Effect Investigation of Important Parameters in Knock Detection Using Cylinder Pressure Data

Several parameters have significant effect on measure and analyzing of cylinder pressure to obtain accurate knock data. Acquisition frequency, number of cycles, transducer location and transducer mounting are measuring parameters. Filtering frequencies, knock windowing, knock model features and threshold values for knock intensity are important analyzing parameters. Effects of these parameters on knock detection and methods for optimal selection are investigated by analyzing test bench datasets. Note that optimalmethods aren’t for special operating conditions and almost cover all operating points of engine. Important parameters for analyzing of transducers’ data, such as digital filtering and Fast Fourier Transformation (FFT), are introduced. FFT determines the main oscillation frequencies and digital filtering omits undesired frequencies.Investigations of test results show that knock is occurred in special frequencies because of acoustic vibration modes, some methods for determining of these frequencies are introduced. It should be mentioned that the main oscillation frequency due to knock is about 8 kHz. Exact determination of these frequencies is important for investigation of filtering and optimal transducer location.