COMPUTATIONAL WELDING MECHANICS

Fracture Mechanics

8.1 Introduction and Synopsis

The development and production of High Strength Low Alloy (HSLA) linepipe steels has been growing rapidly because of their desirable combination of high strength, low temperature toughness, weldablity and low cost.

With the advent of the welding as the major method of fabrication, cracking in the heat affected zone has become a serious problem particularly in large and continuous structures.

One of the major problems in the welding of steels has been the type of cracking generally known as hydrogen induced cold cracking, which is described in chapter VI.

At high carbon levels, the heat affected zone cracking was severe because of the formation of brittle martensite as a result of rapid cooling in welding. In the presence of hydrogen gross cracking was inevitable. Recognizing the deleterious effect on notch toughness and weldability by carbon in welding, the trend towards newly improved materials has resulted in the development of lower carbon steels. Quenched and tempered steels were produced using alloy elements such as manganese and silicon as the main strengthened in solid solution. These additives increased the hardenability of the steels to make heat treatments possible. With the concurrent development of low hydrogen electrodes, these steels were easily welded without any cracking problem at least in thin sections. In general, preheat is only required in thicker section and when the restraint is high. However, since mechanical properties of these steels were strictly controlled by careful heat treatments during processing, the thermal cycle in welding can significantly impair strength and toughness in the fusion zone and the heat affected zone. This limitation in the use of quenched and tempered steels has paved the way for developing lower cost, high strength and toughness low alloy steels. There are five major factors determining the strength, impact toughness and weldability of low alloy steels. They are carbon content, dislocation density, precipitation hardening, solid solution hardening and grain size. Only reduction of grain size significantly improves both strength and toughness; the other strengthening mechanisms improve strength at the expense of toughness, [5 and 11].

Weld solidification cracking (hot cracking) has been a persistent problem in a variety of engineering alloys. The formation of solidification cracks result from the combined effects of metallurgical and mechanical factors. The metallurgical factors relate to conditions of solidification, grain size, presence of low - melting phases, etc. The mechanical factors relate to conditions of stress/strain/strain rate developed near the trailing edge of a weld pool during solidification. In other words, solidification cracking is a result of the competition between the material resistance to cracking and the mechanical driving force. The material resistance to cracking is primarily influenced by alloy composition, the welding process and heat input. The mechanical driving force depends upon the welding process, heat input, joint configuration and rigidity and thermo mechanical properties of alloys.

Although the precise mechanisms responsible for solidification cracking and the remedies for its prevention are still not clear, research efforts have revealed a more or less qualitative picture about the nature of such cracking. A hot crack requires strain localization. Therefore it must require strain softening. It appears that rate-dependent, i. e. viscous, behavior plays a critical role in hot cracking. The solidification of the weld pool is a non-equilibrium process and depending on growth speed and temperature gradient and can result in the formation of columnar microstructure or

Hot cracking of welds is an important problem in the welding industry. It is thought to occur in a temperature range that begins at the start of solidification and ends soon after solidification is entirely complete.

During the later stage of solidification, there exists a brittle temperature range (BTR) in which the strength and ductility of the alloy are very low as some low melting-point constituents segregate between dendrites and form liquid films. At the same time, stresses/strains arise from solidification shrinkage, thermal contraction of parent metal and external restrains.

Like many other cracking problems, solidification cracking occurs when the mechanical driving force exceeds the material resistance to cracking. This concept is illustrated by Feng et al [7] in Figure 8-1.

Hot cracking is observed to be sensitive to chemical composition. It is thought that a critical strain and critical strain rate exists. The critical strain for hot cracking vs. temperature at constant strain rate for some 0.2%C steel could be described by the points (0.1, 1490),
(0.04, 1480), (0.1, 1470), see 5-6-2. As the strain rate is decreased, the critical strain decreases but below a critical strain rate, hot cracking does not occur, [3].

Hot cracking has been studied in depth from the viewpoint of the effect of metallurgical variables. In addition a number of experimental techniques have been developed to measure sensitivity to hot cracking and study the phenomenon of hot cracking. These include the Houldcroft, Varestraint and Sigmajig tests. The Houldcroft test varies the restraint by sawing slits into the plate in an attempt to vary the strain and strain rate. No external forces are applied. The Sigmajig test which is commonly used for sheet metal applies a transverse force or stress. The Varestraint test applies a displacement that strains the weld pool region, [3]. In the 1980s, Matsuda et al [16] published a series of basic research results on weld metal solidification cracking phenomenon based on a new technique, Measurement by means of In-Situ Observation (MISO).With assistance of the MISO technique, Matsuda was able to directly measure the ductility around the solidification crack tip on a very local scale (about 1mm in gauge length). Recently, Lin and co­worker [17] improved the measurement of the BTR and demonstrated that, by using the new measurement methodology developed, the BTR is material specific and independent of the testing conditions.

In plain carbon steel structures, it has long been recognized that brittle fractures almost never propagate along the heat affected zone (HAZ) of welds; consequently little effort was made to measure the toughness of the HAZ. However, with the increasing use of medium and high strength steel and the development of high heat welding processes, the toughness of the HAZ was questioned and a number of papers were published on the subject.

The Charpy F-notch test is a standard laboratory test for welded joints. Goldak and Nguyen [5] present data proving that conventional Charpy-V notch (CVN) testing of narrow zones in electron beam (EB) welds can produce dangerously misleading results and describe a simple economical technique for measuring the true ductile brittle transition temperature of narrow zones in welds. They suggest the Cross Weld Charpy Test (CWCT), which actually does measure CVN FATT (Fracture Appearance Transition Temperature) in EB weld metal. In addition to CVN tests in which the weld plane is parallel to the fracture plane, Figure 8-2, the CWCT uses a conventional Charpy - V notch specimen except that the weld plane is perpendicular to the fracture plane to force the fracture to pass through all zones of the weld.

The competition between brittle and ductile modes of failure in a weld specimen under dynamic loading has been analyzed by Needleman and Tvergaard [20], in terms of a micromechanically based material model. In addition the ductile-brittle transition for different weld joints has also been investigated by plane strain numerical analyses of Charpy impact specimens by the same authors [4].

In the 1970s, Chihoski published three papers [2, 14 and 15] on an experimental study of displacements near the weld pool. His objective was to understand the strain field around a weld pool and the effect of welding variables, such as welding speed, on the strain field and the susceptibility to hot cracking. He used a Moire’ Fringe

method. He studied three edge welds and three bead-on-plate welds. These were TIG welds in thin aluminum-copper alloy, Al 2024, plates. He used welding speeds of 6, 13 and 20 ipm. He found a compressive region near the weld pool. Roughly the thermal

expansion near the weld pool causes a compression region that is later overwhelmed by the shrinkage stresses as the weld cools, Chihoski observed that the location of this compression region relative to the weld pool was a function of welding speed. By

placing this compression region in the region sensitive to hot

cracking, he was able to weld Al 2024, an alloy sensitive to hot cracking, without hot cracking.

In addition to the study of metallurgical affects and

microstructure, which are clearly important to hot cracking in welds, it is also necessary to develop and test a capability for stress analysis near the weld pool. Chapter V presents a preliminary attempt to develop a capability to do quantitative analysis of the stress and strain near the weld pool. In addition experimentally determined constitutive equations or other forms of experiments such as those of [14 and 19] will be needed to validate the analysis. Ultimately microstructure models of dendrites and interdendritic liquids will be needed. Even with these limitations, the model introduced in chapter V appears to be a useful step towards quantitative analysis of stress and strain in the region susceptible to hot cracking in welds.

Feng and co-worker [7] present also the development of a finite element analysis procedure and the calculated dynamic stress/strain evolution that contributes to the formation of solidification cracks in the cracking susceptible temperature range.

As interest in enlarging the gas throughput increases, the use of larger diameter and higher pressure gas transmission pipelines will rise. There will then be an increasing need for reliable pipeline design, inspection and maintenance procedures that will minimize service failures. Concern for the possibility of ductile fracture propagation in gas transmission pipelines stems from two main sources. With the ever-expanding gas transmission system, the probability rises of a third party inflicting damage severe enough to initiate fracture. Further, as the current systems age, the probability of insidious corrosion damage growth producing a local rupture event increases. Failure through either of these mechanisms can result in the initiation of a long-running ductile fracture. A methodology has been developed by O’Donoghue et al [6] that can be used on a routine basis to predict the possibility of the occurrence of long-running cracks in gas transmission pipelines.