Influence of the real energy input on the detection limit of thermographic testing (TT) in case of GFRP

Speaker:
Müller, J.P.; Federal Institute for Materials Research and Testing (BAM); Germany

Authors:
Müller, J.P.; Federal Institute for Materials Research and Testing (BAM); Germany
Krankenhagen, R.; Federal Institute for Materials Research and Testing (BAM); Germany

ID: ECNDT-0247-2018
Download: PDF
Session: Thermography and Thermosonics 1
Room: H1
Date: 2018-06-14
Time: 09:00 - 09:20

Thermographic testing (TT) is an upcoming nondestructive method, which requires no contact at all to the specimen and can be applied on larger areas simultaneously. The measurement concept is based on the production of a thermal imbalance at the surface of the object under test. When the surface of this object is heated by an external source for a certain time, the surface temperature drops subsequently, influenced by inner defects of the sample. This leads to thermal contrasts at the surface. It is crucial that those contrasts are large enough to be detectable above the noise level. In a first approximation, the observed temperature contrast at a defect is proportional to the energy which was really introduced into the specimen during the heating period [1]. However, the real energy input in a TT experiment is almost always unknown due to distinct parameters of the experimental setup or the material investigated. Typically, only the power consumption of the heating sources is reported, sometimes combined with the distance to the specimen surface.
This contribution describes the thermographic inspection of a rear side thickness variation from 1 to 2 cm at GFRP. This could represent a rear side adhesive bond i.e. in a wind turbine rotor blade. The front side heating was realized by usual halogen lamps with variable radiation power. The detected temperature contrast at the front side will be related to the different energy inputs determined by means of a simple analytical model applied to the experimental data. Additionally, the experimental data are compared with results of 1D simulations.
The results clearly demonstrate the key role of the real energy input in a real TT setup, if detection limits have to be evaluated.

[1] V. Vavilov, NDT&E International 61 (2014), 16-23