The track inspection vehicle data is utilized for fault feature extraction of jointless track circuit compensation capacitor using CEEMDAN

In order to keep pace with the advancement of track inspection vehicle operation and maintenance technology, this paper proposes a fault feature extraction technique for compensation capacitor based on the induced voltage data acquired by the track inspection vehicle. Firstly, a finite element model is developed to simulate the induced voltage acquisition process, utilizing the principles of high-speed track inspection vehicle and electromagnetic field theory. Next, the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) method is chosen to extract the fault features by comparing the effectiveness and orthogonalization of three signal decomposition methods, namely Empirical Mode Decomposition (EMD), Ensemble Empirical Mode Decomposition (EEMD), and CEEMDAN. The decomposed modal components are used to calculate the energy value, which is subsequently used as feature information for fault classification using a neural network. Finally, the proposed method is tested on actual railway field data, and statistical analysis shows an accuracy of 87%.