Thermal-based damage detection in porous materials

Abstract

We report here on the use of the heat equation to simulate a thermal interrogation method for detecting damage in a heterogeneous porous material. We first use probability schemes to randomly generate pores in a sample material; then we simulate flash heating of the compartment along one of its boundaries. Temperature data along the source and back boundaries are recorded and then analyzed to distinguish differences between the undamaged and damaged materials. These results suggest that it is possible to detect damage of a certain size within a porous medium using thermal interrogation.

Modeling the flash-heat experiment on porous domains

“…We first recall the system (1) which corresponds to the physical flash-heat experiment described in [1]. This experiment assumes that the temperature of the specimen is within the solid state phase and the boundaries are perfectly insulated [13].…”

“…This porosity causes a certain amount of noise in the interrogation signal of any NDE technique. In [1], where we investigated using active thermography to detect damage in porous domains, we found that we could detect damage of a certain size within a porous medium using thermal interrogation. However, the methods developed in [1] were too computationally intensive to use in the sophisticated 54 H. T. BANKS, D. CIORANESCU, A. K. CRINER, AND W. P. WINFREE parameter estimation routines needed to characterize damage.…”

“…In [1], where we investigated using active thermography to detect damage in porous domains, we found that we could detect damage of a certain size within a porous medium using thermal interrogation. However, the methods developed in [1] were too computationally intensive to use in the sophisticated 54 H. T. BANKS, D. CIORANESCU, A. K. CRINER, AND W. P. WINFREE parameter estimation routines needed to characterize damage. As in [1] we treat here the problem of modeling the flow of heat in a porous domain but in the current work we focus on the techniques of homogenization theory (which are less computationally intensive than the methods used in [1]) to capture the behavior.…”

“…Now that we have considered a model of the flash-heat experiment on a homogeneous domain, we will model the flash-heat experiment on more complicated domains than the rectangle Ω. We developed a method to model the flash-heat experiment on porous domains in [1]. Though these simulations were useful in detecting damage, they were too computationally intensive for the more sophisticated parameter estimation procedures needed to characterize damage.…”

“…We will use (8) to simulate data that one might expect from the flash-heat experiment performed on porous specimens. We generate these geometries and solve the partial differential equations using the methods developed in [1].…”