Control of cyclic variations by computing the cycle-to-cycle air-fuel ratio by neuro fuzzy technique

ISSN： 0264-4401

Source： Engineering Computations: Int J for Computer-Aided Engineering, Vol.24, Iss.8, 2007-11, pp. : 780-792

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

Purpose ‐ The problem of cyclic variation has been an interesting area of research and has been investigated by many researchers. It is more severe in the case of two-stroke engines compared with four-stroke engines. One of the reasons for these cycle-to-cycle variations is the variations in the air-fuel ratios of individual cycles and, if these values of individual cycle air-fuel ratios are available by some means, they can be used for controlling the cyclic variations. The purpose of this paper is to find a technique to predict the air-fuel ratio of the individual cycles and use the same for reducing cyclic variations. Design/methodology/approach ‐ In this work, a neuro-fuzzy model was developed using MATLAB software to compute the air-fuel ratio of the individual cycles based on the relationship between the air-fuel ratio and the combustion parameters such as those indicating mean effective pressure (IMEP), crank angle occurrence of peak pressure, and angles of different percentages of heat releases. In-cylinder pressure traces of 1,000 continuous cycles were measured using a Personal Computer (PC)-based data acquisition system and an investigation was carried out. The readings were taken for two modes of operations, namely gasoline carburetion and electronic gasoline injection. The engine was loaded by an eddy current dynamometer. The air-fuel ratio was varied from rich to lean by adjusting the fuel quantity in the carburetion mode and adjusting the pulse width (measure of quantity of fuel to be injected) in the injection mode, at constant throttle. The cyclic variation was identified by the variations in the peak pressures and IMEPs of the individual cycles. The stored data were given as input to the developed neuro-fuzzy model and, using SIMULINK, the air-fuel ratios of individual cycles were obtained. These predicted values are fed to the electronic control module (ECM) (meant for injecting the fuel) for refining the pulse width to get cyclic variations reduced. Findings ‐ Results show that cyclic variation increases when the mixture becomes lean. It was also found that cyclic variation in an injected engine was less in comparison with the carbureted engine, as the precise control of air-fuel mixture was possible in the case of the injected engine. Research limitations/implications ‐ The technique used in this work may be modified to give more precise pulse width by incorporating various other parameters like exhaust temperature, etc. Future research may be focused to incorporate this system in a moving vehicle to get more fuel efficiency and fewer emissions. Practical implications ‐ The design of vehicle and engine should be slightly modified to incorporate the ECM and various sensors. Originality/value ‐ The originality in this paper is that a new technique was developed to find the air-fuel ratio of individual cycles. This will be useful for the engine manufacturers and for those researchers doing research on the engine side.