Filtern
Erscheinungsjahr
Dokumenttyp
- Zeitschriftenartikel (23) (entfernen)
Schlagworte
- Non-metallic inclusions (4)
- Fatigue (2)
- Fatigue crack (2)
- Pores (2)
- Safe life design (2)
- Schadensanalyse (2)
- Undamped structural elements (2)
- Vibrationsprüfung (2)
- Accompanying measures (1)
- Austenitic steel (1)
Organisationseinheit der BAM
- 9 Komponentensicherheit (5)
- 9.0 Abteilungsleitung und andere (4)
- 9.4 Integrität von Schweißverbindungen (4)
- 5 Werkstofftechnik (3)
- 5.1 Mikrostruktur Design und Degradation (3)
- 7 Bauwerkssicherheit (1)
- 7.6 Korrosion und Korrosionsschutz (1)
- 8 Zerstörungsfreie Prüfung (1)
- 8.1 Sensorik, mess- und prüftechnische Verfahren (1)
The 100th anniversary of the death of Adolf Martens will be commemorated on July 24th, 2014. He is eponymously remembered today through the term martensite, which was first used by Floris Osmond as a name for the metastable phase that results from rapid quenching of carbon steels. Born in 1850 near to Hagenow in the region Mecklenburg, Germany, Martens was one of the pioneers of materials engineering in 19th century Europe.
Martens began his career working for the Prussian Eastern Railway before joining the Royal Industrial Academy in Berlin in 1880. In 1884, he was appointed director of the Royal Mechanical Experimental Station, a small institution associated to the academy. Failure analysis was continuously practiced at this institution, which became later the nucleus of the German Federal Institute for Materials Research and Testing (BAM), for the 110 years since. The history of Martens will be dealt with in an in-depth article in this special issue.
Since the 19th century, failure analysis techniques have been refined, and new methods of chemical analysis and non-destructive testing have been developed; however, the basic approach to failure analysis has not changed much since Martens' days. The basic tenets of failure analysis remain things like on-site inspection, extensive visual 'non-destructive' inspection, developing an understanding of the background story, performing materials testing, and 'connecting the dots.' Martens introduced and developed experimental techniques like macro photography, fractography, metallography, hardness measurements, and mechanical testing. Modern failure analysts continue adding even more techniques to this list, leading to a more interdisciplinary approach which many would say is the only way to find the root causes of complex failure events.
The present special issue of EFA presents an overview of more than 100 years of failure analysis at BAM and its predecessors, closing the circle from the beginnings of modern failure analysis done by Martens himself in the 1890s to its present-day application.
This special issue starts with an excursion back to Martens' work and innovations and presents a newly translated original manuscript of Martens from 1890. Whereas some papers of Martens and his co-workers are well documented, only little can be found about failure analysis in the period from 1914 to the 1950s. Most documents of this period have not survived until today. Beginning in the 1960s more and more significant works are preserved, which were using the interdisciplinary approach of Martens. Since the beginning of the digital age in the 1980s almost all text documents are accessible, whereas digital images were stored since the 1990s. Since then the problem is no longer accessibility but copyright issues that prevent many interesting case studies from being published. Maybe the next generation of failure scientists can reveal some of them later on.
Investigations on the breakdown of a heat recovery steam generator during the initial operation run
(2014)
Several years ago, in 2003, an industrial heat recovery steam generator in charge of generating process steam in a petrochemical refinery was installed and prepared for initial operation. The steam generator enclosed an evaporator section and a superheater section consisting mainly of bundles of tubes with the longest up to several metres in length. During initial operation test runs severe leakages in the evaporator and superheater modules became noticeable. The test runs were stopped and after disassembly, through-wall cracks in several tube bends were found. BAM was commissioned to carry out the investigations in order to find out the reasons for the failure of the tubes. During on-site inspection a number of relevant damaged components and parts were selected and taken away to the laboratories for detailed inspection. Planned analyses were to comprise metallographic as well as fractographic investigations, mechanical fatigue testing and experimental as well as finite element vibration analyses on specimens and components. Soon, the fracture mechanism was found to be mechanical fatigue due to the fact that the examined fracture surfaces showed the very characteristic beachmarks and colouring patterns. To identify the particular loading and time at which crack initiation and crack propagation took place, experimental and numerical vibration analyses of specific tubes as well as mechanical fatigue tests on tube bends were carried out. Thus it was possible to identify the eigenfrequencies of individual tubes, to estimate the dynamic response as well as the nominal stresses and, hence, experimentally characterise the in-service fatigue strength of the components. Mechanical and thermal comparative tests on tube bends were performed simulating the conditions during the initial test run in order to get crack surfaces comparable to that of the originally damaged components. Thereby it became obvious that the fatigue cracks were initiated by vibrations the tube bundles were exposed to during rail transport from the manufacturers site to the place of installation. Based on these results, the damaged components could be repaired or exchanged without modification of the construction, but the rules relating packaging and securing for shipping had to be revised.
The roll-on-roll-off passenger ferry MV ESTONIA sank during the night of 28 September 1994 in the Baltic Sea. In October 2000, divers recovered two palm-sized test pieces from the front bulkhead of the wreck. The investigators analysed these specimens to determine whether there were any indications of deliberate blasting. Since the wreck had been submerged for almost six years, it was clear that chemical traces would not be present on the surfaces of the test pieces. Therefore, the investigators performed comparative tests on shipbuilding steel to find a microstructural criterion that exclusively characterises a blast. The shipyard Jos. L. Meyer, Germany, had built the ESTONIA and supplied shipbuilding steel plates similar to that used for building the vessel in 1979/1980. The comparative tests comprised mechanical tests, shot peening tests and blasting tests using different explosives. Testing demonstrated that blasting always formed twinned ferrite grains in the microstructure over the whole cross-section of each of the 8 mm thick comparative plates. Although one of the original test pieces of the ESTONIA showed deformation twins, this was only confined up to 0.4 mm underneath the surfaces and not spread over the whole cross-section. Comparative shot peening tests produced the very same pattern of subsurface deformation twins. Therefore, the twins detected in the microsection of the test pieces of the ESTONIA wreck traced back to the shot peening process performed by the shipyard in 1979/1980 and not to a deliberate blast.
Self-excited vibrations with large amplitudes in natural wind may occur at slender structural elements with low damping. Because of the different designs (e. g. using solid sections today instead of cables for tension elements in the past) the susceptibility to wind induced oscillations has increased. Those wind induced vibrations of profiles with specific cross section geometry which are motion induced and therefore self-exciting are called 'galloping vibrations'. Especially systems with elements that are highly tensile loaded and undamped, like hangers of bridges or tension bars of cranes, are sensitive to wind induced vibrations. Therefore more and more fatigue problems caused by galloping oscillations have occurred in the 1990s. This paper describes exemplary the collapses of two modern cranes of different design and manufacturers. During standstill periods, both cranes suffered from wind induced vibrations of the tension bars, which bear up the counterweights. The failure analysis process to identify and explain the fatigue fractures as well as the comparative experiments and simulation to verify that they were caused by wind induced galloping-vibrations is described. It is shown, which parameters led to galloping-vibrations of the tension bars and how their onset wind speed and the amplitudes can be estimated with more accuracy by a non-linear and non-stationary approach. Furthermore it is shown that such dynamic stresses caused fatigue failure of the tension bars for the counter weights and subsequently collapsing of the cranes.
For loss prevention knowledge and results gained by these investigations should be put at disposal to engineers working on this field of design. In the meantime, a contribution to development of appropriate technical standards on structural steelwork was given by the research works on galloping. Although new standards were introduced, which consider wind induced vibrations, such failures still occur. (Reference to the paper 'Fatigue crack in railway bridge hanger due to wind induced vibrations – failure analysis, measures and remaining service life estimation' in this same Special Issue 'A tribute to A. Martens' 2014).
In July 2008 an ICE3 high speed train rated for 330 km/h service speed derailed during departure from Cologne, Central Station, Germany, due to fatigue failure of one of the driving axles. The train was emergency stopped immediately and, due to low travel speed at this point, no serious injuries occurred to passengers. Referring to public interest, the public attorneys office solicited the German Federal Institute for Materials Research and Testing (BAM) for the analysis of the root cause.
No deviations from specification were found in the geometries of the basic parts of the bogie or the wheelset assembly. Inspection of the axle fragments using standard acoustic non-destructive testing (NDT) techniques revealed no additional cracks and no indications of oversized discontinuities. Metallographic and chemical inspection of the axle material and its microstructure revealed all parameters to be acceptable except for an elevated impurity level.
The fracture surfaces of the axle fragments were heavily damaged due to some continued travel after final breakage on the high speed line before Cologne Central Station. Extensive visual inspection of the remaining beachmarks was carried out to find the origin of the fatigue crack. The region of the crack origin was located near the axle surface but could not be analysed in detail due to secondary damage. Fatigue was identified as the mechanism of crack growth until final fracture, but the reasons for crack initiation initially remained unclear.
Neither standard NDT techniques nor metallography according to the relevant axle specifications were able to identify inclusions in the material that could have served as crack initiation sites. However, discontinuities were detected near the crack origin in micro computer tomography and ultrasonic immersion testing. Subsequent metallographic sample preparation was targeted to specific areas based on the location coordinates of the flaws identified by these NDT techniques. These revealed non-metallic inclusions that were much larger than admissible for the relevant specifications. It is likely that the fatigue crack in the highly loaded axle volume initiated at those non-metallic inclusions.
Shortly after midnight on 15th November 1992, a severe train crash was caused by the break off of a buffer from a tool and gear wagon of a freight train. The buffer fell between the rails, tumbled, and then lifted one axle of a following freight wagon off the rails. Several freight wagons derailed in the station of Northeim, Germany, blocking the adjacent track. A night train travelling on that blocked track was already too close (only 180 m) to stop, despite emergency braking initiated immediately and automatically by the track signalling system. When the night train crashed into the freight wagons, some of its coaches were heavily damaged. Eleven people were killed, and 51 injured, some of them seriously.
The task for failure analysts was to find out why the buffer broke off from the front frame of the tool and gear wagon. Therefore, the material of the attachment, the front frame, and the welds were tested. Chemical composition, hardness, and microstructure were allowable. The welds showed imperfections which were determined not to be causal. The fracture surfaces had graded corrosion which again was not causal and beach marks that depicted fatigue cracks before final fracture.
The cause for the break off was the fact that the front frame had been repaired at this same buffer attachment some years before and that the repair patch used was thinner than the original front frame profile.
Additionally, the force characteristics of the buffer showed reduced stroke but increased forces which was due to several friction springs that had been broken for a long time. The buffer forces in service were assessed from standards, literature, and the deformations of the attachment. The in-service stresses at the welds were estimated using FEM analysis. High cyclic stresses at the welds resulted in fatigue cracking and finally break off of the attachment with the buffer.
An einer Eisenbahnbrücke über die Elbe bei Lutherstadt Wittenberg führten windinduzierte Schwingungen zu einer unplanmäßigen und frühzeitigen Ermüdung ausgewählter Bauteile und Verbindungen des Stahltragwerks.
Für die zyklisch bereits vorbelasteten und noch im Bauwerk befindlichen Bauteile konnte rechnerisch eine ausreichende Restnutzungsdauer nicht ermittelt werden. Auf Basis von experimentellen Untersuchungen wurde daher ein ingenieurmäßiges und baustellengeeignetes Sanierungskonzept für die vorbelasteten Bauteile entwickelt, deren Hauptbestandteil das Beschleifen der betroffenen Oberflächen darstellt.
Weiterhin wurden für die ermüdeten Schweißverbindungen bruchmechanische Berechnungen angestellt, um auch deren Restnutzungsdauer abschätzen zu können. Hierfür wurden den im Bereich der Deutschen Bahn AG geltenden Rechenansätzen vergleichend weitere Ansätze aus dem Maschinenbau gegenübergestellt.
Durch die angestellten Berechnungen und Untersuchungen konnte nachgewiesen werden, dass für die betroffenen Bauteile nach deren Sanierung ausreichende Tragsicherheit und Ermüdungssicherheit vorschriftengemäß eingehalten werden.
End of November 2005 strong south-west wind and heavy snowfall were predominant in the region Münsterland, north-western part of Germany. This led to accretion of a considerable quantity of wet snow to overhead electrical lines in form of snow rolls on the conductors. Eighty-two transmission towers failed catastrophically, most of them by buckling, however some by brittle fracture. As a consequence nearly 250,000 people have been cut off from electrical power supply for several days with major media attention.
This paper describes the forensic analysis in order to investigate the failure cause. Therefore extensive materials investigations, mechanical testing of original components and specimens thereof, estimations for the real wind and snow loads and their combinations, structural analyses as well as detailed evaluations on the basis of previous investigations, literature and regulations were conducted. It was revealed that some of the examined components were manufactured from Thomas steel which was partially in embrittled condition. The investigated towers fulfilled the design codes valid at the time of erection. However the present line loads of the wet snow rolls on the conductors exceeded by far the ones given in the design codes valid at that time.
The load case leading to failure was reconstructed by the derived positions of loads mainly caused by unequal and asymmetric distribution of snow rolls on left and right electrical system. The loads and corresponding stresses acting on the structure before failure were estimated. By comparison with the fracture forces from mechanical testing of original members of the collapsed tower the component that primarily failed was localised. The primary fracture occurred on a diagonal member under tension made of Thomas steel which was weakened by embrittlement. The failure cause was a combination of heavy weather conditions (storm, approx. 0 °C and wet snowfall leading to heavy snow rolls on conductors), asymmetric loading conditions and the usage of Thomas steel which was partially embrittled. Finally, recommendations for avoiding future failures are given.