Search Results

The fiber optic spark plug was used in conjunction with a piezoelectric pressure transducer to collect combustion diagnostic data on four production engines designed to generate quiescent, swirl, and tumble charge motions. Spark advance was varied under low speed, low load conditions to investigate changes in flame kernel behavior and in-cylinder charge motion as functions of crank angle and spark advance. Two flame kernel models were filled to the data and a critical comparison of the models was conducted. Flame kernel behavior was represented by three values: convection velocity, growth rate, and convection direction. Convection velocity was highest in the swirl chambers. It also varied considerably among cylinders in the same engine. Growth rate correlated well with 0-2% burn but showed negligible correlations with later burn or IMEP. Convection direction proved useful in determining flow direction near the plug.

One of the characteristics of nominally homogeneous charge spark ignition engines is a pronounced variation in the combustion rate from cycle to cycle. Many theories have been advanced which attempt to explain the fundamental origin for differences on a cyclic basis. In the present work, some of the suspected causes or their manifestations have been incorporated into Ford's engine combustion model with the intention of determining if their impact on the combustion rate is as theorized. It has been found that initial spark kernel burn rate, the displacement of the spark kernel from the spark plug gap, and the turbulence intensity must all be perturbed simultaneously on a cycle-by-cycle basis to cause the cycle simulation program to mimic the experimentally determined burn parameters with respect to their averages and distributions.

Current demands for high fuel efficiency and low emissions in automotive powerplants have drawn attention to the two-stroke engine configuration. The present study measured trapping and scavenging efficiencies of a firing two-stroke spark-ignition engine by in-cylinder gas composition analysis. Intermediate results of the procedure included the trapped air-fuel ratio and residual exhaust gas fraction. Samples, acquired with a fast-acting electromagnetic valve installed in the cylinder head, were taken of the unburned mixture without fuel injection and of the burned gases prior to exhaust port opening, at engine speeds of 1000 to 3000 rpm and at 10 to 100% of full load. A semi-empirical, zero-dimensional scavenging model was developed based on modification of the non-isothermal, perfect-mixing model. Comparison to the experimental data shows good agreement.