Filtern
Dokumenttyp
- Beitrag zu einem Tagungsband (6)
- Beitrag zu einem Sammelband (5)
- Zeitschriftenartikel (1)
- Buchkapitel (1)
- Vortrag (1)
- Posterpräsentation (1)
Sprache
- Englisch (15) (entfernen)
Referierte Publikation
- nein (15) (entfernen)
Schlagworte
Organisationseinheit der BAM
Eingeladener Vortrag
- nein (1)
The present contribution highlights recent trends in investigating hydrogen diffusion and cracking in steel welds. For such studies, supermartensitic stainless steels (SMSS) have exemplarily been selected. These materials have been used for offshore and marine constructions for about two decades now. They present a versatility of improved properties such as high strength to thickness ratio, good weldability and good corrosion resistance. However, as shown by respective failure cases, SMSS welds might become prone to hydrogen assisted cracking and the degradation phenomena are more easily to investigate due to the nearly homogeneous martensitic microstructure than in other materials. Generally, it has to be distinguished between cracking that occurs during or shortly after fabrication welding, or during subsequent operation of SMSS components. In order to achieve crack avoidance during fabrication and service, conclusive test sequences have to be applied, ranging from field tests at real components and full scale tests investigating the materials behavior under real service conditions to basic and small scale tests, such as tensile and corrosion tests, oriented more towards a materials ranking. Considerable testing of SMSS welds has been carried out and the present paper particularly summarizes spotlights on 1:1 scale component testing of welded tubulars, slow strain rate testing and basic tests oriented to elucidate the hydrogen behavior and degradation in SMSS weld microstructures. Also, permeation tests, hydrogen dependent degradation of mechanical properties and thermal desorption spectroscopy are adressed. As a specific item, first results of lately conducted investigations for tracking hydrogen movement in such weld microstructures by using high energy synchrotron radiation are elucidated.
Duplex stainless steels (DSS) are used in various industrial applications, e.g. in offshore construction as well as in chemical industry. But, at specific conditions, as for instance arc welding fabrication, cathodic protection or exposure to sour service environments, such materials can take up hydrogen which may cause significant property degradation particularly in terms of ductility losses which, in turn, may entail hydrogen assisted cracking (HAC). If HAC occurs in a duplex stainless steel the cracking mechanism is different from steels having only one phase, because hydrogen diffusion, stress-strain distribution and crack propagation are totally different in the austenite or ferrite phase. Whereas, the mechanism of hydrogen assisted crack initiation and propagation as well as hydrogen trapping in DSS have not been fully understood up to the present, as in most two-phase microstructures. In order to better understand the mechanisms of hydrogen assisted cracking in duplex stainless steels, knowledge of the diffusion behavior and of the stress-strain distribution in the ferritic and austenitic phase is of great interest. A numerical mesoscale model was created with a view to studying the hydrogen transport and the stresses and strains in each phase of duplex stainless steel. The material investigated in this work was DSS 1.4462, consisting of approximately equal portions of ferrite and austenite. Hydrogen diffusion in the duplex base metal was studied using the finite element program ANSYS. Stress-strain distribution as well as hydrogen assisted cracking in the ferritic and austenitic phase fractions were additionally investigated. The results of numerical simulation of the hydrogen diffusion process as well as structural analyses enable the identification of crack critical areas in the DSS microstructure. Numerical simulations qualitatively reflect the crack initiation and propagation process in ferrite. Crack critical combinations of hydrogen concentrations and local mechanical loads initiating HAC can be identified.
Offshore wind turbines are an important goal in national energy strategies worldwide. Foundation structures are manufactured from submerged arc welded (SAW) plates with thicknesses up to 200 mm. In that connection, high-strength steels like the S420G2+M are more and more applied offering the possibility for increased stability and load-bearing capacity of the foundations. These offshore steel grades can show a susceptibility for delayed hydrogen assisted cold cracking of the weld joints. For that purpose, a minimum waiting time (MWT) of up to 48 h (dependent on applied standards) is recommended before non-destructive testing is allowed and conducted. But this concept is based on older steel grades that have been used for three or more decades. Nowadays, the metallurgical improvements (clean steels, proper rolling, and heat treatment) of base materials and well as welding consumables must be anticipated. Hence, the MWT concept should be critically discussed as it is assumed to be very conservative. For that reason, the focus of this study was to investigate the diffusion behavior in S420G2+M steel and its multi-layer SAW joint. Electrochemical permeation experiments were carried at room temperature. Boundary conditions were anticipated in terms of using different sample thicknesses. From the experimental data, hydrogen diffusion coefficients and absorbed diffusible hydrogen concentrations had been calculated. It was shown that hydrogen diffusion in the base material is increased compared to the weld metal. In addition, the sample thickness had a significant on the calculated diffusion coefficients. The minimum and maximum diffusion coefficients had been used for numerical modelling of the hydrogen diffusion in the welding joint. It became clear that a MWT must be always regarded together with a critical initial diffusible hydrogen concentration for the evaluation of a possible delayed cracking as diffusion times were mostly > 48 h due to the thick plates.
Duplex stainless steels have been used for a long time in the offshore industry, since they have higher strength than conventional austenitic stainless steels and they exhibit a better ductility as well as an improved corrosion resistance in harsh environments compared to ferritic stainless steels. However, despite these good properties the literature shows some failure cases of duplex stainless steels in which hydrogen plays a crucial role for the cause of the damage. Numerical simulations can give a significant contribution in clarifying the damage mechanisms. Therefore, a numerical model of a duplex stainless steel microstructure was developed enabling simulation of crack initiation and propagation in both phases. The phase specific stress strain analysis revealed that local plastic deformation occurs in both austenite and δ-ferrite already in the macroscopically elastic range. Altogether, phase specific hydrogen-assisted material damage was simulated for the first time taking into account all main factors influencing hydrogen assisted cracking process. The results agree well with experimental observations and thus allow a better insight in the mechanism of hydrogen-assisted material damage.
Modeling of Hydrogen-Assisted Cracking (HAC) in Duplex Stainless Steels (DSS):
- Hydrogen causes significant degradation in DSS and consequently HAC
- Mechanisms of HAC have not been fully understood in two-phase microstructures
- Mesoscale numerical modeling for HAC in DSS was created, coupled to the macroscale, enabling simulation of HAC in both phases independent from predefined crack paths
In the field of modelling hydrogen assisted cracking (HAC) phenomenon, hydrogen diffusivity is an important input parameter for numerical simulation. In terms of hydrogen diffusion coefficients, they have great impact on realistic assessment of the evolution of possible crack critical hydrogen concentrations. In addition, the chemical compositions of steels can have a strong effect on hydrogen diffusion. Unfortunately, literature provides a wide range of available hydrogen diffusion coefficients even for similar microstructures and equal temperatures. The scattering of the data can lead to significant deviations in the results of simulating the evolving hydrogen concentrations due to hydrogen uptake (by fabrication or service). Thus, the application of such data to crack-models or for component life tie predictions can be realized up to the present only by considering envelope curves of such value, corresponding to a work or bench case scenario, respectively. For improved reliability of numerical simulaitons, it is necessary to minimize the mentioned deviation of these data. Hence, this work focuses on the validation of hydrogen diffusion coefficients obtained from permeation experiments at room temperature. Two baintic steels with different alloying concepts were investigated, the creep-resistant 7CrMoVTiB10-10 and the reactor pressure vessel grade 20MnMoNi4-5. A numerical model is presented for simulation of the corresponding hydrogen diffusion during permeation experiments using the finite element software ANSYS. Three different diffusion coefficients (obtained from different common calculation methods) are considered and compared to numerical results. The vases of thes calculation methods are permeation transients which are a direct measure for hydrogen. The results of the simulated hydrogen diffusion coefficients show that only one procedure for calculation of diffusion coefficitnes is suitable in comparision to the experimental values. Thus, it is suggested to use this method for analysis of experimental results in case of hydrogen diffusion during permeation experiments. Furthermore, this work supplies validated values for the hydrogen diffusion coefficients of both steel grades.
A two-dimensional numerical mesoscale model has been created representing a microstructure of a typical 2205/1.4462 duplex stainless steel to further elucidate the mechanisms of hydrogen-assisted crack initiation and propagation in multiphase non-hydride forming metallic microstructures. Hydrogen-assisted cracking (HAC) was simulated by considering different stress and strain behavior as well as different diffusion behavior in both phases. For simulation of crack initiation and propagation, the element elimination technique has been applied. The model allows the simulation of path-free crack propagation which contributes to a better understanding of the HAC process in two-phase microstructures. As a particular result, the analyses revealed that a global macroscopic elastic deformation might already cause plastic deformation in both phases entailing respective HAC.
Numerical investigations on hydrogen-assisted cracking in duplex stainless steel microstructures
(2016)
Duplex stainless steels (DSS) are used in various industrial applications, e.g. in offshore constructions as well as in chemical industry. DSS reach higher strength than commercial austenitic stainless steels at still acceptable ductility. Additionally, they exhibit an improved corrosion resistance against pitting corrosion and corrosion cracking in harsh environments. Nevertheless, at specific conditions, as for instance arc welding, cathodic protection or exposure to sour service environments, such materials can take up hydrogen which may cause significant property degradation particularly in terms of ductility losses which, in turn, may entail hydrogen-assisted cracking (HAC). The cracking mechanism in DSS is different from steels having only a single phase, because hydrogen diffusion, stress-strain distribution and crack propagation are different in the austenite or ferrite phase. Therefore, the mechanism of HAC initiation and propagation as well as hydrogen trapping in DSS have not been fully clarified up to the present, as for most of the two-phase microstructures. At this point the numerical simulation can bridge the gap to a better insight in the cracking mechanism regarding the stress-strain distribution as well as hydrogen distribution between the phases, both austenite and ferrite, of the DSS. For that purpose, a two dimensional numerical mesoscale model was created representing the microstructure of the duplex stainless steel 1.4462, consisting of approximately equal portions of austenite and ferrite. Hydrogen assisted cracking was simulated considering stresses and strains as well as hydrogen concentration in both phases. Regarding the mechanical properties of austenite and ferrite different statements can be found in the literature, dependent on chemical composition and thermal treatment. Thus, various stress-strain curves were applied for austenite and ferrite simulating the HAC process in the DSS microstructure. By using the element elimination technique crack critical areas can be identified in both phases of the DSS regarding the local hydrogen concentration and the local mechanical load. The results clearly show different cracking behavior with varying mechanical properties of austenite and ferrite. Comparison of the results of the numerical simulation to those of experimental investigations on DSS will improve understanding of the HAC process in two phase microstructures.
Since the mid-nineties, supermartensitic stainless steels (SMSS) have increasingly been applied to welded subsea-pipeline systems in the North Sea oil and gas fields, especially to flowlines at mild sour service conditions. However, in 2001 cracking and leaks occurred during installation and service start-up of two SMSS flowlines in the Norwegian Tune gas condensate field, welded with a new developed matching filler wire. Brittle transgranular cracking started especially at inter-run lack of fusion and propagated brittle, predominantly through the weld metal. The present paper provides a brief overview of the original failure case and respective sequence of events leading to complete replacement of the SMSS by carbon steel flowlines in 2002. Then, detailed investigations of a circumferential weld sample of the failed Tune flowline are highlighted, targeted at comparison of the failure appearance to previous investigations of this filler material type and to search for possible explanations for the brittle fracture at the crack initiation area. SEM investigations of the fracture surface revealed brittle areas only in the direction towards the top side of the weld while the major part of the investigated surface exhibited ductile fracture. As an approach to clarify, if the fracture was a consequence of hydrogen assisted cracking, five small sized specimens have been cut out of the original sample. Cracking has been introduced parallel to the original fracture surface in these specimens at respective saw cuts and bending. The results show that brittle transgranular cracking appeared only in the specimen cooled down to very low temperatures by liquid nitrogen and in the sample charged with hydrogen to an average concentration of about 15 ml/100 g. However, a fracture similar to the original surface was observed only in the hydrogenized specimen. As a further result, very similar fracture surfaces of supermartensitic stainless steel weld metals had been observed on specimens subjected to hydrogen assisted cold cracking (HACC) as well as to hydrogen assisted stress corrosion cracking (HASCC). In total, the results indicate that brittle fracture starting at the inter-run lack of fusion were not initiated by high notch tip deformation rates, but rather influenced by hydrogen, probably taken up during welding.
Hydrogen assisted stress corrosion and cold cracking represent still a major topic regarding the safety of welded steel components against failure in many industrial branches. Hydrogen might be introduced during fabrication welding or might be taken up from an environment during sour service or at cathodic protection. Additionally, understanding and avoidance of hydrogen entry into weld microstructures from gaseous pressurized environments becomes increasingly important for renewable energy components. There are two types of metallurgical mechanisms associated with hydrogen assisted cracking, i.e. the cracking as well as hydrogen transport and trapping mechanisms. For numerical modelling, it has to be considered that both types are not independent of each other, that the mechanisms are not yet completely clarified and that validation of such models strongly depends on implementation of the correct hydrogen related materials properties. However, quite significant achievements have been made in modelling of hydrogen assisted cracking by indirect coupling of thermal, stress-strain as well as hydrogen uptake and diffusion analyses. After a brief introduction into the subject and by revisiting various proposed cracking mechanisms, the present contribution focuses on recent developments of a numerical model based on a comparison of actual hydrogen concentrations and mechanical loads with respective hydrogen dependent material properties as crack initiation and propagation criteria. The basic procedure for numerical simulation of crack initiation and propagation is outlined and it is shown how such numerical simulations can be validated experimentally. Furthermore, it is highlighted how such a procedure has been extended to a comprehensive model for life time prediction of welded steel pipeline components and experimentally verified. Finally, it is outlined how the model can be extended to simulate cracking in heterogeneous steel microstructures on the different scales.
Supermartensitic stainless steels (SMSS) are a commonly used material nowadays for building offshore structures, i.e. pipelines in the oil and gas industry. The harsh and corrosive environments in oil and gas applications require the correct combination of alloys to attain the desired properties of steel, including high strength and good corrosion properties, even in severe sour service conditions. Welding is the most commonly used method in joining offshore components, depending on requirements requiring strength or fitting. It has been shown that the heat affected zone (HAZ) is more susceptible to certain types of corrosion, including pitting corrosion, especially during severe sour service where a high pH and lower H2S values in the flow medium can lead to pitting corrosion in the HAZ of welded structures. Subsequent hydrogen uptake in the pits can cause cracks to initiate and propagate, leading to rupture of pipelines or catastrophic failures of structures, even at low mechanical loads. Offshore standards allow a certain amount of corrosion, including pitting, to be present before action is required, however the extent of pitting corrosion is not identified by performing visual inspection alone as the subsurface pit diameter may be vastly greater than the pit diameter at the surface. The critical conditions which lead to crack initiation and propagation from a pit with hydrogen uptake are currently not known. Therefore, pitting corrosion and subsequent crack initiation are a danger to the safety of structures. The interest in this phenomenon has resulted in many experimental studies and numerical simulations.
Several numerical models of pitting corrosion and hydrogen uptake resulting in crack initiation are already in existence, but these two phenomena are regularly modelled individually. Thus, a model enabling simulation of both phenomena simultaneously would be of great benefit. Hence, the goal of this study is to develop a model enabling simulation of pit growth and crack initiation, considering hydrogen uptake in the pit from a corrosive environment.
As a first step, this paper presents an investigation into various parameters, which influence crack initiation at pits. These crack critical parameters include: pit geometry, pit location, mechanical load and hydrogen transport into the microstructure. The results will help to identify critical conditions for crack initiation starting at the pit and developing measures to avoid hydrogen assisted cracking (HAC).
Literature provides a wide range of hydrogen diffusion coefficients for low alloyed steels used in power plants. In fact, experimental boundary conditions and calculation methods have influence on the determination of these coefficients. The diffusion and trapping behavior in creep-resistant steel 7CrMoVTiB10-10 has been studied. Based on experimental carrier gas hot extraction (CGHE) data, a numerical model has been developed to describe the hydrogen transport and respective hydrogen distribution at elevated temperatures. The numerical results suggest that common calculation methods for diffusion coefficients are limited for experimental data analysis. The sample preparation time before CGHE experiment influences the determined diffusion coefficients with the consequence that non-homogeneous hydrogen concentration profiles have to be considered in the simulations.
This study was carried out to investigate the neutron transmission signal as a function of sample temperature during a welding process. A theoretical description that includes the Debye-Waller factor was used to describe the temperature influence on the neutron crosssections. Neutron imaging using a monochromatic beam helps to observe transmission variations related to the material temperature. In-situ neutron imaging of welding experiments show the distribution of the temperature in bulk steel samples. The performed finite element modelling of expected temperature distributions shows good agreement with the obtained experimental data.
In Neutron-Bragg-Edge Imaging (NBEI) experiments, we studied the phase transition during butt-welding of supermartensitic steel plates. Gas tungsten arc welding (GTAW) was used with a motorized torch allowing for automated weldments. The austenitization in the heat affected zone (HAZ) underneath the welding head could be clearly visualized at λ = 0.39 nm, a wavelength smaller than the Bragg edge wavelengths of both austenite and martensite. Also, the re-transformation into the martensitic phase upon cooling was detected. However, we observed an unexpected additional change in transmission at λ = 0.44 nm that is a wavelength larger than the wavelength of the Bragg edges of both the martensitic and austenitic phases. We attribute this change to the Deybe-Waller-Factor that describes the temperature dependence of coherent scattering at a crystal lattice. The observed two-dimensional attenuation map corresponds well with a temperature distribution modelling by software macros in ANSYS. Here, the absolute temperature values could be achieved by calibrating the modelled attenuation with help of a thermocouple placed at the steel plate. This allows in return for a direct two-dimensional temperature reading based on the Debye-Waller-relation between neutron attenuation and sample temperature.