• losses reduction
• CFD
• 2D channel
• optimization

In the present work the first results are described of an investigation on the possibility of reducing the dissipation in constant-section two-dimensional channels by introducing suitably-shaped transverse cavities in the walls. The reduction of the pressure losses in ducts and channels is of great interest for engineering applications and may also be seen as a first step towards the conception of control strategies that may eventually be applied to more complex configurations.
To this aim, a possible method to reduce these losses, is to devise modifications of the geometry that may alter the flow field so that the wall viscous stresses are decreased more than the pressure drag is increased.
The considered geometry is a two-dimensional channel with a laminar flow at Reynolds numbers Re = hu_bulk/nu = 250 and 500, where u_bulk is the bulk velocity of the flow, h is the channel half-width and nu is the kinematic viscosity, and two different velocity inlet conditions have been considered, viz. a Poiseuille flow profile and a constant velocity profile. Steady-state incompressible simulations have been carried out by using a commercial code and a finite-volume space discretization.
The introduction of semi-elliptical cavities having depths b = 0.05h and 0.025h and semi-axis ratios b/a ranging from 1:1.125 to 1:10 in a channel with total length l = 50h (leaving two 5h long flat-wall portions at both extreme ends) leads to a reduction of the losses inside the channel from 2% to 8%. The greatest reductions are obtained by introducing 40 cavities, whose length is equal to h for both the considered depths (b = 0.05h and b = 0.025h). The enhanced performance is due both to a widening of the channel local cross-section and to a reduction of the viscous losses in the cavity region. The cavity region is indeed characterized by low values of velocity and vorticity.
In this latter more challenging case, in order to distinguish the effect of the viscous losses reduction in the cavity region from the one of the widening of the channel, we considered the case in which the introduction of cavities do not alter the mean cross-section area. An optimization procedure is carried out in order to identify the shape of the transverse cavity allowing the losses inside the channel to be minimized. This optimization leads to a reduction of the dissipation in the channel of the order of the 1%. The optimized shape is identified and it produces a local widening of the channel cross section of the order of 0.016h and a maximum contraction of the 1%, compared to the wall of the reference channel.