Use this URL to cite or link to this record in EThOS:
Title: Neural network estimation of air-fuel ratio in internal combustion engines
Author: De Zoysa, Merrenna Manula.
ISNI:       0000 0001 3420 9134
Awarding Body: University of Brighton
Current Institution: University of Brighton
Date of Award: 2003
Availability of Full Text:
Access from EThOS:
This thesis presents an investigation into a novel method of estimating the air-fuel ratio of a gasoline-fuelled spark-ignition internal combustion engine. The measurement of the air-fuel ratio is important for controlling an engine to reduce exhaust emissions. In production vehicles, the air-fuel ratio is measured using an exhaust gas analyser and the exhaust emissions are reduced by using electronically controlled three-way catalytic converters, which are expensive and subject to operationallimitations such as, requiring the engine to operate with a stoichiometric air-fuel ratio. A micro-processor based engine management system monitors the engine performance and controls various engine parameters - the fuel pulse width, ignition timing, exhaust gas re-circulation etc. - to maintain strict control of the engine and ensure optimum engine performance. In the USA and UK the engine management system is also responsible for performing on-board diagnostics and warns the driver of any problems such as misfire, knocking combustion and failure of the catalytic converter. The method of measuring the air-fuel ratio presented in this thesis, termed Spark Voltage Characterization (SVC), uses neural networks to analyse the time varying spark voltage waveform to estimate the air-fuel ratio. The spark plug is in direct contact with the combustion itself, thus making it is an excellent candidate for use as a combustion sensor. As it is already installed in the engine, no modifications are required to the engine block itself. The method uses few external components making it cheaper to implement. Preliminary investigations on this method showed that it was possible to estimate the air-fuel ratio by neural network analysis of the spark voltage waveform. As different engines are equipped with different types of ignition systems, it is important that the sensor is independent of the ignition system thus ensuring that it is able to operate with any type of ignition system. The work presented in this thesis includes: i) an extensive review of other methods of measuring the air-fuel ratio, noting the advantages and disadvantages of each method and how the SVC sensor overcome these disadvantages; ii) a description of the theoretical operation of the sensor; iii) investigation of the effects of various engine parameters on the performance of the sensor; iv) suggestions for further work to improve the sensor performance.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available