TNO Industrie en Techniek
|Source:||36th Annual Review of Progress in Quantitative Nondestructive Evaluation, QNDE 2009, 26-31 July 2009, Kingston, RI, USA, 1211, 774-781|
|AIP Conference Proceedings|
Electronics · Ultrasonic · Tomography · Inversion · Guided wave · Corrosion monitoring · Travel time · Parameterization · Dispersion
The concept of predictive maintenance using permanent sensors that monitor the integrity of an installation is an interesting addition to the current method of periodic inspections. The method should be capable of providing quantitative wall thickness information for both straight pipes and bends. The wave propagation in bends is far more complicated than in straight pipes because natural focusing occurs due to geometrical path differences. Numerical simulations clearly show this effect. Travel time tomography requires accurate modeling of travel times that can be translated to spatial wall thickness variations. Therefore, a ray tracing algorithm has been developed to calculate travel times as part of the tomographic inversion kernel. Numerical results show that a tomographic inversion on simple simulated data provides accurate results. The focusing effect due to the shape of the bend yields a phase rotation of the wavelet, which complicates accurate timing picking. This effect was excluded in the simulated ray tracing data. Based on these observations it is concluded that a more accurate, wave equation based forward modeling algorithm is required to obtain accurate inversion results on realistic data.