PROJECTIVE NOISE REDUCTION ALGORITHM FOR NEGATIVE PRESSURE WAVE SIGNAL PROCESSING

摘要

The appearance of a rupture, leak or damage in the long-distance oil & gas pipeline, which could cause a leak, usually generates a non-linear & chaotic negative pressure wave signal. By properly interpreting the negative pressure wave signature, it is possible to detect a leak along the pipeline. Most traditional noise reduction methods are established based on the linear system, which are not in line with the actual non-linear & chaotic situation. Therefore, the weak negative pressure wave signals, generated by small leaks, are often filtered out and cause false alarm and failure alarm. In order to resolve the problem, this paper uses the non-linear projective algorithm for noise reduction. First, the weak negative pressure wave signal series would be reconstructed using delay coordinates, in the high dimensional phase space, the background signal, the negative pressure wave signal and the noise signal are separated into different sub-spaces. Through the reconstruction of sub-spaces, the weak pressure wave signal can be isolated from the background signal as well as the random noise component reduced.