ABSTRACT

The thermo-microstructural model TENWELD©, originally developed to predict the hardness and microstructure in the heat affected zone (HAZ) of multipass girth joints of X-65 steels for seamless pipes, has been up-dated. New continuous cooling transformation diagrams of high strength steels have been determined and added to the database of the artificial neural network (ANN), which can now predict the decomposition of austenite to low-transformation temperature microstructural constituents, even in the case of steel grades up to X-100. In the model, analytical equations are used to describe the thermal evolution and a series of physical models have been tuned to evaluate grain growth during austenitization, taking also into account the inhibition effect of microalloying precipitates. Empirical relationships are applied to evaluate the fraction of high-C martensite formed in infra-critical reheating and to predict the hardness after each pass, knowing the volume fraction of the various microstructural constituents formed. In the last release, an algorithm has been developed which fills-up the bevel automatically calculating the position of each bead (i.e. heat source) to be used in the thermal model. This feature is particularly useful to speed-up the input procedure when joints with a large number of passes have to be simulated. The shape of the bevel is established through a few geometrical parameters that are set by the user. As the shape of the two half-bevels can be fixed independently, it is possible to model various joint configurations. The model has been validated through laboratory experiments as well as industrial welding trials, showing acceptable prediction capabilities. The present approach offers high flexibility and the possibility to obtain useful information on the relevant effect of steel chemical composition and welding procedure on hardness and microstructure of the HAZ.