Detection of Automotive Emissions Status using Fuzzy Inference System

Abstract

The adoption of a fuzzy control system to determine the emission values of CO, CO2, NOx, HC and O2 from an SI engine is studied depending upon the air fuel ratio (A/F), as known that the emission values are highly related to A/F ratio, the target of the combustion process in the internal combustion engine is to be as close to stoichiometric as possible, however the driving conditions play a significant role in the value of this ratio. Fuzzy logic is used to represent, interpret, and compute vague and/or subjective information of automotive emissions. Fuzzy logic approach uses human language and thinking strategies in a way for decision making based on domain expert knowledge, and rules of thumb even for systems without any mathematical model. I. Introduction The study of the SI engine's emissions is an important issue to the automotive manufacturers as these emissions affect the environment from one side and because the limits of these emissions is controlled according to the regulations set by many agencies like EPA (1). Fuel consumption and emission models using vehicle speed and acceleration levels as input variables were studied in (2), the models were found to produce vehicle emissions that were consistent with the ORNL data. The use of 3 way catalytic converter which promotes the reduction of nitrogen oxides and the oxidation of carbon monoxide and hydrocarbons in gasoline-fueled vehicles was discussed in (3). Controlling the air-fuel ratio with an oxygen sensor used with a three-way catalyst showed a significant reduction of CO, NOx, emissions (4). The effect of the change in driver's behavior and the variability associated with individual driving styles like duration of acceleration events showed significant differences in measured emissions (5). The effect of the air-fuel equivalence ratio and ethanol content in gasoline showed that CO and HC emission content were reduced with an increase of ethanol content, this was due to oxygen enrichment coming from ethanol (6). It was noticed in (7) that NOx emission was closely related to the equivalence ratio, such that NOx emission reaches a maximum near the stoichiometric condition and NOx emission depends on the engine operating condition rather than the ethanol content in gasoline. Finally , The cold starting emission levels were studied in (8) using ethanol blended with gasoline, the study showed that for best cold-start emissions the ethanol content in gasoline should be at least 20 percent but no greater than 30 per cent. Fuzzy logic (9) is a well-established methodology that is effective for systematic handling of deterministic uncertainty and subjective information. Fuzzy logic is useful especially when a mathematical model of a system is not available and rules of the thumb from domain experts are available (10). It has been successfully used to solve challenging industrial and medical problems in practice, some of which are very difficult to solve without it. Using Fuzzy rule based approach in this paper is dependent upon air fuel ratio as an input to the system to well define the values of the five major engine emissions. This system makes it easy to know exactly the values at any operating condition which can be used in the future with the ECM to reduce emissions. In this paper the reasoning is implemented using MATLAB Fuzzy toolbox. The detecting task of the proposed system is performed by the Inference Engine that evaluates all the rules in a knowledge Base and combines the weighted consequents of all relevant rules into a single output fuzzy set that shows the status of an emission. That set is then defuzzified to produce a crisp emission status value. The automotive emission status detector is experience-based as experience plays a key role in the design of the Detector.