Abstract
The shape, curvature and diffusivity of the intake and exhaust manifolds in the reciprocating engines have a strong influence on the charge exchange and so the performance, fuel consumption and the emission of the vehicle. The pressure and velocity variation in the manifolds are controlled by the internal solid walls and these surfaces are not necessarily aligned with the flow. Several stagnation points, separations, chocked flow conditions can be evolved in the complex geometry. Hence, a surface morphing method is developed, tested and presented herein to improve design specifications by means of reaching the optimal wall shape belongs to the previously imposed favourable graduated pressure distribution.
The iterative CFD (Computational Fluid Dynamics) based process contains three main steps, which are repeated till the target conditions are reached. The predefinition of the smooth target pressure distribution is based on the engineering experiences or from theory. The first step is a direct analysis, which expects an initial geometry to be available. The second step is the opening wall analysis. The wall becomes an inlet or outlet locally and the flow can enter or leave the computational domain depends on the local pressure gradient between the solid boundary and the inner flow field. The velocity distribution is the outcome of this process in the first computational cell row. This data will be used in the third step, in which the geometry modification algorithm is executed. The procedure starts from a specified grid point, usually near from the upstream stagnation point. The wall surface is modified on such a way, that the new cell interface becomes parallel with the local velocity vector and it is continued generally till the downstream stagnation point. The calculation loop based on the direct, opening wall analysis and the wall modification algorithm and it is repeated till the target pressure distribution and so the final wall shape is reached. The mentioned procedure generally takes 3-10 steps till the expected geometry is evolved. However, several other constraints (available geometrical space, wall thickness, producibility etc.) must be kept in order to avoid unrealistic solutions.
The CFD solver is based on the compressible Euler equations. Cell centre finite volume approach has been used for discretizing governing equations. The convective terms are discretized by Roe approximated Riemann method. The MUSCL (Monotone Upstream Schemes for Conservation Laws) approach is implemented for higher order spatial reconstruction with Mulder limiter for monotonicity preserving. The system of equations is solved by the 4th order Runge-Kutta method.
Keywords: CFD, optimisation, inverse design, internal flows, manifolds