21.08: Evaluation of steel buildings by means of non‐destructive testing methods

Abstract

ABSTRACTNon‐destructive testing (NDT) methods have become increasingly attractive for verification of existing steel infrastructure including buildings. However destructive testing is still state of the art combined with load condition monitoring and yield limit determination of the steel elements incorporated in buildings.In many of the existing buildings, the characteristics of the steels used are often not known anymore. A statement of the residual load‐bearing capacity such as of incorporated steel beams within a building can be made if there is a means to determine it reliably. This opportunity principally exists via micromagnetic as well as ultrasonic measurements to be shown here. Magnetism even allows the yield strength of ferromagnetic steel beams to be determined. An ultimate limit state design can be proven by real material characteristics where model uncertainties have been excluded and not by selected minimum values only.If the residual load‐bearing capacity of steel beams cannot be guaranteed today, higher static loads can only be transferred by strengthening measures, such as welding or bolting additional parts. However, if the true loading as well as material condition could be determined those measures might be avoidable. A means on how to do this is by ultrasonic measurement since the propagation speed of ultrasonic waves in a steel is influenced by its strain and stress conditions. Acousto‐elastic characteristics show the correlation between the propagation of speed and the current state of stress. Those characteristics have therefore been determined on beams in lab tests and compared to strain gauge measurements, where the acoustic wave speeds have been determined first as a speed without the influence of load, before loaded conditions were considered.Material texture due to rolling processes plays a significant role in steel structures too. It has a ten times higher influence on the ultrasonic measurements than stresses due to loading have. Hence a clear separation between texture and stress related influences is therefore essential. Particularly since all those measurement results may further be exposed to significant scatter, statements regarding confidence bands of the determined stresses have to be made. This therefore includes statistical tests that will provide the respective reliability required for the concept to be competitive. The respective NDT techniques combined with suitable stochastic methods are shown below.