A detailed mathematical model of the thermodynamic events of a crankcase scavenged, two-stroke, SI engine is described. The engine is divided into three thermodynamic systems: the cylinder gases, the crankcase gases, and the inlet system gases. Energy balances, mass continuity equations, the ideal gas law, and thermodynamic property relationships are combined to give a set of coupled ordinary differential equations which describe the thermodynamic states encountered by the systems of the engine during one cycle of operation. A computer program is used to integrate the equations, starting with estimated initial thermodynamic conditions and estimated metal surface temperatures. The program iterates the cycle, adjusting the initial estimates, until the final conditions agree with the beginning conditions, that is, until a cycle results.The combustion process is described by dividing the cylinder gases into two systems, one corresponding to the burned region and the other to the unburned region. As combustion takes place, the boundary between the two systems moves, and mass is transferred from the unburned system to the burned system. The effects of dissociation of the combustion products are included by using empirical curve fits for the equilibrium thermodynamic properties of the combustion products of air and CnH2n fuel.The instantaneous mass flow through the engine is computed from stagnation conditions upstream and downstream of each flow restriction using steady flow coefficients. The motion of the reed is described by the solution of an equivalent spring-mass system.A theoretical relationship between scavenging efficiency and delivery ratio assuming perfect mixing is employed. The delivery ratio actually used in this expression is multiplied by a factor chosen to give the best agreement between computed and experimental BHP.An arbitrary model which has characteristics in common with the real process is used to compute the vaporization rates during the cycle.The heat transfer surfaces for each system are divided into several different areas, each considered to be at one uniform temperature for the cycle. At the end of each iteration, a one-dimensional heat transfer rate balance is performed on each surface to obtain better estimates of the metal temperatures for the next cycle.Experimental data from the simulated engine are compared with the computed results.