier Brugière et Edgar Cortey, CFD Numerics Nicolas-Yoan François, Valeo Thermal Systems BG dossier Mesure acoustique et vibratoire : état de l’art À l'heure où s'ouvre nouvelle édition du salon Enova – la dernière en ce qui concerne Paris, laquelle sera remplacée par l'événement biennal Measurement World dès l'automne 2019 –, la rédaction du magazine Essais & Simulations entame la réalisation d'un nouveau dossier consacré à la mesure et à la métrologie dans l'industrie et les laboratoires d'essai. À l'occasion de ce numéro spécial, la rédaction a souhaité mettre l'accent sur la question de la vibration et de l'acoustique, thème également d'un autre grand rendez-vous, le congrès Automotive NVH Comfort. Celui-ci se déroulera quant à lui au palais des congrès du Mans et accueillera des experts venus des quatre coins du globe dans l’acoustique appliqué aux véhicules. Ce focus consacré aux applications et aux mises en pratique de solutions innovantes dans le domaine vibratoire et acoustique fait suite au dossier spécialement réalisé sur l'automobile à l'occasion du mondial de Paris. Il entend répondre aux attentes des lecteurs à la fois dans le domaine de la mesure et de toute personne intéressée par les applications dans l'automobile. Au programme de ce dossier, des avis d'expert, un reportage, des interviews et la mise en lumière d'outils innovants, pour beaucoup appliqués à l'automobile qui ne cesse d'évoluer depuis plusieurs années. ● Olivier Guillon Figure 1 Charge Air cooler description The inlet box (green part) represents the maximum volume (constraint due to the packaging of engine environment) to be optimized. The cooler (brown part) and outlet box (grey part) are frozen. Figure 2 Final design of the geometry Références [1] C. Othmer, “A continuous adjoint formulation for the computation of topological and surface sensitivities of ducted flows”, Int. J. Numer. Meth. Fluids, 2008 [2] R.P. Dwight and J. Brezillon, “Effect of various approximations of the discrete adjoint on gradientbased optimization”, AIAA-2006-0690, 2006 [3] www.openfoam.org 36I ESSAIS & SIMULATIONS • N°134 • septembre 2018 ESSAIS & SIMULATIONS • N°134 • septembre 2018 I37