Modelo de estimación diferencial de la propagación térmica en 3D mediante diferencias finitas

Abstract

In this paper, a new mathematical model based on finite difference discretization of the Fourier’s 3D heat propagation model, is shown. From this one, a new technique is proposed to detect flaws in composite thin layers, under evaluation by a Pulsed Active Thermography experiment. The discrete model defined is easily adaptable to a spatial filter structure, which can be applied to the infrared sequence of images acquired from that thermography experiment, to obtain a better contrast between possible internal flaws and sound regions of material. The performance of the technique proposed is evaluated using artificial thermal sequences generated by ThermoCalc6L, software that is able to compute dynamic thermal distributions in anisotropic layered solids, simulating defects and different excitation sets. Results show that this technique offers a better contrast between defect and background thermal information, than other common technique like Differential Absolute Contrast (DAC), and it runs faster than the classic 3D thermal filtering method. Palabras clave Termografía activa; termografía pulsada; propagación térmica; contraste térmico; filtro espacial. Keywords Active thermography; pulsed thermography; thermal propagation; thermal contrast; spatial filter. 1 Fecha de recepción: Abril 13 de 2012 Fecha de aceptación: Mayo 14 de 2012 Andrés David Restrepo Girón, MSc adareg378@gmail.com Grupo de Investigación en Instrumentación Electrónica (GIE) Universidad Santiago de Cali, Colombia Humberto Loaiza Correa, PhD humberto.loaiza@correounivalle.edu.co Grupo de Investigación en Percepción y Sistemas inteligentes (PSI) Universidad del Valle, Colombia Citación: Restrepo, A. & Loaiza, H.(2012). Modelo de estimación diferencial de la propagación térmica en 3D mediante diferencias finitas. Revista S&T, 10(21), 9-26 Modelo de estimación diferencial de la propagación térmica en 3D mediante diferencias finitas A model for differential estimation of 3D thermal