This article addresses the study of crack behavior elicited on axial fatigue in specimens joined by butt weld made of steel ASTM A36 by using fracture mechanics and simulation software of finite elements (Ansys APDL, Franc3D). The computational model was initially to define the geometry model by using CAD software. Specimens with Weld Reinforcement of 2 mm and 3mm were simulated. Subsequently, the type of element for the mesh, the information inclusion concerning material mechanical properties and load conditions were selected. By using Franc3D software, the crack propagation phenomenon is analyzed, and its growth parameters have been established. In this way, it is possible to calculate the magnitude of stress intensity factor (SIF) along the crack front. It is concluded that the stress located in the weld toe is maximized proportionately to the size of the weld reinforcement due to the concentration effect of geometric stress. In addition, it is observed that the propagation rate obtained from Paris law has a similar behavior for the studied weld reinforcements; the latter as there were short cracks.
Fatigue Crack Propagation Ppt Viewer
The integrity of welded joints has always been a concern in the engineering field. Most of these joints are found in elements subjected to cyclical loads, thus the incidence of flaws such as cracks are inevitable. These cracks emerge because of different factors among which changes in microstructures in steels are highlighted, inclusions that can generate stress concentration or simply operation conditions. This phenomenon behavior is studied by Fracture Mechanics, which is defined as the study of crack propagation in an elastic material through determination of critical conditions (for instance, magnitude and type of load and defect size) for which their growth is produced [1].
Regarding the analysis of fracture mechanics of welded joints, several authors have researched the issue in the last years through the selection of a combination of factors, for example, material, geometry of the model, and load conditions. However, most have chosen the use of computational tools and the finite elements method (FEM). For instance, Berrios [2] has found a prominent correspondence among experimental trajectories of crack propagation in high-strength steel and trajectories of propagation obtained by using ANSYS APDL software.
However, there are some international standards that must be considered in the design of metal structures (where there are welded joints). Standards such as BS 7910 [3] that establish the failure criteria, as, fracture, fatigue, creep and other types of failure (corrosion and buckling), as well as to replace traditional standards in welded joints. This allows it for greater discontinuities in areas subjected to small stress which saves resources without compromising the safety of the structure. From another perspective, Hobbacher [4] establishes a basis for the design and analysis of welded joints whose stress can be considered fluctuating to avoid fatigue failures. The evaluation procedures determine the information related to the action and fatigue resistance. However, these methods depend, in turn, on available data of the welded joint and the forces involved. On the other hand, Newman [5] has proposed a series of empirical equations of stress intensity factors for elliptical, semi-elliptical cracks, among others, which are embedded in a finite body subjected to axial loads, They are especially useful when crack propagation rates are to be analyzed and the calculation of fracture toughness in the types of faults mentioned above. Likewise, Bowness [6] has developed a mathematical model for the determination of the magnifying factors (Mk) in the welding foot for semi-elliptical cracks in welded joints in T. The author concluded that the equations described should not be applied to the calculation of K deep in the tip of the crack due to a singularity fault of r.
Within the same research field, Lewandowski [12] has analyzed crack growth behavior in ferrite-perlite structures subjected to cyclic bending and has determined that, in all cases, life in fatigue of welded specimens is lower than those fully solid due to dissimilar mechanical properties of characteristic zones of welded joints (base, bead, and ZAT). In addition, Zerbst [13] has researched fracture mechanics implementation in the establishment of fatigue strength in welded joints whose cracks were originated in the weld toe. He has found that the mechanical properties and S355NL and S960QL steels in the heat-affected zones differ in the correspondence between simulations and experiments. This fundamentally occurs because of higher-strength steel weld toe deformation is still elastic for applied higher stress.
A new finite elements method with interface elements has been developed by Serizawa [14] with the purpose of examining microstructural fracture behavior and where the anisotropy of grain was modeled by ordinary finite elements. It has been found that anisotropic mechanical properties of grain boundary are a prevailing factor in the fracture process. Following this line of work, Salcedo-Mora [15] has exposed a meshfree microstructural elements method to analyze the microstructure effect in quasi-fragile properties within numerical simulations of damage, improving computational accuracy and cost in engineering applications; the researcher has demonstrated a method for released-energy standardization as for calculations made from the Finite Element Method on thick mesh. It allows the conducting of simulations for Finite Elements for deformations development on thick mesh without losing accuracy in the result. Similarly, Szávai [16] has proposed a thermal-metallurgical-mechanical tridimensional model of finite elements with the aim of researching the residual stress microstructure and distribution of a welded joint among dissimilar metals in a pressurized vessel. This author has found there is an acceptable agreement between predicted and measured data. Besides, both the numerical model and the experiment show that strengthening by deformation is the cause of final residual stress. Comparably, Guo [17] has developed a numerical model of tridimensional fracture by using the combined finite-discrete element method with the purpose of providing a base for engineering applications. It has been demonstrated that accuracy in tridimensional fracture modelling depends on the size of the element around crack fronts. One of the most used models for crack propagation in mechanical lineal fracture is Paris law on which researchers as Ciavarella [18], Ancona [19], Carrascal [20], Kirane [21] [22], Rajabipour [23] and Toribio [24] have carried out their studies from different perspectives.
As stated above, research in Fracture Mechanics field is based on the evaluation of specific phenomena enabled by certain fractal mechanical parameters such us the stress intensity factor. The aim of this research is the calculation of crack propagation in welded joints subjected to cyclic axial load; as main characteristics the use of a none-standardized geometry is node, commonly used in structural joints and the inclusion in the simulation model of two welding leg sizes.
The methodology used in this research consisted in the creation of CAD modeling software, as for the purpose of stress-deformation analysis along the specimen before crack growth, incorporation and crack growth. Finally, the construction of Paris curve from crack stable propagation, from stress intensity factors detected by the software (FRANC3D).
To estimate the rate of crack propagation for a given initial geometry, the Paris-Erdogan law is used [33].(14)Where is the crack growth rate and it is expressed in mm/cycle; ΔKI is the rise of SIF and it is expressed in ; C and m are the constant of Paris Law, which depend on the material and in this case they have been considered as y 3, respectively.
From the computational simulation it is found that the largest stress concentration is in the region near the weld bead (yellow region), with a maximum stress located at the start of the crack propagation of 549 MPa. Once the area where the crack might be located (As shown in Fig 8), it was written, using ANSYS, a file type CDB, hence the format is compatible with FRANC3D.
After having imported the model of this software, the crack is inserted as shown in Fig 9 and the simulation of the propagation of this is initiated, having set growth parameters (growth model, failure size, number of steps, etc.).
In this section analysis results of propagation carried out under loads exclusively of tensile for two sizes of weld reinforcement are presented. The growth rate for each case corresponds to the following characteristics: the initial crack width has 1.2 mm (measure in the semi-major axis of the ellipse) located approximately in the half of the weld toe length and the configuration corresponds to a linear increase depth of 0.12 mm applied to a total of 50 steps of length. For the specimens whose weld bead had a reinforcement of 2mm and 3mm; the crack propagation was just as it is shown in Fig 10.
On the other hand, the former graphs show that there was always a change of surface in the propagation for specimens of 2 mm and 3 mm of weld reinforcement, the magnitude of the stress intensity factor for this crack front becomes significantly higher when the crack front distances more from the initial surface of propagation [34].
In Fig 13 crack growth step is illustrated and the rise of stress intensity factor for each step. It is shown how the distribution of stress is practically the same in the propagation until step number 22 where a gradual break of KI values starts for both geometric configurations.
Paris law allows to estimate the number of produced cycles to reach the preset progress in each iteration, in which considering front shape variation of an iteration to the following determines the crack propagation rate for the studied models. It is observed that weld reinforcement had no significant effect in the crack growth speed, for the two types of weld reinforcement considered in this study. The above is related to the short crack size of the subjects of study models. When analyzing Fig 11, discrepancies are observed in the growth of the crack (a) in the last steps of the simulation that correspond to the change of the propagation surface (see Fig 10). Therefore, Regarding the initial size and the growth rate of the crack remained constant for both cases, it is possible to affirm that these are small numerical errors of the software. 2ff7e9595c
تعليقات