Engine knock detection and intensity evaluation based on sparse maximum correlation kurtosis devonvolution

In order to solve the problem of light knock detection, the research is focused on engine knock feature extraction and intensity evaluation with engine block vibration signal. A signal processing method based on sparse maximum correlated kurtosis deconvolution (SMCKD) is proposed. Firstly, the sparse representation is used to extract the internal signal characters. Meanwhile, a sparse representation redundancy dictionary covering the knock features is established. Secondly, the maximum correlation kurtosis deconvolution (MCKD) is used to identify the knock features and eliminate the irrelevant features. Therefore, the knock state can be accurately distinguished. An engine knock intensity index based on SMCKD is also presented. And the knock intensity can accurately distinguish normal combustion, light knock and strong knock. It is verified that the method of extracting knock feature and the evaluation index of knock intensity presented in the paper have a certain value in engineering application.