Shape optimization of an inlet pipe to an engine re-circulator cooler using the adjoint method is presented. The method uses surface sensitivities calculated from an adjoint flow field implemented in the finite volume CFD solver OpenFOAM® [1]. This method allows for computation of the whole sensitivity field with only two solver calls, a primal and an adjoint solver call. A RANS solver with the standard k-epsilon turbulence model applying standard wall functions was used for the primal flow solver. The adjoint surface sensitivities are calculated from the adjoint and the primal flow fields and give information about how the objective function is affected by normal motion of the surface. The surface sensitivities are coupled to a mesh morphing library in OpenFOAM diffusing the motion of the boundary nodes to the internal cells of the mesh. The resulting geometry gave a 6.5% decrease in the total pressure drop through the pipe.

No abstract

A sensitivity analysis of a particle distribution at the outlet of a 2D-channel is presented. A convection-diffusion equation is used to represent the distribution of particles in a flow. The flow is governed by the Reynolds-Averaged Navier-Stokes equations. The particles follow the medium to the outlet, where the goal is to obtain uniform distribution of the particles. A goal function for the particle distribution at the outlet is presented, and the gradient of the goal function with respect to the normal motion of the surface is calculated. The calculation of the gradient is performed by applying the so-called adjoint method. For this, the adjoint scalar transport equation and the relevant adjoint RANS equations are implemented. The calculation cost of the entire sensitivity field is roughly twice the cost of a standard convection-diffusion particle simulation. This results in computationally cheap gradients compared to traditional methods. The results are verified by comparing the gradients calculated using the adjoint method to gradients obtained using numerical differentiation.

The paper presents results of simulations of the flow around rudimentary landing gear using an improved version of the Partially-Averaged Navier Stokes model (PANS ζ-f ). The results are validated against time-averaged flow data such as pressure field and oil-film visualizations as well as quantities such as sound pressure level and surface pressure spectra that are relevant for aero acoustic noise generation. The results of PANS ζ-f are found to be in good agreement with the experimental data and observations. The PANS ζ-f presented here shows a clear advantage in the prediction of the flow compared with the reference LES simulation on an identical grid.

This paper presents results from adjoint-based optimization processes applied to an inlet pipe of an exhaust gas recirculation cooler in a diesel engine. The boundary conditions applied resemble those of a truck at cruising speed. Three implementations are considered for the gradient calculations with the objective of minimizing the total pressure drop through the pipe. In the first implementation the gradients are evaluated with respect to the motion of the center of the cell using a newly presented implementation based on the ALE formulation of the Navier-Stokes equations. The results are compared to the surface sensitivities, where the gradient of the cost function is evaluated with respect to the normal motion of the surface of the pipe. In the last approach a topological optimization is performed where the gradients are evaluated with respect to a momentum loss in each cell. This gives information that is used when blocking the cells.