A 1D/3D Methodology for the Prediction and Calibration of a High Performance Motorcycle SI Engine

AbstractNowadays internal combustion engines development is widely supported by 1D and 3D codes. On the other hand, an extensive experimental activity at test bench involves increased development cost and time-to-market of a new engine. For this reason, a numerical methodology capable to provide a reliable estimation of engine performance starting from a reduced set of measured data represents a very promising approach.In this paper, a hierarchical 1D/3D numerical procedure is proposed with reference to a motorcycle naturally aspirated spark ignition engine to predict its performance even in absence of experimental data.To this aim, the real engine is geometrically characterized through a reverse engineering process. Different measurements including 3D cylinder geometry, main lengths and diameters of intake /exhaust systems and valve lift profiles, are carried out. Then, a 1D model of the whole engine is realized within the GT-Power™ code, while a 3D model of the sole cylinder is developed within ANSYS Fluent™ environment.The exchange of data between 1D and 3D models starts with preliminary 3D CFD analyses, performed to evaluate the discharge coefficients of intake and exhaust valves. The latter are passed to 1D model to compute the time-varying boundary conditions at the intake and exhaust head ducts, under motored operation. Multi-cycle 3D CFD analyses are hence carried out to describe the in-cylinder mean and turbulent flow fields in motored conditions. The mass-averaged 3D results are then used to tune the turbulence sub-model included in the 1D engine model.The last step of the procedure is the computation of the engine performance under fired conditions at full load by means of the 1D simulation. The numerical/experimental comparison of performance parameters demonstrates that the proposed methodology is capable to satisfactory describe the overall engine behavior even in absence of detailed experimental data