For thermo-inductive crack detection, a metallic work-piece is placed in a high frequency magnetic field which induces eddy currents in a very thin layer of the surface. This eddy current heats up the sample and the emitted infrared radiation is viewed by an infrared sensitive camera. An inhomogeneous temperature distribution on the surface corresponds to inhomogeneities and cracks in the material. The main goal of the thermo-inductive crack detection is on the one side to find cracks and on the other side to determine their depths. For this purpose an examination of all parameters affecting the result of the measurements has to be made.
In previous publications it has been shown how the thermal quotient Tcrack/Tsurf depends on several parameters (i.e.: time, pulse length, penetration depth of the eddy current and crack depth). All these investigations were made for rectangular shaped cracks. But metallographic cross-sections show that real cracks have different shapes and different angles depending on the circumstances of the origin of the crack. In this paper results of finite element simulations are presented demonstrating what kind of influence the different shapes have to the thermal contrast. It is also shown in which way the crack geometry affects the temperature distribution on the crack near surface. The calculations take into consideration the distribution of the eddy currents around the crack for both magnetic and non-magnetic materials. The simulations are based on coupled modeling of magnetic and thermal phenomena. The calculated results are in very good agreement with the measurements.
In the case of thermo-inductive probing the material is heated by HF-induced eddy currents and the emission from the material surface is detected by an infrared camera. Anomalies in the surface temperature correspond to in-homogeneities in the material. Due to the high excitation frequency (200 kHz) and the magnetic properties of the material, the penetration depth of the current is very small (about 0.03 mm). Therefore the eddy current 'flows around' surface cracks with a depth of 0.1-1 mm. This causes a higher current density and higher temperature around the failures, which are made visible by the infrared camera. Experiments have been carried out on steel wires with a diameter of 4.5-10 mm and with longitudinal surface cracks with a depth of 0.1-0.2 mm. Due to the high heat conductivity of the material, the temperature difference diminishes very quickly. Therefore short heating pulses with duration of 0.1-0.5 sec have been applied. The measurement result shows, that the thermo-inductive method is well suited to detect such shallow flaws. An analytical model has been derived, to calculate the temperature distribution in the wire and around the failure. The model also shows the dependence of the temperature distribution on the parameters of the experiments, as e.g. the length of the heating pulse, which helps to optimize the measurement setup. Additionally, finite element simulations have been carried out. The results of the model-calculations and the simulations are successfully compared with the experimental results.