Automobile Independent Fault Detection based on Acoustic Emission Using Wavelet

Wavelet transform is a well-known signal processing technique which has proven to be a successful technique in Non-Destructive Testing (NDT). Adopting this technique by Acoustic Emission (AE) method could result in better interpretation of data that are used for fault detection. In this study, AE signals emitted from automobile engines in both faulty and healthy conditions are analyzed using wavelet transform. Here, intension is to categorize the AE signals into healthy and faulty classes. The investigated fault is within the ignition system of the engines beside other possible problems that may affect the generated acoustic signals. A set of features are extracted from signals. Correlation-based Feature Selection (CFS) algorithm is used to reduce the dimensionality of the dataset. The case study is carried-out on 4 different types of automobiles using 480 automobiles to prove the independency of the proposed approach on the type of the automobile. Classification results are reported to be around 80%.