engleski

Pulse thermography applied on a complex structure sample: comparison and analysis of numerical and experimental results

The purpose of this work was to determine the possibility of modeling complex structure samples inspected by infrared thermography, as well as the possibility of identifying the defect characteristics, such as the defect type, using the results obtained. The paper presents the analysis of results obtained by pulse thermography experiments on a complex structure sample containing defects. The sample consists of two different honeycomb panels as a core between two graphite-epoxy layers joined with appropriate specified adhesive. Each of the two sample parts contains the same number of defects of different type, size and depth. A pulse thermography experiment was conducted and the results were recorded by high resolution infrared camera as presented in the article. The images obtained were processed so as to enhance the defect visibility. Once the location of the defects was determined, surface temperature evolution curves were extracted above each of the defects as well as in non-defective regions. These curves were used afterwards for numerical model evaluation. As mentioned, the numerical model of the tested sample was made in order to simulate the behavior of the sample under inspection. COMSOL software based on Finite Element Method (FEM) was used to make a 3D model of transient heat transfer. In a similar way as for the results obtained experimentally, the surface temperature evolution curves were extracted from the results of the simulation. Since many of the model parameters were not precisely known (power density of the heat source used in experiment, convective heat transfer coefficients, sample surface emissivity...), they were adjusted to obtain the results of numerical simulation as close as possible to those obtained experimentally. Nevertheless, all of the parameter values determined in such a way were within the range of the expected values. In general, the behavior of the temperature evolution curves obtained numerically fit the experimental results. Some difficulties in comparison aroused as the nonuniform heating was present in the experiment while the model assumed a uniform heating lamp power density distribution over the whole sample surface. This required a careful choice of the reference defect-free regions that was made for each of the defects separately in the recorded thermograms. A difference in behavior obtained experimentally and numerically was observed for only one of the defects. The comparisons performed in this paper enable us to have confidence in the developed model and thus to evaluate, through modelling, the potential performance of infrared thermography on similar specimens.

NDT; Infrared Thermography; Pulse Thermography; Finite Element Method;

nije evidentirano