We demonstrate a procedure for designing and optimizing the scavenging properties of a two-stroke engine. A method of surface design, known as the partial differential equation method, is used to represent the internal geometry of the engine. We then run steady-state computational uid dynamicscalculations to assess the scavenging ow inside the engine. To quantify the scavenging, the trapping ef ciency is used, and calculated values of this measure are obtained for different geometrical designs. We use a locally weighted regression-smoothing model to t the data, and then we optimize the resulting response surface. The optimization captured two local minima, one of which corresponds to accepted manufacturing practice and the other one is a new and counter-intuitive way of scavenging the engine, having optimum scavenging characteristics.
Nomenclature
C= contour M = mass of air n = number of particles r 0 = bore of the cylinder u, v = surface parametric coordinates V = volume of air X = vector of Cartesian coordinates of surface points X n = normal derivative along the boundary x, y, z = Cartesian coordinates xsu = derivative parameter of the transfer port, in x direction, on C 0 ysl = derivative parameter of the transfer port, in y direction, on C 1 ysu = derivative parameter of the transfer port, in y direction, on C 0 zsu = derivative parameter of the transfer port, in z direction, on C 0 a = smoothing parameter q = density Subscripts s = supplied t = trapped 0, 1 = upper and lower surface boundaries, respectively