The present invention relates to a method for improving the fatigue properties in a component or structure by a local heat treatment of a workpiece and a workpiece treated by such method.

In many engineering applications components and parts are fabricated using a range of processing techniques. The techniques used are dependent upon the material from which the components are constructed and their intended application. For example, engineering components fabricated from metals and alloys are often processed using techniques such as extrusion, forging, drawing, bending and rolling as well as joining methods like fusion welding, brazing and clinching.

Such operations introduce residual stresses within the structure, which may be added up during a multi-stage process route. It is well known that the resulting residual stress distribution has a significant influence on the cyclic loading capacity of fatigue exposed structural members or parts.

The present invention aims to modify the residual stress distribution in a way which enhances the capability of the structural member or part to withstand repeated cyclic loading with constant or variable amplitude and frequency. This is achieved by deliberately superimposing compression residual stresses in regions that are susceptible for initiation or growth of fatigue cracks. The superimposed stresses are thermally induced stresses resulting from a local thermal heat treatment.

The method requires a heat source, e.g. an induction coil, which generates heat at certain well defined positions within the material, and preferably a numerical simulation program, i.e. a FE-program for simulating the influence of the different processing steps, included the local heat treatment, on the resulting residual stress distribution. By proper choice of positioning and strength of the heat source, it is possible to obtain compression stresses in regions that are susceptible for initiation or growth of fatigue cracks. The result of the local heat treatment will be a component or part possessing significantly improved resistance against cyclic loading, since it is well known that compression stresses generally show a positive effect on the fatigue properties.

One particular favourable application of this invention is in improving the fatigue properties of fusion welded joints, which are subjected to cyclic loading during the service life. Fusion welds, like MIG or TIG weldments, are particularly susceptible for fatigue cracking. This is because it is almost impossible to avoid small imperfections at the fusion boundary (i.e. the weld toe), where the stress concentrations are high. At the same time, unfavourable residual tensile stresses often results due to the inhomogeneous thermal expansion and contraction of the material taking place during the weld thermal cycle.

It is well known that the residual stress distribution has a significant influence on the cyclic loading capacity of fatigue exposed structural components and parts. Furthermore, it is also known that the fatigue crack growth propagation can either be slowed down or accelerated by manipulation of the residual stress distribution.

It is also well known that compression stresses have a positive effect on the fatigue properties of a component by slowing down the cyclic crack growth rate or by increasing the number of cyclic loadings needed before a crack starts to grow.

Different techniques are used to impose compression stresses in critical regions of a component where fatigue cracking is likely to occur. These include different types of local mechanical treatments like grinding or hammer-peening which both introduce compression stresses in the surface. The known thermal methods usually aim to remove the existing residual stresses by annealing. This is usually done by heating the whole component to sufficiently high temperatures where local plastic yielding occurs.

Rough guidelines for local post-weld heating also exist for certain idealised geometries, such as semi-infinite butt weld geometries and circular tubes. However, for a person skilled in the art, it is obvious that these guidelines are not valid for components in general with geometries that derive from the ideal cases.

Common for all existing mechanical methods is that they are generally labour- intensive, time-consuming and costly and therefore usually not feasible in high- volume productions, e.g. in fabrication of fatigue exposed automotive structures

The methods are rather used as a last resort against poor fatigue design at the end of the manufacturing process, when no other options are available.

For the known thermal methods, the residual tensile stresses are attempted reduced or eliminated by annealing. These methods often involve heating of the whole component, which can be highly inconvenient and costly. Moreover, such heat treatments will usually not bring the critical regions, where fatigue cracks are likely to appear, into compression. Hence, heating of the whole component does not utilise the full potential of manipulation of the residual stress field.

Furthermore, if a local heat treatment is applied, it will be very difficult to prescribe an optimum process, which gives the required effect on the fatigue properties if no simulations are applied. Actually, an inadequate local heat treatment, e.g. by a poor choice of positioning of the heat source, can even worsen the fatigue properties, by increasing the tensile stresses rather than reducing their magnitude at the critical regions of the workpiece.

The present invention represents a simple, flexible method that can be integrated directly into the manufacture process or at an early stage of the design process in order to improve the fatigue properties of structural components and parts which will be exposed for repeated cyclic loading.

The improved fatigue properties are obtained by slowing down the cyclic crack growth rate or by increasing the number of cyclic loadings needed before a crack starts to grow. This can be achieved in a cheap and efficient way by thermal manipulation of the residual stress distribution at fatigue exposed regions of the structural component or part.

The method requires a heat source, e.g. an induction coil, which generates heat at certain well defined positions within the material, and preferably a numerical simulation program, for simulating the influence of the different processing steps, included the local heat treatment, on the resulting residual stress distribution. The concept involves an initial estimation of the residual stress distribution resulting from fabrication of the component or part, and a further estimation of the effect of the local heat treatment by different alternative processing conditions and positions of the heat source. The estimated residual stresses are usually most conveniently obtained by applications of a FE-model.

The local heating can be carried out by a range of different heat sources, e.g. induction heating, resistance heating, arc heating (e.g. applying a MIG or TIG apparatus). The heat source can be stationary or moving. In the latter case, the heat source can be either manually guided or automatically guided by some sort of device, e.g. a robot. It is also possible to apply several heat sources simultaneously.

The invention does not require any mechanical tooling or device that impinge on the surface of the workpiece since the plastic strains required to change the residual stresses are imposed by the use of a heat source. The method is a non-contact method, which means that the wear of the equipment and the maintenance costs can be kept very low, while the up-time in serial production will be correspondingly high. Furthermore, since the heating can be performed with access from just one side, regions of the workpiece can be reached which could not be accessed by using some of the traditional mechanical techniques, like hammer-peening, which may require a backing tool.

Another advantage of using the present concept is that the structure just needs to be heated locally rather than heating the whole structure, which is a much more comprehensive operation that is not even possible for large structures.

Contrary to empirical based methods for determining the local heat treatment, the application of numerical simulation tools like FE-models ensures that the method will be more reliable, and yield the required effect with respect to improved fatigue properties.

The basic idea is to change the direction of the residual stresses at fatigue exposed regions from tension to compression by the application of local heating at some carefully selected positions. These positions are not straightforward to determine, and the same is the case for the power intensity and the duration of the local heat treatment, which are not known a priori.

The present invention preferably utilizes numerical simulation models like e.g. FE- models in order to determine an adequate local heat treatment, which gives a positive effect on the residual stress distribution.

More than one FE-simulation may be necessary to run before an adequate result is obtained and an acceptable residual stress distribution in the fatigue-exposed regions is achieved. The FE-simulations are typically run in a systematic manner, where the positioning and power of a moving or stationary heat source is varied in the simulations. The present concept also applies to situations where large series of components that exhibit similar geometries or common characteristics concerning the residual stress fields may provide basis for QA guidelines for local heat treatment for a family of components by interpolation between simulation results. Such documentation may also serve as input for future design rules and common industry standards with respect to fatigue.

The FE-simulation can alternatively be partly or completely replaced by physical measurements of the residual stresses, although this is usually not preferred due to cost and time considerations.

The invention will now be explained by means of an example with reference to the accompanying drawings, in which

Figs, la-c show stepwise the treatment of a workpiece, Fig. 2 shows a welded structure, Fig. 3 shows a top view of the welded structure where the welding sequence is indicated, Fig. 4 shows the regions where a local heat treatment is applied, Fig. 5 shows the predicted residual stress distribution along the line A-A shown in Fig. 2, after welding, and after welding and local heat treatment, Fig. 6 shows an apparatus for fatigue testing, Fig. 7 shows examples on different welded joints where the invention can be applied, Fig. 8 shows examples on different geometries where application of the invention can improve the fatigue properties.

Figs, la-c show stepwise the treatment of a workpiece. The method according to the present invention consists of predicting the location of zones which contain unacceptably high tensile stresses at fatigue exposed regions. The next step will then be to seek one or several local heat treatments which change these residual stresses to an adequate level and finally apply these heat treatments on the structure.

Fig. Ia shows an example of a workpiece 1 where local zones 2 possessing unacceptably high tensile stresses are located. Fig. Ib shows the location of the regions 3 where a local heat treatment is preferably applied in order to change the direction of the residual stresses from tension to compression as shown at the heat treated local zones 2 in Fig. Ic. The location of the locally heat treated regions 3 is preferably calculated in a simulation procedure.

The present example relates to a welded component although the invention is equally applicable for other structural components or parts fabricated by other operations or processes where residual stresses represent a problem. The structure according to the present example consists of two identical rectangular hollow sections.

As shown in Fig. 2, the butt end 5 of a tube 6 is welded against a flat side 7 of another tube 6. The resulting welded structure will be referred to as a Tee-joint 8. During welding, the area close to the weld 9 will be heated, leading to a Heat Affected Zone (HAZ) 10. The inhomogeneous and rapid temperature fluctuations, which take place during the welding, lead to residual stresses both in the solidified weld metal in the weld 9 as well as in the HAZ 10 after the welded component or part has cooled down.

In the present example, the weld deposition is carried out along the circumference of the butt end 5 of the tube in two passes, as shown in Fig. 3, which also shows the locations of the start and stop positions of the welds in the present example. From the starting position 11 two welds Wl and W2 are made, encircling the circumference of the butt end 5. The two welds will meet at a stop position, thereby ensuring a continuous weld around the butt end 5 of the tube. It should be understood that the weld can also be performed in only one pass or more than two passes if workpieces of a different design are welded. The present method can be used on any welded structure independently on how the welding operation is performed. After welding is completed and the components are completely cooled down, residual stresses will occur in the welded area, as explained above.

In experiments carried out some of the welded structures were given a local heat treatment to modify the residual stress distribution resulting from the welding process, while others were kept in the as-welded condition. The heat treated regions 3 in the present example are shown on Fig. 4. Fig. 4 shows the Tee-joint 8 in a side view. Several heat treated regions 3 are located in the proximity to the weld 9.

In these experiments, structures were used where the cross section dimensions are 60 mm x 40 mm with 3 mm wall thickness. The structures are made of the aluminium alloy 6082 in the peak aged, i.e. T6 temper condition. The resulting welded structure is shown on Fig. 2. The local heat treatment, which was imposed to some of the welded samples was carried out applying a standard induction heating equipment where the footprint of the induction coil facing the workpiece was 10 mm x 30 mm.

Based on a systematic series of FE-simulations, which are described below, induction heating was conducted applying a net power and an operation time of 1.1 kW and 6 seconds, respectively. During the heat treatment, the temperatures in the local heated regions are raised significantly. However, the selected combination of power and operation time ensures that no positions within the local heated regions reach the temperature where incipient melting may occur.

Due to the temperature increase associated with the local heat treatment, thermal expansion will take place in certain regions of the component during the heating stage. This, in turn, may lead to local plastic yielding which, in turn, will change the residual stress distribution after the component has been completely cooled down. The actual choice of parameters and location of the induction coil, as shown in Fig. 4, were based on a systematic series of FE-simulations, where these parameters were varied and the calculated resulting residual stress distribution was registered. As described above, a FE-program was used to predict the residual stresses following the welding process as well as the resulting residual stress distribution following both the welding and the subsequent local heat treatment. The FE-program WeldSim was chosen for this purpose, since this code is especially suitable for simulations involving welding and heat treatment of aluminium alloys.

The calculated residual stress distributions are shown in Fig. 5. The figure shows the residual stress distribution along the line A-A as defined in Fig. 2 after welding, and after welding and subsequent local heat treatment. The figure shows the normal stress component at the surface in the longitudinal direction of the horizontal profile, which is expected to be the critical stress component for the actual cyclic loading condition, which will be described below.

Fig. 5 shows two curves where the solid line represents the residual stresses after welding, while the broken line represents the residual stresses after welding and subsequent local heat treatment.

After welding the stresses are tensile stresses approximately in the range 50MPa to 100MPa, while after the subsequent local heat treatment, the same stress component is changed to compression in the range -5MPa to -50MPa. Hence, the direction of the zz-stress component along the weld toe is fully reversed along the whole section A-A after the local heat treatment, according to the FE-simulation.

The FE-simulations indicate that it is possible to obtain a significant difference in the residual stress distribution between as welded Tee-joints, and the same Tee-joints given a subsequent local heat treatment.

Fig. 6 shows a test rig 13 for fatigue testing. The Tee-joints 9 were tested in 4-point bending, that is with a constant bending moment across the weld section. The loading used in the fatigue testing was sinusoidal constant amplitude loading at R = 0.1 [-]. The testing was performed in laboratory air at ambient temperature with a loading frequency ranging from 5 [HZ] to 8 [HZ] depending on the applied load level. The fatigue test results are shown in Fig. 7 for two differently treated T-joints, i.e. as welded and as welded and local heat treated (LHT), respectively. The test S-N data are plotted using the number of cycles, N, to a complete loss of the load-bearing capacity of the T-joint versus the nominal stress range, Snom, obtained by means of elastic beam theory. As can be seen from Fig. 7, the LHT T-joints respond significantly better than the as welded T-joints to the fatigue tests. This is easily seen by comparing the stress range at 106 cycles, which give Δσ -values of approximately 57MPa and 82MPa for as welded, and local heat treated samples, respectively.

In general, welding processes introduce stress concentrations and small dis¬ continuities in and around the joint, which both yield a detrimental effect on the fatigue properties. The present method is applicable for any welded component where cracks are likely to occur.

Figs. 8a-d show examples of such welded components and indicate typical regions where fatigue cracks 14 are often observed. The figures show that cracks 14 are likely to be found close to the fusion boundary at the so-called weld toe, where the combination of high residual tensile stresses and high stress concentrations are particularly detrimental with respect to the evolution of fatigue cracking.

Figs. 9a-d show schematically the stress distribution in some simple geometries. Depending on the geometry and the actual loading situation, certain regions of the workpiece, such as sharp edges, will be subjected to high stress concentrations. An object of the present invention is to predict the location where high residual tensile stresses exist, and to advise a procedure for local heat treatment which will change the residual stresses at these critical regions into a less harmful state. By the use of an appropriate analysing tool, such as a FE-programme, this can be done for any component or part independent of the geometric complexity.