In recent years, the industrial implementation of composite materials has spread to a wide range of manufacturing industries. The automotive industry has also recognized the advantages of lightweight engineering, classifying it as a key technology for the manufacture of vehicles of tomorrow. Composite materials are not only lightweight and flexible but also able to withstand high stress levels and maintain their stability despite fluctuations in temperature. For this reason, they are increasingly being used in small-series productions. To enable the widespread use of these materials, for example in the mass production of complete carbon fiber car bodies, the inspection processes implemented need to be validated. These must be affordable, non-destructive methods that are capable of comprehensively characterizing materials and identifying any defects.
The department “Machine Vision and Signal Processing” has developed a thermographic system for evaluating carbon-fiber reinforced plastics (CFRP) and glass-fiber reinforced plastics (GFRP). The system has been constructed on a modular basis. Different IR cameras (e.g. Velox 65kM Pro from IRCam) and excitation units (e.g. flash light) can be mounted regardless of the geometry of the component requiring inspection. The analysis method is based on the principle of lock-in thermography. Hidden structures or defects alter the course of temperature distribution over time. This enables them to be detected by means of phase images generated by the sensitive lock-in thermography system.
The results of the thermographic assessments are analyzed with the aid of diverse in-house image-processing algorithms to automatically obtain IO/NIO information. Any areas containing defects are shown in color in the resulting image.
The system developed at IPA is suitable for rapid analysis and supplies sound qualitative information. There is also the option of evaluating the ROI on further sensory analyses (e.g. CT), thus reducing the amount of time and money required for quality control.
In principle, the following classes of defect can be fully tested with automated IO/NIO analysis: