Enhancement of knocking detection accuracy by an ion-current sensor integrated in the ignition system

Abstract

Abnormal combustion such as pre-ignition and knocking is becoming one of the biggest problems in the latest gasoline engines that have a higher compression ratio and boosting for higher efficiency. An ion-current sensor integrated in an ignition system is used for accurately detecting knocking cycles. First, the problem for accurate knocking detection with an ion-current sensor was clarified in the test engine. The oscillation in the ion-current signal was observed in knocking cycles as is commonly known in the previous research. However, heavy oscillation in the ion-current signal can be observed occasionally even in the small knocking cycles. This phenomenon leads to the misdetection of knocking cycles with the conventional signal-processing method, which defines the oscillation intensity of the change amount in the ion-current signal as a knocking indicator. Second, to solve the problem mentioned above, a new signal-processing method is proposed on the basis of the thermal characteristics of ion-current signals. This method defines the oscillation intensity of the “normalized ion-signal change rate” as a knock indicator in order to suppress the effect of temperature dependency in ion-current signals. Finally, the proposed method was applied to an actual gasoline engine, and the knocking detection performance was evaluated. The method enabled the misdetection of the knocking cycles to be avoided and enhanced the correlation factor with knock intensity compared with the conventional method. vibration sensor equipped on the engine block. Yang et al. (2010) developed a signal-processing method to detect the low intensity vibration from short data series, on the basis of kurtosis of vibration intensity. Momeni et al. (2016) proposed a normalization model that enables a fixed detection threshold in all head can be changed, and the signal intensity of the ion current can be adjusted.