Finite element simulation of high speed pulse welding of high specific strength metal alloys

Finite element simulation of high speed pulse welding of high specific strength metal alloys

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 1 9 7 ( 2 0 0 8 ) 301–305 journal homepage:

932KB Sizes 0 Downloads 11 Views

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 1 9 7 ( 2 0 0 8 ) 301–305

journal homepage:

Finite element simulation of high speed pulse welding of high specific strength metal alloys G. Casalino ∗ , A.D. Ludovico DIMeG, Politecnico di Bari, Viale Iapigia, 186 Bari, Italy

a r t i c l e

i n f o

a b s t r a c t

Article history:

High specific strength metal alloys as aluminium, nickel and titanium alloys are suitable to

Received 28 November 2006

reduce weight in components for transportation industry. Beside component performances,

Received in revised form

manufacturability and production efficiency must be considered when substituting classical

6 March 2007

steels with high specific strength metal alloys.

Accepted 17 June 2007

In a way similar to resistance welding a sound autogenously joint can be achieved by discharging a very high current pulse over a very short time that is called high pulsed welding or capacitative discharge welding.


In this paper a numerical model was built with the aim to support the evaluation, via a

High speed pulse welding

computer aided approach, of the effects of this technique on the high pulsed weldability of

High specific strength alloys

some lightweight metal alloys.

Finite element analysis

The model aimed to reproduce the electrical, thermal and mechanical aspects of the high pulse welding process and to evaluate the temperature profile, the piece displacement, the effects of different welding parameters on the weld seam. © 2007 Elsevier B.V. All rights reserved.



High specific strength metal alloys provide dent resistance, increased load carrying capability, improved crash energy management, or mass reduction through a reduction in sheet metal thickness, or gauge. In steels those properties are achieved through strength improvement, in the aluminium and magnesium alloys the density is greatly lower than steels and it permits to achieve large strength-to-density ratio. That property is of crucial interest when masses have to be moved and lightweight components are desirable. In the transportation industry high stiffness and/or strength combined with low weight is required in order to reduce fuel costs and polluting consequences for the whole environment. From a manufacturing point of view, all those alloys must be joined into devices or components to form packages, assemblies, and structures to achieve functions they were

Corresponding author. E-mail address: [email protected] (G. Casalino). 0924-0136/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2007.06.049

designed for. In this process welding methods and techniques are critical and must evolve together with the innovative materials and applications. Nowadays the welding techniques are numerous but old and new ones have to be deeply understood in order to provide a better knowledge of the opportunities, hindrances, options that a sound welding of lightweight materials can provide for the sustainability of the transportation industry (Messler, 1995). With the advances of computer technologies and finite element method it is now possible to model many different welding processes such as arc welding, friction welding, laser welding, and projection welding. This numerical technique has proven to be suitable to reproduce the physics and provide useful quantitative information on those welding processes. In this paper the results of a fully coupled electrical and thermal analysis were successively incrementally coupled with a mechanical analysis that took into account the


j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 1 9 7 ( 2 0 0 8 ) 301–305

evolution of the contact area between the welding edges, which is fundamental to the correct simulation of the current discharge (Karlsson, 1997). This model, which was already successfully applied to the investigation of the high pulsed weldability of stainless steel (Casalino and Panella, 2006), is here applied to the Ti6Al4V titanium alloy and to the INCONEL 718 nickel alloy. The FEM model was developed on the ANSYS platform. The effects of the geometry and of the interactions between the electrical and mechanical parameters on the temperature profile, the displacements and the seam length were studied.


The high speed pulse welding

High speed pulse welding (HSPW) is a form of resistance welding but is achieved within milliseconds at very high current levels by utilizing capacitor-stored energy. Therefore a sound autogenously joint can be achieved by discharging a very high current pulse at low voltage over a short welding time. Welding parts are put in contact and pressed together with high force the high current pulse is released through the parts and a fine grain diffusion bond forms with low total energy input (Messler, 1995). Fig. 1 for three-dimensional (3D) and Fig. 2 for twodimensional (2D) meshed profile show the wave profile that was the studied contact geometry. The model dimensions were 12 mm diameter and 10 mm height. The radius of each protrusion was 2 mm. The contact developed along three lines that reduced in three points in the diametrical transverse cross-section, which is showed in Fig. 2. The multipoint contact geometry (Casalino and Panella, 2006) was a first attempt to improve the weldability of the single point contact, which was the first historical proposed geometry (Wilson, 1994). The energy for the pulse is derived totally from charged capacitors which is why HSPW is sometimes called capacitor discharge (or CD) welding, which are the most known welding techniques that use the power of high current discharge stored in capacitors. It is able to produce butt welded joints at cooling rates greater than 106 K/s. Therefore, extremely narrow heat affected zones are produced as well as rapidly solidified welds

Fig. 1 – 3D geometry of the weld.

Fig. 2 – Bi-dimensional mesh.

without losing metastable crystalline structures, good grain refinements and reduced segregation. It is also worth to point out that the deformations induced by this process are negligible (Kristiansen, 1995). HSPW opens up many potential joining applications such as welding together different materials like powder metals, high carbon steel, brass, copper and others. The pulse welding process can be described as follows. As soon as current begins to discharge, following the diagram in Fig. 3 against the time, the components edges begin to fuse. The weld energy is focused onto the contact region by an appropriate protrusion (tip) design and the material in this region heats up rapidly. Due to the high surface pressure the heated material is squeezed out from the contact area and the components displace together (upset) and the resulting contact area is forged welded together. Due to the low heat input and rapid cool-

Fig. 3 – Current discharging diagram.

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 1 9 7 ( 2 0 0 8 ) 301–305

ing rates in HSPW the heat affected zone is extremely small (0.5 mm and lower) relative to other welding processes. This localized heating enables welding near heat sensitive materials and results in no weld zone alteration or part distortion. Typical steels require energy from 4 to 8 J/mm3 to melt, but the designated weld energy settings range from 80 to 200 J/mm3 . The surface contact pressure (system preload) is also designed in proportion to the weld volume and the weld energy and is typically about 200 N/mm3 (Casalino et al., 2003). In theory, lower contact pressure results in greater contact resistance which gives rise to higher localized heat input in the weld zone. If the pressure is too low, the weld head follows up too slowly and the welding materials will be hurled out of the welding zone by rotational electromagnetic fields. For this reason, it is necessary to find an optimum relationship between pressure and energy. To contain the plasticized material of the projection, adequate follow-up of the welding head is achieved by compensating for the effective inertia of the weld head by changing its mass or adjusting the contact pressure. The geometry of the welding edges was also very important for the weld efficacy and soundness.


The finite element model

The FEM model was developed on ANSYS platform. It is a 2D simulation of the cylindrical joint. Since the sample was axial symmetric only one half of the diametrical section was drawn and meshed. The thermal–electrical–mechanical problem was split in two parts. The thermal and electrical calculations were fully coupled. The mechanical part was incrementally coupled with the previous ones (Casalino and Panella, 2006; Sun, 2000). This means that the contact geometry was updated after each step, which lasted 0.001 s each, by means of a user’s subroutine. The sensitiveness of the model towards the current peak, the upsetting force, and the discharge time were tested. The material properties varied with the temperature. They were the specific heat, the emissivity, the thermal conductivity, the convection, the electric resistance, the mechanical properties (elastic, plastic), and the friction condition. The source of the material properties data were the Aerospace structural metals handbook (Maykuth, 1980). The input for the thermal and electric part of the simulation was the peak current and the discharge time. All the surfaces that were not meant to come in contact were assumed to have free convection with the surrounding air. Zero electrical potential was imposed on the top of the lower electrode as a boundary condition for solving the electrical problem. For the structural part the load was the upsetting pressure on top of the upper electrode. PLANE67 2D thermal–electric element, which has thermal and electrical conduction capability, was the mesh element for the bi-dimensional model. It has two degrees of freedom, temperature and voltage at each node, applicable to plane or axe-symmetric, although no transient electrical capacitance or inductance effects are included. The element includes the Joule heating effect in the thermal solution.


For the mechanical calculation the mesh is replaced by the VISCO 106 element with three displacements degrees of freedom at each node. This element is designed to solve both isochoric (volume preserving), large strain plasticity problems due to contact and thermal relaxation. Iterative solution procedures were used since highly non-linear behaviour is expected to occur. For the structural steps, the VISCO 107 is superimposed in the same way of the 2D correspondent element. After that the contact surfaces were identified, TARGE170 and CONTA174 elements, which form a “target-and-contact” pair, suitable for most contact problems solved with ANSYS, were used. One of the contact surfaces was conventionally established as the “target” surface and the other as the “contact” surface. So the contact cinematic was defined. The “contact” elements could not penetrate the target surface. The conductivity and thermal exchange varied with the temperature and the contact profile. However, the non-linear target element penetration and the contact behaviour had to be controlled. ANSYS employs the “Newton–Raphson” approach to solve non-linear problems like contact one. Before each solution, the Newton–Raphson method evaluates the out-of-balance load vector, which is the difference between the restoring forces (the loads corresponding to the element stresses) and the applied loads. The program then performs a linear solution, using the outof-balance loads, and checks for convergence. If convergence criteria are not satisfied, the out-of-balance load vector is reevaluated, the stiffness matrix is updated, and a new solution is obtained. This iterative procedure continues until the problem converges. The model’s outputs were the temperature along the seam and the displacements of the joint during and over the discharge time.


Results for the Ti6Al4V

Fig. 4 shows the displacement of the nodes at the top of the upper electrode along the sample axis and in the upsetting force direction. The lower electrode bottom nodes were fixed. A rising in the discharging time and in the current peak had the effect of increasing the electrode’s displacement. Fig. 5 displays the temperature profile for the case with 6 ms discharge time, 80 kN upsetting force, and 69 kA current peak.

Fig. 4 – Displacement against discharge time for the Ti6Al4V samples.


j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 1 9 7 ( 2 0 0 8 ) 301–305

Fig. 5 – Temperature distribution along the seam for the TiAl64V sample.

The melting temperature about 1650 ◦ C was reached only in the proximity of the peaks of the profile even if the upper and lower electrodes were in contact after the application of the upsetting force. Along the rest of the profile the temperature was largely lower than the melting one. In fact a few welds were carried out and they confirmed the melting of the material in the place where the calculated temperature was above 1650 ◦ C.


Results for the INCONEL 718

The displacements of the upper electrode were in the same order of those of the Ti6Al4V. Fig. 6 shows the displacements against the current peak for different discharge time. The upsetting force was 135 kN. Fig. 7 shows the temperature distribution along the seam of the weld for an INCONEL 718 simulation whose inputs were 6 ms discharge time, 125 kN upsetting force, and 110 kA peak current.

Fig. 7 – Temperature distribution along the seam for the INCONEL 718 sample.

Again it is a lower-than-melting temperature along the protrusions’ sides. Nevertheless, the applied upsetting force and peak currents had the effect to bring the protrusion peak areas well above the melting temperature, which for the INCONEL 718 is about 1300 ◦ C.



The FEM numerical model took into account the complicated electrical, thermal and mechanical phenomena that occur during the high speed pulse welding. The electrical and thermal phenomena were directly coupled during the calculations and the mechanical one was coupled to the previous via an incremental procedure. The sensitiveness of the model towards the peak current, the discharge time, and upsetting force on the upper electrode was proved. The displacement of the upper electrode and temperature distribution in the weld proximity were calculated. The results of the simulation showed the inadequacy of the waved profile to produce a whole seam weld even for tiny pieces.


Fig. 6 – Displacement against the current peak (kA) for the INCONEL 718 samples.

ANSYS 8.0. 80 Quick Start Licensing Guide.htm. Casalino, G., Panella, F.W., 2006. Multipoints capacitator discharge welding of AISI 304 bars. J. Eng. Manuf. 220, 647–655. Casalino, G., Ludovico, A.D., Dattoma, V., Panella, F.W., 2003. In: Katalinic, B. (Ed.), The Use of Discharge Energy for Rapid Welding. DAAAM Int. Scientific Book, Vienna, pp. 113–124. Karlsson, L., 1997. Modelling in Welding, Hot Powder Forming, and Casting. ASM International, Materials Park, OH.

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 1 9 7 ( 2 0 0 8 ) 301–305

Kristiansen, M., 1995. Pulsed power applications. In: Conf. Rec. JIEE-IPEC ’95, Yokohama, Japan, April, pp. 1391–1396. Maykuth, D.J., 1980. Aerospace Structural Metals Handbook. Battelle Columbus Laboratories, USA. Messler Jr., R.W., 1995. The challenge of joining to keep pace with advancing materials and design. Mater. Des. 16 (5), 261–269.


Sun, X., 2000. Modelling of projection welding processes using coupled finite element analysis. Weld. J. Res. Suppl. 79, 244–251. Wilson, R.D., 1994. Explore the potential of capacitor discharge welding. Adv. Mater. Process. 145 (6), 93–94.