The automatic, intelligent data analysis is an essential part of inspection tasks. Previous projects focused on the development of new functions and techniques for the inspection of technical facilities. An example of use is the metal loss inspection of oil and gas pipelines with passive robots (so called inspection pigs). For this a learning diagnosis system was developed. Based on ultrasonic data it detects and classifies defects like corrosion, dents and laminations.
Our current work is aimed at the consolidation of the developed prototype towards a product fully useable as a matter of routine. Moreover research is done in the field of crack detection for pipelines. In addition to the defect types stated above here also crack- and notch-like defects are to be found. This method is also based on ultrasonic techniques, but with a more complex measurement process.
Another field of work is the integration of the data analysis into the inspection tool (the pig) itself. By this a detailed description of the pipeline will be available directly after receiving the tool from the pipeline.