Blind source extraction of acoustic emission signals for rail cracks based on ensemble empirical mode decomposition and constrained independent component analysis

In order to detect rail cracks by acoustic emission (AE) technology, a constrained independent component analysis algorithm is proposed to extract investigated components sequentially. Standard independent component analysis is modified through the iterative constraints of higher-order statistics, including the recursive kurtosis and waveform factor, to extract the required signals more accurately and efficiently. The proposed method is combined with ensemble empirical mode decomposition and initially verified by a single-channel simulation, which manifests a better separation efficiency and a higher robustness in recovering sources. To further testify the practicability, constrained independent component analysis is applied in single-channel experiment in an actual operating railway. It can be concluded that the proposed method is effective to detect crack signals in real noise environment and occurrence of crack signals can be distinguished from noisy mixtures, which can provide a guidance in the actual application of AE detection of rail cracks.