Wissenschaftliche Artikel der BAM
Classical film radiography is a well-established NDT technique and it is most commonly used for testing weld seams and corroded pipes e.g. in oil and gas industry or in nuclear power plants. In the course of digitization, digital detector arrays (DDA) are finding their way into industrial applications and are replacing film radiography step by step. This study deals with the latest generation of DDAs, the photon counting and energy resolving detectors (PCD), and their characteristics compared to charge integrating detectors (CID). No matter which technology to use, radiography still lacks a general issue: A three-dimensional object is projected onto a two dimensional image. Of course, advanced computed tomography (CT) algorithms exist since many years, but if the object to investigate is too large to fit into the manipulation system or its shape is not appropriate, CT is not feasible or sensible to be applied. To overcome this limitation, numerous laminographic algorithms have been developed in the past. In this study, photon counting detectors are used in combination with co-planar translational laminography to gain reconstructed three-dimensional volumes. Both laminographic testing and PCDs require a serious knowledge of many parameters that can influence the image quality in the resulting datasets. These are e.g. the detector efficiency and calibration procedure, setting of energy thresholds, exposure data, number of projections, beam length correction and spatial resolution. The use of PCDs yields more variables to be considered compared to CIDs. The most important parameters in laminographic testing and in the use of PCDs are described in this study and limits are discussed.
X5CrNiCuNb16-4 has been proven to be sufficient resistant in corrosive environments, but shows rather unusual corrosion behaviour in CCS environment. Therefore a series of 30 specimens was tested at stress amplitudes between 150 MPa and 500 MPa (sinusoidal dynamic test loads, R=-1; resonant frequency ∼ 30 Hz). Due to the rather heterogeneous fine machined surfaces (Rz=4) the specimens are comparable with prefabricated parts. X5CrNiCuNb16-4 reached the maximum number of cycles (10 x 106) at a stress amplitude of 150 MPa and lies 60% below the stress amplitude measured in air. The scatter range TN = 1:34 is disproportionately large. Although the fracture surface exhibited the typical striations and corroded surface areas no significant differences were found. The hardness was found to be homogeneous in all specimens tested at 335 HV10. Non-metallic inclusions were found within the microstructure, but no correlation could be found between the inclusions and early rupture. Still specimens that showed inclusions at the fracture surface and its cross section endured lower number of cycles. Additionally Aluminium was analysed in specimens with low number of cycles and may be cause for early rupture during corrosion fatigue tests. These findings reveal a very high sensitivity on a homogeneous microstructure upon the corrosion and corrosion fatigue behaviour of X5CrNiCuNb16-4 and needs to be taken into account when regarding this steel as pipe steel during injection of CO2 into saline aquifers.
The vast experience with the automated, ultrasonic system for the inspection of hollow railway axles used by Deutsche Bahn shows that much smaller flaws are detectable than required. This results in a number of false calls. False calls lead to unnecessary demounting and disassembling of wheelsets, which generates unnecessary additional costs. In order to adjust the sensitivity of the inspection system to reduce the number of false calls without compromising safety, the capability of the system to detect cracks needs to be comprehensively established. This capability can be quantified by using probability of detection (POD) curves for the system. The multi-parameter POD model makes it possible to include several factors that influence the crack detection in the analysis. The analysis presented in this paper shows that crack position, orientation, depth extension, and shape as well as the geometry of the axle all have influence on the ultrasonic response amplitude. For future work, calculation of the POD using multi-parameter POD model with these parameters is planned.
Dass die Zuverlässigkeit der ZfP eine sehr wichtige Rolle bei der Bewertung sicherheitsrelevanter Systeme spielt, ist bekannt. Laut dem modularen Zuverlässigkeitsmodell für zerstörungsfreie Prüfprozesse hängt die Zuverlässigkeit von der innewohnenden (physikalisch-technischen) Fähigkeit der Prüfsystems, den Anwendungsparametern, den menschlichen Faktoren und dem organisatorischen Kontext ab. Seit Einführung des Modells während des ersten Workshops zur Zuverlässigkeit der ZfP (European-American Workshop on Reliability of NDE) wurde das Modell in weiteren fünf Workshops stetig weiterentwickelt, diskutiert und an den Stand der Wissenschaft angepasst. Das Ziel dieser Publikation ist es, den aktuellen Stand der Diskussion und die neusten Fragestellungen der internationalen Zuverlässigkeitsgemeinschaft vorzustellen. Im Einzelnen werden die Entwicklung fortgeschrittener Modelle zur Bewertung der Probability of Detection (POD), Bewertungsansätze im Bereich Structural Health Monitoring und Fragestellungen zu den menschlichen Faktoren dargestellt. Besonderer Wert wurde sowohl auf die Entwicklung, als auch auf die Anwendung der Ansätze in der industriellen Praxis gelegt.