Thermography for defect detection and structural integrity analysis: comparison of new and established methods and novel procedures for a robust quantitative assessment

Subsurface defects can be detected by pulsed thermography, starting from the evaluation of the temperature response at the surface. However, it is necessary a post-processing of the raw thermal data, in order to provide quantitative information, such as defect depth and size. In the past, many algorithms have been developed to localize defects with a very good contrast and also to estimate their depths such as Thermal Signal Reconstruction (TSR®) and Pulsed Phase Thermography (PPT). There are also different numerical models, one, two or three dimensional, that describe the thermal behavior after a pulsed test of a certain duration that help to have an estimation about the correct thermographic parameter to perform a quantitative analysis. Therefore, a lot of very important parameters, such as the acquisition frequency, the power of the heat sources, the truncation window size in terms of analysis window and also the required geometrical resolution change with the used material, the sample geometry and also the type of expected defect (delamination, crack, porosity, flat bottom hole); for this reason the same have to be selected carefully. Even if many publications have been already dealing with these topics, within this dissertation a new approach is proposed for performing a quantitative and correct analysis after a pulsed test. In particular, a very critical type of material has been studied, such as the aluminum one, because of the major part of the past works regards the composite materials, which are not affected by the problems due to the high diffusivity when a pulsed test is performed. A very large investigation concern the possibility to detect the defects in an aluminum sample with different flat bottom holes, having also very critical aspect ratio values less than 2, by applying different post-processing algorithms such as Principal Component Thermography (PCT), Pulsed Phase Thermography (PPT), Thermal Signal Reconstruction (TSR®), Slope and Square Correlation Coefficient (R2). The influence of the truncation window size, the flash power, and the acquisition frequency has been investigated and also different set-ups, with very different peculiarities, have been used in order to compare the possible achieved results. The new proposed quantitative approach starts from the evidence that exists a linear correlation between the defect aspect ratios and their relative contrasts, shown after the application of different post-processing algorithms and a suitable truncation window size. This experimental evidence has been also explained by simulating the thermal behavior with different well-established models. In this way, the possibility to have an estimation of defect depth and size is however demonstrated by using a very low-cost set-up with very competitive acquisition parameters. The limits of PPT and TSR® algorithms are shown when the aim is to have quantitative information in very high diffusivity materials. An important part of this dissertation regards the application of the pulsed thermography to investigate the thermal behavior of different materials and typical defects involved within several industrial and research applications. In this part of the dissertation, the results in terms of “how to analyze a thermographic data to characterize a defect and quantify it” obtained within the first part, will be used and further studies to solve and investigate components with real and particular types of defects. Finally, new procedures very similar to the one proposed in the first part to quantify imposed defects, will be outlined and described, for evaluating the quality of industrial processes of spread interest, that require not only the identification of the defect, but the control of decisive process parameters . In particular, a lot of pulsed tests, by using different energy sources and IR cameras, and also different thermographic parameters in terms of frame rate and pulse duration have been performed to detect typical defects in a Metal Additive Manufacturing process. This particular type of defects is not simple to detect because they are very small and superficial and also their typical irregular 3D shape involves small air pockets air pockets inside the material. A new and suitable approach has been proposed to control the integrity of a real component by using the thermography as a non-destructive technique. The starting point of this procedure is the comparison between the results reached by using two very different non-destructive techniques, such as the pulsed thermography in terms of Pulsed Phase Thermography and the ultrasonic one in terms of C-scan map with a Phased Array Technology, in the estimation of delamination depth in a composite material. The proposed approach allows for extracting the sound area from the same real component to get information about defect size and depth of a defective part having the same its geometry. Novel procedures for a robust quantitative assessment have been carried out. The first one regards the application of a very simple transmission set-up in order to perform pulsed tests to control the quality of different RSW (Resistance Spot Welding) joints. Different steel joints were obtained from the RSW process by varying the main process parameters such as current and time. By studying the thermal behavior in correspondence of the nugget of these welded joints, it is possible to find different thermographic “indexes”, capable of assessing the quality of joints. A further investigation regards the possibility to use a thermographic method for estimating the coating thickness by using a calibration curve, like the approach shown in the case of an aluminum sample with different flat bottom holes. The investigation regards a coating and a basic material in steel with very close thermophysical properties. For this reason, also in this sac, the limits that shown the classical methods to do a quantitative analysis in this type of application are shown in detail. However, by using a very low-cost equipment, mainly for the used infrared camera, that is a microbolometer one, it will be shown how these limits can be overcome, by developing a new thermographic procedure capable to provide an estimation of the coating thickness from 1 to 10 mm. All the investigated applications are related by a main topic: the possibility and also the necessity to use the thermography as a quantitative control in a lot of industrial and research applications. It is worth to underline that, it is necessary to have a standard as a reference, such as a sample with different imposed defects, or a sample with quality feature well known, or again another type of non destructive technique, in order to calibrate pre and post-processing thermographic parameters before to move on to the analysis of a real component. The variety of the treated real problems such as the detection of typical metal additive manufacturing defects, shows once again as thermography is a versatile and competitive type of non destructive control in terms of time and costs and as a more that promising technique to talk about of “integrity analysis” in the field of engineering.