Search Results

Experimental data was obtained from a Rover K4 optical access engine and analyzed with a combustion analysis package. Cyclic NO values were calculated by mass averaging the measurements obtained by a fast NO analyzer. While the mass averaged results were used as the basis of comparison for the model, results indicate that mass averaging a fast NO signal is not nearly as critical as mass averaging a fast FID signal. A computer simulation (ISIS - Integrated Spark Ignition engine Simulation) was used to model the NO formation on a cyclic basis by means of the extended Zeldovich equations. The model achieves its cyclic variability through the input of experimentally derived burn rates and a completeness of combustion parameter, which is based on the Rassweiler and Withrow method of calculating mass fraction burned and is derived from the pressure-crank angle record of the engine.

This work examines the possibility of detecting misfires via measurements of the angular acceleration of the engine block. Measurements were taken on a production 4-cylinder engine which was modeled as a single degree of freedom torsional oscillator. The torque waveform was estimated and compared to the torque calculated via cylinder pressure measurements. Further analysis was conducted in the frequency domain. Results indicate that metrics based on low frequency information were most reliable, but this is impractical for vehicular applications. The accuracy of high frequency metrics was degraded due to the limitations of the model and the non-rigid behavior of the block at high engine speeds.