Use of partial metallurgical injection to control the solidification forces in fusion welding processes
In welding technology, the local thermal load leads to high temperature gradients and consequently to the formation of residual stresses in the component during and after cooling. As a result of thermal expansion and microstructural changes, thermal and transformation residual stresses arise which influence the plastic component distortion and have a detrimental effect on the precision of the component. If the total stresses exceed the temperature-dependent yield point, plastic distortion occurs. For this reason, the sub-project is concerned with controlling the residual stresses that build up in a component by metallurgical injection. For this purpose, the compressive stress build-up during the γ-α-phase transformation is maximally exploited by reducing the martensite start temperature (Ms) to a lower temperature by manipulating the chemical composition in the weld. This reduction is called the low transformation temperature (LTT) effect. In this process, the volume expansion of the martensitic phase transformation is used after the majority of the tensile stresses have built up in the weld. Thus, the compressive stresses created by the phase transformation counteract the thermally induced tensile stresses during cooling and prevent the yield strength from being exceeded. The temperature distribution and the development of the residual stresses and phase transformations during cooling can be mapped using a simulation model. Here, the development of residual stresses over time during the welding process and the effect of phase transformations on the residual stresses can be depicted, the time of the maximum tensile stress formed can be represented and the ideal transformation temperature for the LTT effect can be derived.
While in the literature the LTT effect is mostly investigated in the arc process, subproject A07 has proven in the course of SFB 1120 that this effect can also be successfully used in the beam welding process.
In phase 1, the LTT effect was demonstrated in electron beam and laser beam welding. In phase 2, the influence of different alloy compositions on the built-up residual stresses in the component was quantified and a reduction in component distortion was demonstrated. Here, special filler wires were not used as in the literature, but the desired alloy composition was produced by mixing dissimilar materials during welding in the molten pool itself. The effect could be demonstrated not only in a structural steel but also in an austenitic steel.
In phase 3, the focus is on the process control of martensitic phase transformation for targeted thermal strain control in the beam welding process. This is supported by numerical structure simulation in order to specifically address the temperature range by means of martensitic transformation, which is decisive for the residual stresses that determine precision.
Simulative mapping of the welding heat cycle and the formation of the welding residual stresses as well as the component distortion
In order to gain an understanding of the welding process during cooling, a simulation model was built for the ferritic-pearlitic base material S235JR. The first step was to create a temperature field model by using equivalent heat sources, Figure 1 a). With the thermal model, a description of the residual stress development over time was established, Figure 1 b). The phase simulation could be implemented and the influence of this could be clearly seen in the residual stress development. The numerical calculation showed that the tensile residual stresses form at a maximum up to 270 °C. From this, it was deduced that the tensile residual stresses should not exceed the maximum values. From this it was deduced that the target Ms temperature should be below this value. When observing the residual stress in the component, the influence of phase transformations could be observed, Figure 1 c). In the cooled state, the phase transformation leads to a pressure induction in the weld seam, which is why lower residual stresses occur. This shows the importance of considering the phase transformation. With this model, the phase fractions in the weld seam could finally be determined, Figure 1 d). The structure of the simulation model for phase transformation and residual stresses was presented in some publications of this subproject.Copyright: © SFB 1120
Figure 1: Simulation model of the base material S235JR. a) Simulation of the thermal field, which serves as the basis for the residual stress and phase simulation; b) Longitudina l residual stress development in the weld (with and without phase transformation); c) Measured and simulated longitudinal residual stresses in the weld after welding (with and without phase transformation); d) Simulation of the phase fractions in the weld.
Investigation of the volume expansion behaviour of LTT alloys
If the Ms temperature is shifted to low values, the amount of compressive stress build-up also changes. In order to investigate the influence of different transformation temperatures on this, varying alloy compositions were created with different filler wires. Subsequently, residual stress measurements were carried out using the drill hole method, Figure 2 left. It was found that with higher alloy content, the compressive stress build-up becomes increasingly larger. Dilatometer samples were created from samples melted in the furnace and the martensite start temperature was measured, Figure 2 right. The compressive stress level increases with increasing strain, but it can be observed that from 200 °C and below the expansion increases only minimally. A further reduction of the Ms temperature above this temperature would therefore no longer make sense.
Figure 2: Residual stress measurement and dilatometry on varying alloy compositions in an S235JR. Left: Higher alloy contents lead to a larger compressive stress level in the weld, Right: The higher alloy contents lead to a larger expansion at a lower martensite start temperature. Bottom: Fully martensitic weld in an austenitic CrNi steel (1.4307).