PSI - Issue 79

Andrea Avanzini et al. / Procedia Structural Integrity 79 (2026) 88–96

91

2.3. Thermal acquisition and data processing For thermal acquisition, a Testo 890 thermal camera equipped with 25° optics was employed. This camera uses an FPA sensor with resolution of 640x480 pixels, thermal sensitivity lower than 40 mK, and acquisition frame rate of approximately 33 Hz. The selected optics have a minimum focal distance of 20 cm, which was adopted to maximize the camera's performance and achieve optimal spatial resolution, ensuring the highest possible pixel density over the specimen. Prior to testing the specimens were painted with a black spray of known emissivity (0.95). A painted but unloaded (dumb) sample of the same material was placed close to the loaded specimen to serve as a reference for environmental condition monitoring, allowing correction of measured data in the post-processing phase. An in-house custom software was developed in LabVIEW environment, with a dedicated interface to manage acquisition and storage of thermal images during the various stages of the tests. An example of the acquisition is provided in Fig. 2a for AN condition, including temperature maps for the rupture frame (Fig. 2b). The software allowed the definition of multiple regions of interest (ROI) for simultaneous thermal acquisition of tested specimen (gauge region) and dumb specimen (reference). Within the ROI it is possible to automatically track the maximum, minimum and average value during the test. Another series of scripts were implemented using MATLAB to automatize extraction of the desired thermal indexes from the ∆ T-N curves and their processing. Different methods, such as the one-curve (OCM) or two curve (TCM) have been proposed in literature (Luong, 1998) (La Rosa, 2000) (Cura et al., 2005), as well as different algorithms, to estimate the stress level associated with fatigue, such as iterative methods and minimization fitting procedures. For comparison purposes, these were all implemented in the scripts according to the workflows schematically described in Fig. 2c.

Fig. 2. (a) Example of temperature distribution (b) Fracture frame for AN (x and y axes indicate pixel number, and colors represent the temperature difference in °C over ambient), (c) Worflow for Two-Curves methods.