摘要

Composite materials are being increasingly used in high-tech industries, such as aerospace, automotive manufacture and building inspection. Thermal nondestructive testing (TNDT) has become an accepted method for composite inspection. However, the majority of investigations have dealt with flat or slightly-curved composite components with a thickness of up to 5 mm. Particular studies have been devoted either to NDT modeling with an emphasis on some theoretical issues, or they have been based exclusively on experimental results. There has been some recent interest in the use of composite materials in the nuclear industry. Some critical parts, including centrifuge components, have been made of carbon fiber reinforced polymer (CFRP) composites. The working conditions in a centrifuge include radioactivity and high rotational speed, and the composites used in centrifuges must have very uniform thermal properties and must be free of defects. This paper describes a complex approach to the TNDT of cylindrical parts made of CFRP by starting from thermal properties measurement, theoretical modeling and preliminary experiments, and finishing with the technical requirements for the development of practical equipment capable of operating in both laboratory and industrial conditions. The objects tested were CFRP cylinders with a diameter of 150 mm and a wall thickness of 4-6 mm, and they contained some artificial defects of varying size and depth. Both one- and two-sided test procedures have been analyzed for spot, line and uniform heating. Ultrasonic excitation has also been used as an alternative stimulation technique. In a one-sided test, the depth detection limit has been about 4 mm. Similar results have been observed in the case of ultrasonic stimulation, but the practical implementation of ultrasonic IR thermography to the inspection of cylindrical parts requires further exploration. In a two-sided test, even fairly mild heating resulted in the reliable detection of all defects independent of their size and depth. In all test cases, the highest signal-to-noise ratio occurred after applying the technique of principal component analysis.