A Deterministic Method for Real Time Detection of Misfire for Smaller Capacity Spark Ignition Engine

• Main Authors: Bagade, Monika Jayprakash, Das, Himadri, Jabez Dhinagar
• Format: Report
• Language: English
• ISSN: 0148-7191; 2688-3627
• DOI: 10.4271/2021-24-0031

As per the OBD II regulations, it is essential to detect and monitor the misfire event in an I.C. engine. Misfire in the engine causes incomplete combustion or no combustion, which results into increase in emission of exhaust gases as well as early aging of 3W catalytic converter. Misfire event can be categorized as partial or full, based on amount of combustion occurred during that particular engine cycle. There are several techniques available in previously published work for the detection of misfire based on measurements of crankshaft speed, in-cylinder pressure, crankshaft torque, etc. However, low mechanical inertia and stochastic load disturbances of smaller capacity S.I. engine causes several challenges for these techniques and may not be accurate for detection of misfire in all the operating regions of the engine. Ion current sensing can be one of the most prominent method for detection misfire for smaller capacity S.I. engine. The time-varying ion current signal is captured with the help of ion current measurement circuit proposed by the author in her previous publication. The captured signal needs to be processed to detect misfire. Multiple techniques are explored in time domain to differentiate the misfire from normal combustion event with the help of the ion current signal. Real time Fast Fourier transform (RTFFT) is the method proposed in this paper to detect any kind of misfire for the mentioned engine type. The novelty of the proposed method lies in processing the signal information within a very short time window for the identification and classification of misfire events. Ion current signal shows lot of dynamics in normal combustion condition compared to that of misfire conditions. Frequency spectrum of the signal can be used to extract the essential information and to differentiate between no misfire and misfire conditions. RTFFT is performed to observe the features of signal in frequency domain. The misfire can be detected by extracting the information from the FFT spectrum of the ion current signal in every spark event. In order to capture misfire event at all engine speed, large sample sizes are required in traditional frequency domain approach. A design of FFT for real time processing of ion current signal demands very high speed and large memory hardware system. This paper investigate and propose a method to efficiently implement the RTFFT algorithm. Hardware implementation of RTFFT algorithm on the field programmable gate array (FPG