Misfire is the condition where the engine does not fire correctly due to an ignition miss or poor combustion of the air fuel mixture, resulting in serious deterioration of tailpipe emissions due to the discharge of unburned gas. In order to prevent further exacerbating environmental problems, misfire detection is obligatory in On Board Diagnosis (OBD) II systems.
OBD II technology for passenger cars cannot be easily adopted to motorcycles for several reasons. However, very little research has been reported on misfire detection for an unevenly firing engine in which the degree of contribution to engine output and the variation pattern of angular velocity show a large difference between cylinders, an aspect that is unique to motorcycles.
This research focuses on uneven firing V-twin motorcycle engines, to explore misfire detection techniques using variation characters in crank angular velocity.
Firstly, algorithms were proposed to be applied either individually or in combination utilizing methodology such as, identifying the portion of the combustion cycle in which combustion characteristics are exhibited for each cylinder from the crank angular velocity variation and to define judgment parameters by calculation of the differences, or setting angular velocity at ignition timing of each cylinder as a reference, and calculating the accumulative relative angular velocity up to another identified point to be used as an index parameter.
For the next step, performance of the proposed algorithms were evaluated efficiently and quantitatively, using a simulated environment constructed from various measurement data.
The algorithm which showed higher potential in the simulation were then entered into an Electronic Control Unit (ECU) in order to confirm their effectiveness through vehicle testing.
Finally, this paper summarizes our recent progress in researching misfire detection algorithms for motorcycles, and suggests directions for future research.