Infrared thermography as a non destructive technique for the detection of titanium casting defects / Messa a punto di tecniche termografiche per il rilievo di difetti di fusione in getti di titanio commercialmente puro

This paper analyzes the potential application of an experimental quality assessment of pure titanium specimens, produced by investment casting, by means of active infrared thermography (AIT). This technique uses thermal waves to excite the specimen in order to induce temperature differences witnessing the presence of anomalies. Thus, AIT permits the analysis of both surface and subsurface integrity, contrary to other non-destructive testing (NDT) method used in industrial applications, such as penetrant liquid testing. In last decades, the use of pure titanium in biomedical field has been increased noticeably, thanks to its biocompatibility and optimal thermo-physical characteristics. Low thermal conductivity and density in respect to conventional Au and Ni-Cr [1,2] alloys are only two examples. However, Ti casting shows some difficulties due to Ti high melting point (1668 °C) and high chemical reactivity. In particular, problems might result in the presence of internal pits, porosity and discontinuities [3], with defects dimension and number depending on shape, gas permeability [4,5,6], casting geometry and methods [7]. In this work AIT measurements have been performed on commercially pure titanium (Ti CP grade 2) specimens, parallelepiped shaped, with high base areas, in order to obtain a better thermal exchange, resulting in a faster solidification. This is one of priTitanio e leghe La Metallurgia Italiana - n. 11-12/2011 29 mary causes of the presence of internal defects. In particular, AIT measurements have been performed on three different samples (fig. 1): - Sample A: dimensions 35x35 mm2, thickness 5 mm - Sample B: dimensions 20x20 mm2, thickness 2 mm - Sample C: dimensions 20x20 mm2, thickness 2 mm Plank’s law asserts that each body, having a temperature above the absolute zero, emits an infrared wave depending on its temperature and surface conditions (emissivity ε = 0÷1 ). Infrared thermography is based on the detection of this emission, in order to measure temperature, and to generate a false colors thermal map of the analyzed surface (thermogram). Moreover, any thermally excited surface shows back an infrared wave that may be considered as the sum of two components: a reflection component (of excitation source) and an emission one (due to the heating). The quality assessment of a sample is based on the emitted wave, that permits to obtain information about subsurface conditions. Thus, the reflection component induces a noise that may affect results. In order to study only the emission, emissivity becomes a fundamental parameter, because the higher the emissivity, the lower the value of reflected energy. As given in table 1, dark, oxidized, raw surfaces are characterized by high value of ε (0.6 - 0.98), on the contrary shiny or smooth surfaces show low values of ε (0.2 - 0.4) [8]. Active infrared thermography uses an external periodic or impulsive optical excitation on a test sample surface, in order to produce a thermal transient state in the specimen [9,10]. The presence of an internal discontinuity changes the thermal flow conduction in the sample, thus causing a local alteration of surface temperature distribution, during the heating or cooling transitional stage [11,12]. This paper analyzes three different active techniques: - “lock-in-thermography”, - “pulse thermography”, - “pulse phase thermography”. In the “lock-in-thermography” technique [13,14], sample is excited using a sine thermal wave, produced by an optical source (lamp). The thermal flow propagating in the specimen is reflected by internal defects, establishing a wave opposed to the excitation one. The transient thermal state is, afterwards, evaluated by a discrete Fourier analysis (dft) in respect to the excitation frequency, in order to obtain amplitude and phase thermograms. The use of dft makes necessary an exact time synchronization between excitation and acquisition signals. Detection depth is a fundamental parameter depending, in inverse proportion, on the excitation frequency. Using a mono-dimensional thermal transmission model it’s possible to assert that the lower the excitation frequency, the higher the detection depth, the lower the thermal surface resolution [15]. “Pulse thermography” technique uses an impulsive high energy excitation, in order to evaluate the thermal transient in cooling conditions [16,17]. An internal defect appears as a surface zone showing a temperature difference in respect to surrounding areas. “Pulse phase thermography” technique is an evolution of the previous one, and it’s based on the theory that an impulsive wave is ideally composed by a wide number of different frequency waves [14,18,19]. Thus, in this case, it’s possible to carry out a dft analysis at various frequencies, reducing the test time in respect to lock-in technique. Results are amplitude and phase thermograms. Pulse thermography has been led using an oven to heat the specimen, and monitoring the cooling phase. Thermograms elaboration has been carried out by a commercial software. Lock-in and pulse-phase techniques have been performed both in reflection and in transmission mode, using 1 lamp (1 kW power) and 1 flash (1 kW) respectively as excitation sources. Each specimen has been examined also by radiographic and tomographic analysis, in order to compare with thermographic results. Radiographic tests have been conducted by “F” class (Iridium192 isotope, with 0,3- 0,6 MeV energy) equipment [20]; tomographic tests have been carried out by a variable focus station [21]. Pulse thermography results are showed in figure 3. Cooling phase has been monitored for 6 minutes. As can be observed analyzing the thermograms, surface thermal gradients, located in correspondence to internal defects, show a lower temperature respect to bulk areas. Those thermal discontinuities are more visible at the beginning of the cooling process, with higher mean temperature. In particular, the first thermogram (20 s) shows a pronounced gradient, that gradually disappears. A comparison between pulse thermography and radiography is illustrated in figure 4. Resultant thermograms suggest that “pulse” technique doesn’t permit the location of internal discontinuities smaller than few millimeters. Moreover, this minimal detectable defect dimension depends on the deepness of the discontinuities, as well as on the specimen thickness. A tomographic image of the A sample is given in figure 5. According to this, results obtained on the A sample, thicker than others, are not clear. For this reason, this specimen has been tested also with “lock-in” and “pulse phase” techniques, both in reflection and in transmission mode. Results are reported in figures 6-7. In figures 8-9 phase thermograms, obtained at different lock-in or dft frequencies, are showed, respectively for lock-in an pulse phase thermography. Conclusions can be summarized as follows: - thermographic test method permits the identification of internal macro-discontinuities; - pulse thermography technique lead to good results only at the beginning of transient state; - positive results have been obtained with reflection lock-in thermography; - good results have been obtained with pulse phase technique, both in reflection and transmission set-up.