Valve clearance fault diagnosis of an internal combustion engine based on wavelet packets and k-nearest neighbors

At present, in order to improve the diagnostic accuracy, many machinery diagnosis systems with intelligent classification have been developed, such as Support Vector Machine (SVM) (Yang, 2014; Kankar, 2011; Wu, 2012; Wu, 2012) and the Artificial Neural Network techniques (ANN). They, especially SVM, are powerful classification tools for small sampling and nonlinear data. However, SVM is sensitive to missing data and has no certain solution for nonlinear problems. ANN needs a large number of parameters and its learning process cannot be observed, which will affect the credibility and acceptable degree of the results.