Comparison of Boundary Condition Models and Numerical Methods for the Gas Dynamic Effect of a Two-Stroke Engine 950985

To formulate a computer model for a two-stroke engine, the boundary conditions have to be solved for cylinder or crankcase to pipe flow, pipe flow to atmosphere. The models to deal with the former and the latter are often called partially open end model and open end model respectively. In this paper, two types of models for partially open end and two for open end are investigated by comparing the predicted pressure data in the exhaust pipe. Also, two different ways to model transfer pipe are considered. One treats the flow in transfer pipes as quasi-steady flow while the other simulates that with an one-dimensional gas dynamic model. The simulation is based on a zero-dimensional thermodynamic model for cylinder as well as crankcase and an one-dimesional gas dynamic model for pipes. The former uses 4th-order Runge-Kutta method and the latter uses a combined 4th-order Runge-Kutta and two-step Lax-Wendroff (4RK/2LW) method. The results of calculation compared with the measured data show that the difference between the partially open end model with constant-pressure assumption and that with a momentum equation instead is not obvious. As for the two open models, the one using finite-difference method can give a better agreement in terms of the pressure histories. Furthermore, the attemp to simulate the unsteady nature of gas flow in the transfer pipe using an one-dimensional gas dynamic model does not bring forth better results as expected. On the contrary, it shows a better agreement with the measured data to treat the gas flow in transfer pipe as quasi-steady flow.