Modifying residual stress levels in rail flash-butt welds using localised rapid post-weld heat treatment and accelerated cooling

Abstract

Flash-butt welding is used in the manufacture of continuously-welded rails. Finished welds typically exhibit high tensile residual stresses in the rail web and at the upper surface of the rail foot, which may increase the risk of fatigue failure in service. An understanding of the influence of the welding process, including post-weld cooling, on the residual stress distribution is necessary to improve the performance of flash-butt welds by post-weld heat treatment (PWHT), since incorrect treatment may have adverse effects on both residual stress and weld material characteristics. A finite element model has been developed to simulate post-weld cooling in flash-butt welded AS60 kg m^–1 rail. Computed thermal histories for normal (air) cooling, rapid PWHT, and accelerated cooling (water spray) were used as inputs to calculate sequentially coupled stress–time histories, including phase transformations. In addition, the localised influence of the initiation time for rapid PWHT, after final upset, on the reduction of tensile residual stresses was investigated. Heating the rail foot immediately after final upset reduced tensile residual stresses in the web region of the weld. Preliminary numerical predictions showed that water quenching the entire weld region too soon after the austenite–pearlite transformation is completed can induce further tensile residual stresses without affecting the microstructure. The results of the numerical analysis can be used to modify the flash-butt welding procedure to lower residual stress levels, and hence improve weld performance.     

Integrated fracture mechanics approach to analyse fatigue behaviour of welded joints

Abstract

Current fracture mechanics methods for fatigue assessment, including those that consider thresholds for crack propagation, are based on long crack behaviour. The present work is concerned with an attempt to predict the fatigue strength of welded joints using a fracture mechanics approach that takes into account the fatigue behaviour of short cracks. The methodology estimates the fatigue crack propagation rate as a function of the difference between the applied driving force and the material threshold for crack propagation, which is a function of crack length. The fatigue strength of butt welded specimens stressed transversely was analysed. Experimental results from the literature were used for comparison. Estimations are obtained by using only the fatigue limit and the fatigue propagation threshold for long cracks, and the applied stress distribution along the crack path obtained from simple finite element models. The influence of plate thickness, initial crack length, and reinforcement angle on fatigue strength of butt welded joints was analysed. Results show good agreement with experimental trends.     

Welding distortion investigation in fillet welded joint and structure based on iterative substructure method

Abstract

In this paper, prediction and controlling angular distortion in fillet welded joint and structure were investigated. First, two methods to reduce angular distortion in fillet welded joint were investigated by experimental and numerical analysis. One was to apply a constant external force in-process and the other was rigid clamping. In numerical analysis, a new in-house finite element code has been developed based on the idea of iterative substructure method (ISM) to calculate welding distortion in rational time. During experimental analysis, the constant external force was designed and applied in-process to reduce angular distortion of fillet welded joint. The results showed that the distortion can be efficiently predicted by ISM, which were in good agreement with the experimental ones. Applying constant external force in process was a more effective method to reduce distortion than using rigid clamping. In addition, with a constant load distance from weld bead, the locations of the applied constant external force and rigid clamping along the longitudinal direction (welding direction) have little influence on the magnitude of welding angular distortion. Finally, the angular distortion of a large fillet welded structure was predicted with ISM and also controlled with applying a constant external force based on the simulation results of the fillet welded joint.     

Evaluation of through thickness residual stress in multipass butt welded joint by using inherent strain

Abstract

The through thickness residual stress of an eight pass butt welded plate joint is evaluated using inherent strain analysis. The residual stress distribution is obtained in detail along the thickness direction from measurements using multiple strain gauges. The residual stresses agree with the results of the thermal elastic–plastic analysis as well as the values obtained by direct measurement of the specimen surface, which is not used in inherent strain analysis. These results indicate that both inherent strain analysis and thermal elastic–plastic analysis are effective in evaluating through thickness residual stress. Therefore, each analysis method should be chosen after considering the object to be evaluated and the characteristics to be analysed.     

Suppressing type IV failure via modification of heat affected zone microstructures using high boron content in 9Cr heat resistant steel welded joints

Abstract

Creep rupture strength at 923 K and microstructural evolution of welded joints have been investigated for high boron–low nitrogen–9Cr heat resistant steels developed at the National Institute for Materials Science (Japan). Welded joints were prepared from plates containing 47–180 ppm boron using gas tungsten arc welding and Inconel type filler metal, and showed superior creep properties to those of welded joints of conventional high chromium steels such as P92 and P122. No type IV failure was observed in the boron steel welded joints. A large grained microstructure was observed in the heat affected zone heated to Ac ₃ (Ac ₃ HAZ) during welding, whereas the grains are refined at the same location in conventional steel welded joints. The simulated Ac ₃ HAZ structures of the boron steels have a creep life almost equal to that of the base metal. Large grained HAZ microstructures and stabilisation of M₂₃C₆ precipitates are probable reasons for suppression of type IV failure and improved creep resistance of the boron steel welded joints.     

Microstructure and residual stress variations in weld zone of flash-butt welded railroads

Abstract

In the weld zone and base metal contiguous to the weld zone in flash-butt welded rail, the web of the rail has high tensile residual stresses and the head and base of the rail have compressive residual stresses. The web region is susceptible to failure since most of the weld zone of the rail is coarse grained and has porosity, inclusions and defects resulting from rapid solidification of molten metal entrapment in the weld. Efforts to reduce the amount of these tensile residual stresses require recognition of their causes. In this research, microscopic and macroscopic studies were carried out on vertical and horizontal sections of the weld zone in the head, web and base of the rail. Just after flash-butt welding, the temperature of the web between the current carrying copper electrodes is higher than the temperature of the head and base of the rail. Therefore, by cooling the weld zone to room temperature, the amount of web contraction between the electrodes is higher than the amount of base and head contraction and consequently tensile residual stresses are produced in the web at and near the weld zone. In the head and base of the rail, compressive residual stresses are developed.     

Residual stresses and distortions in welded structures: a perspective for engineering applications

Abstract

In this paper, some of the historical developments and recent advances in understanding welding induced residual stresses and distortions are discussed in the context of their impact on today's engineering applications. With recent rapid advances in computational simulation techniques, the complex thermomechanical phenomena associated with typical welding processes can be effectively dissected into solvable problem sets for both fundamental understanding and specific engineering applications. Although a great deal of research is currently still ongoing to advance our fundamental understanding of the complex thermophysical and thermomechanical phenomena, an engineering perspective is provided to demonstrate how an engineer can make use of the current state of knowledge to derive effective practical solutions when dealing with day to day problems in the areas of residual stresses and distortions in welded structures. First, some of the fundamental mechanical considerations associated with residual stress and distortion developments are presented. Then, some of the computational modelling requirements for engineering applications are discussed in the light of recent developments. Finally, application examples are presented to demonstrate how an effective engineering solution can be sought by taking advantage of today's advanced modelling techniques with appropriate engineering assumptions.     

Wood joints and laminated wood beams assembled by mechanically-welded wood dowels

Dowel welding by high-speed rotation was used to join two wood blocks and strong joints were obtained. Dowel angle to the surface of the wood blocks to be joined had a marked influence on the mechanical performance of the joint. When the dowel was inserted at 90° to the substrate, the dowel was subjected to and resisted a shear force only. When introduced at an angle such as 30° or 45°, the dowel was subjected to and resisted both shear and tension forces, resulting in better joint strength. The joint almost always failed by dowel fracture. The dowel/substrate interface was almost always stronger and did not break. Short two-layer beams joined exclusively by a series of welded dowels were prepared, tested in shear according to structural standards, and their performance was compared to those of solid wood and of glued laminated beams (glulam) of the same dimensions. The short two-layer beams prepared for testing met the Eurocode 5 standard requirements when the optimum dowel insertion angle was used. Then 2-m-long two-layer wood beams were prepared, with the two layers connected exclusively by a series of welded dowels, and tested in bending. Their maximum failure strength and stiffness in bending were determined. These beams outperformed both nailed beams and glued-dowel beams. All the beams had the same length and conformation. The number of nails necessary was double the number of dowels used.     

Determination of lower bound fracture toughness of a high strength low alloy steel welded joint

Abstract

The use of high strength low alloy steels for high performance structures (e.g. pressure vessels and pipelines) requires high strength consumables to produce an overmatched welded joint. This globally overmatched multipass welded joint contains two significantly different microstructures, as-welded and reheated. In this paper, the influence of weld metal microstructure on fracture behaviour is estimated in comparison with the fracture behaviour of composite microstructures (as-welded and reheated). The lower bound of fracture toughness for different microstructures was evaluated by using the modified Weibull distribution. The results, obtained using specimens with crack front through the thickness, indicated low fracture toughness, caused by strength mismatching interaction along the crack front. In the case of through thickness specimens, at least one local brittle microstructure is incorporated in the process zone at the vicinity of the crack tip. Hence, unstable fracture occurred with small, or without, stable crack propagation. Despite the fact that the differences between the impact toughness of a weld metal and the that of base metal are insignificant, the fracture toughness of a weld metal can be significantly lower.     

Development of a 2-d.o.f. finger using load-sensitive continuously variable transmissions and ultrasonic motors

This paper presents a 2-d.o.f. finger using ultrasonic motors and load-sensitive continuously variable transmissions (CVTs) we proposed previously, which consist of a five-bar linkage and a torsion coil spring. The proposed CVT is remarkably simple and enables a finger joint to exert a large fingertip force and to move quickly. For the two joints of the 2-d.o.f. finger, we designed five-bar linkage CVTs which are more compact than the previous ones so that they can be installed on its base. Taking advantage of the geometric feature that ultrasonic motors are thin, we can arrange the drive systems for the two joints, including the two CVTs and the two ultrasonic motors, in a parallel configuration. The experimental results verify that the maximum fingertip force of the finger is as large as 30 N and the maximum angular velocity of its joints is more than 400°/s. These performance results would be impossible without the CVTs. We also verify that the finger can lift up a 1.5-kg load.