Investigating the feasibility of characterising gasoline autoignition using a motored engine apparatus

• Main Author: Demnitz, Simon
• Other Authors: Yates, Andrew
• Format: Dissertation
• Language: English
• Published: University of Cape Town 2015
• Subjects: Mechanical Engineering
• Online Access: http://hdl.handle.net/11427/14645

Development of a predictive octane model is a potentially useful tool for designing fuel blends for meeting octane specifications. One of the approaches adopted is through chemical kinetic modelling of the autoignition properties of constituent compounds. The results obtained from models, however, are dependent on experimental data for validation. It was the intention of this thesis to provide empirical data that could be used confirm a recently proposed autoignition model based upon the results obtained from chemical kinetics modelling. Motored engines have been used extensively for the investigation of autoignition properties of fuels. They are useful in interpreting results from conventional ignition delay measuring systems as well as giving practical insight into the process of autoignition in spark ignition engines. The conditions required for autoignition reactions to take place are easily produced in a motored engine with a suitable compression ratio. A single cylinder engine was modified so that the inlet conditions could be adjusted and n-heptane was tested in the device. Fuelling was controlled with an injection system which was calibrated for n-heptane use in the engine. A range of inlet conditions were determined that would enable peak conditions in the engine to result in autoignition of the fuel. The autoignition data was then used in describing the ignition delay characteristics of the fuel and the range of interest, the so called negative temperature coefficient region. Autoignition experiments were performed in the engine and the data was analysed by the comparison of measured autoignition reactions with predicted reaction times; the predictions were calculated using the new empirical autoignition model. Direct analysis of the model resulted in good correlation of measured and predicted overall autoignition reaction times, with improved correlation of cool flame reaction times with initial temperature adjustment. Modification of initial temperature values in the indirect model application (whereby traces were generated using an engine model with autoignition prediction capabilities) resulted in similar observances. These initial results led to the conclusion that the temperature and Arrhenius parameter adjustments necessary to obtain a perfect fit in the autoignition model were indicative of errors involved in the temperature measurement or in the fuel metering. Recommendations for further work on the engine would be the investigation of a dynamometer system that would be free from noise transmission during operation and that would enable experimentation with lower engine speeds. Further work on the inlet system would be the installation of shielded thermocouples and a quicker acting heater controller. A fundamental change in fuel metering calibration is required. Further recommendation is that a variable compression ratio engine should be used to enable the attainment of a wider range of readings for fuel characterisation and possibly eradicate the problems experienced with fuelling.