Robust Spectral Leak Detection of Complex Pipelines Using Filter Diagonalization Method

The control and managing of pipelines have been assuming a major importance for all kinds of fluids to be conveyed through. When the fluid is like oil, harmful liquid and/or water for human beings necessity, the monitoring of pipelines becomes extremely fundamental. Based on the reflexion according to fast detecting systems, spectral analysis response is a topic of interest. Among spectral analysis response techniques, fast Fourier transform (FFT) is rated. Different other techniques are utilized, but they are costly and difficult to be used. An interesting technique, used in nuclear magnetic resonance data processing, filter diagonalization method (FDM), for tackling FFT limitations, can be used, by considering the pipeline, especially complex configurations, as a vascular apparatus with arteries, veins, capillaries, etc. The thrombosis, for human vascular apparatus, that might be occur, can be considered as a leakage for the complex pipeline. The research proposes the use of FDM according to two sub techniques called algorithm I and algorithm II. The first algorithm is a direct transformation of FDM application, while the second includes robustness and a regularization technique to solve ill-posed problems that may emerge in processing data. The results are encouraging.