TIG and A-TIG welding experimental investigations and comparison with simulation

Flux preparation and material conditions

Before flux application and welding, the workpiece was cleaned with acetone and dried. In the present work, different fluxes were used. The first was a commercial flux composed of TiO₂, Cr₂O₃, V₂O₅, MgF₂ and MgCl₂. The second flux contains the mixture of three oxides, such as TiO₂, Cr₂O₃ and V₂O₅, and TiO₂, Cr₂O₃, K₂Cr₂O₇, MgF₂ and BaF₂ were also studied separately. A planetary crusher was used to grind 15 g of powder (oxides and halides) with 15 mL of ethanol in a stainless steel bowl for 20 min at 360 rev min⁻¹ in order to prepare the activating flux. The flux was deposited on the steel plate before welding by pulverisation using an aerograph. The geometry of the fusion line was produced on austenitic stainless steel 304L rectangular plates (200×50×4 mm). The chemical composition of steel is 0·013C–0·53Si–1·61Mn–0·017P–0·011S–19·67Cr–0·08Mo–9·95Ni–0·005Al–0·072Co–0·11Cu–0·019Ti–0·046V (wt-%).

Welding was conducted using a Fronius Magic Wave 2200 current supply and a Servisoud automatic welding bench for dynamic experiments.

Arc plasma visualisation

During welding, the workpiece was continuously displaced using a movable table, whereas the plasma torch was fixed. Then, all the optical observation can be performed on a fixed spatial position. The ‘shape’ of the plasma column and the arc constriction phenomena were monitored by a high speed camera FASTCAM-1024PCI HW (Photron Technology) with an interference filter of 469·2 nm central wavelength and 3 nm spectral bandwidth. Recorded signals were dominated by continuum radiation originating from free electrons. The plasma images recorded in this way can give a lot of information on the welding process since the radiation directly follows the current.5 Two different camera positions were used, as shown in Fig. 2, to get face (x direction) or side (in y direction) view of the plasma column, in order to see if the arc constriction occurs in two directions.

OPEN IN VIEWER

Figure 2

Experimental conditions for arc constriction identification

Investigations of plasma by optical emission spectroscopy

Optical visible emission spectroscopy was used to characterise the arc plasma and to determine the arc temperature during welding in TIG and A-TIG with commercial fluxes and MgF₂.

During welding, the workpiece was continuously displaced using a movable table. The distance z₀ between the contact tip and the workpiece (the metal plate) was fixed, and the investigations were accomplished using the light emitted from different layers z (origin taken on the workpiece) of the plasma.

The investigated plasma region was imaged onto the entrance slit of an Ebert type spectrograph (0·2 nm mm⁻¹ reciprocal dispersion) by a system of flat and concave mirrors and a Dove prism. As the entrance slit of the spectrometer is vertical, the Dove prism rotated the plasma image by 90°, enabling the registration of the plasma layer at the given distance z from the electrodes. Investigations of the light emitted from different layers of the plasma column were accomplished by vertical translation of a set of two mirrors, which causes translation of the optical axis with no angle modification.

The spectra were collected at wavelengths of 405·5 and 706·8 nm over the spectral range of ∼1·6 nm using a gated two-dimensional intensified charge coupled device camera. The spectral resolution of the optical system varied from 0·0044 to 0·0026 nm per pixel, while the spatial resolution was 50 μm vertically (along the axis of the plasma column) and 45 μm horizontally.

Arc constriction phenomena

To determine the influence of oxides and halides on the arc constriction, a commercial flux constituted of TiO₂, Cr₂O₃, V₂O₅, MgF₂ and MgCl₂ was used as well as the mixture of oxides TiO₂, V₂O₅ and MgF₂. Different oxides such as TiO₂, Cr₂O₃ and K₂Cr₂O₇ were chosen to study their influence on the physics of the arc. MgF₂ and BaF₂ were also studied separately.

The arc shapes (face view) monitored during TIG and A-TIG welding for each flux are reported in Fig. 3. We can observe that the arc constriction phenomenon has occurred while passing from TIG to A-TIG for each flux. However, fluxes act differently on the physics of the arc. Oxides or fluorides used separately have an important influence on the arc compared with the two fluxes composed of the mixture of oxides and halides.

OPEN IN VIEWER

Figure 3

Arc shapes monitored during TIG and A-TIG welding for different fluxes (face view)

The arc constriction phenomenon is different according to the type of oxide or fluoride. When comparing the arc shape during A-TIG welding with the two fluorides, magnesium fluoride decreases the arc length, increasing the penetration depth of the weld beads.

We can see this decrease in the arc length added to the arc constriction for TiO₂ and K₂Cr₂O₇. However, the fluxes constituted of Cr₂O₃ also have an important constriction effect. The results obtained for K₂Cr₂O₇ contradict with the geometry of the weld beads realised in A-TIG with this same flux. Indeed, K₂Cr₂O₇ is not present in commercial fluxes because its activating effect is poor.

The arc shapes (side view) monitored during TIG and A-TIG welding for each flux are reported in Fig. 4. The arc constriction is less important in this case. These observations show that the arc constriction acts only perpendicularly to welding direction. However, the width of the arc is narrower in A-TIG welding, and the decrease in the arc is still present for MgF₂ and K₂Cr₂O₇ but also for BaF₂.

OPEN IN VIEWER

Figure 4

Arc shapes monitored during TIG and A-TIG welding for different fluxes (side view)

To quantify the arc constriction, the ratio between arc width while passing from TIG to A-TIG was calculated for each flux and reported in Fig. 5. The maximum width of the electrical arc in TIG and A-TIG welding was determined by measuring the white region of the photo, and the formula used for the ratio is L_ATIG/L_TIG (Fig. 3). The arc constrictions while passing from TIG to A-TIG for commercial flux and the flux constituted of TiO₂–V₂O₅–MgF₂ are respectively 45 and 40%. The percentage of arc constriction during A-TIG welding with titanium oxide or chromium oxide is more important than a mixture of oxides and halides (55% for TiO₂ and Cr₂O₃). We obtain 56% arc constriction with K₂Cr₂O₇, but this percentage does not represent really in totality but also the difficulty of the arc to be established with this flux.

OPEN IN VIEWER

Figure 5

Arc constriction while passing from TIG to A-TIG welding process for different fluxes

The arc constriction while passing from TIG to A-TIG with MgF₂ and BaF₂ is less important than oxides (40% for MgF₂ and 30% for BaF₂). However, the addition of magnesium fluoride during A-TIG welding leads to arc constriction and also a decrease in the arc length. This is why the penetration depth of the weld beads realised in A-TIG welding with MgF₂ is deeper than the same process with the addition of one oxide, another fluoride or the mixture of them.

Energy density influence

The two-dimensional axial symmetrical model detailed in part I of this paper is used to study the effect of arc constriction on weld bead shapes. In this model, the heat source and electric current are represented respectively by a Gaussian heat source and Gaussian electric current applied on the sample surface (see equations (1) and (2)). In the present work, arc constriction is modelled by decreasing both the heat distribution coefficient r₀ and the current distribution coefficient r_c, which increase the heat and current densities near the axis (1) (2) where η represents the arc efficiency, U is the voltage and I is the electric current.

The parameters used for simulations are I = 100 A, U = 10 V and η = 75%. In the simulations, r₀ varies from 2·5 and 1·6 mm and r_c is from 3 to 1·55 mm with negative and positive Marangoni. Lorentz forces are taken into account or not. It is shown that the weld shapes change significantly with the variation of r₀ and depend on the Lorentz forces only for the negative case (Fig. 6a and b). Without Lorentz force, by decreasing the value of the heat distribution coefficient r₀ from 2·5 to 1·6 mm, the weld penetration is multiplied by a factor of 2 for the negative case and 4 for the positive case (Fig. 6b). The positive case is thus more sensitive to arc constriction with a significant increase in penetration depth. By taking into account the electromagnetic force, the weld shape of the negative case changes significantly with the value of r₀ due to the creation of a second vortex by decreasing r₀. With r₀ = 2·5 mm and r_c = 3 mm, the electromagnetic forces are not sufficiently high to create an inward flow, resulting in a shallow weld pool for the negative case. Hence, the arc constriction has two effects: by increasing the heat density, the surface temperatures of the weld pool will be higher, creating a wider and deeper weld pool, and by increasing the current density, the electromagnetic forces will be larger, resulting in a second vortex that will carry the heat energy towards the centre, causing a deeper weld penetration.

OPEN IN VIEWER

Figure 6

Comparison of weld pool geometry using r₀ = 1·6 or 2·5 mm for negative and positive and a with or b without Lorentz

It is noted that if the surface tension temperature gradient is positive, the electromagnetic force has little influence on the weld pool shape with regards to the variation of r₀ and r_c. However, it should be noticed that the surface tension temperature gradient is assumed here as a constant (positive or negative). In fact, if active elements such as sulphur are present in the material, the surface tension gradient can be negative at low temperatures but positive at higher temperatures, resulting in complex flow patterns. Wang et al. have shown that, depending upon the sulphur concentration, one, two or three vortexes may be found, caused by the interaction between the electromagnetic force and surface tension.6 These numerical results are in good agreement with experimental results and more particularly on the weld bead shape, which is more penetrated in A-TIG welding, as shown in part I of this paper.

Arc temperature

Optical visible emission spectroscopy is used in TIG and A-TIG welding with commercial flux and MgF₂ on austenitic stainless steel 304L. These experiments were carried out for an arc at 80 A and 3 mm of arc length, at different distances z of the workpiece in the plasma column: near the plate, in the middle of the arc and near the cathode.

The recorded spectra are laterally integrated (along the observation direction). The local, radially resolved spectra were obtained assuming the cylindrical symmetry of the plasma column and applying the Abel transformation procedure, as described in Ref. 7. Then, the Abel inverted spectra of emission lines are fitted with the Voigt profiles that are composed of Gaussian and Lorentzian profiles. Under our experimental conditions, the Lorentzian profile was completely dominated by Stark broadening, while the Gaussian profile was essentially generated by instrumental broadening: in our experimental conditions, the Doppler broadening, due to the plasma temperature, is weak and can be neglected. The emission line intensity I_nm is calibrated in energy for the given wavelength λ_mn using a tungsten ribbon lamp.

In the case of our plasma, there is no spectral line of flux elements except TiI. Near the welded material, only FeI spectral lines of metallic vapours were shown. This is why it is impossible to use the method proposed by Sola et al. 8 to determine simultaneously the electron density and the electron temperature without assumptions about the thermodynamic equilibrium of the plasma.7 There is no observation of maximum except the axis of the spectral line of ArI that enabled us to suppose that the temperature does not exceed the norm temperature (15 000 or 16 000 K) in argon plasma at thermodynamic equilibrium. There are little ArII spectral lines, mostly near the cathode, which supposes that the electron temperature does not exceed 18 000 or 20 000 K.

Some previous works have shown that the local thermal equilibrium (LTE) was obtain in plasma, at least at a sufficient distance to the cathode.9, 10 The electronic temperature has been calculated using the Boltzmann plot method, assuming LTE, with argon lines (mainly 415·8 nm ArI, 696·5 nm ArI and 706·8 nm ArI). This method is derived from the Boltzmann equation (3) where E_m is the upper level energy, g_m is the statistical weight, A_mn is the transition probability, λ_mn is the wavelength, I_mn is the emission line intensity, k_B is Boltzmann's constant, h is Planck's constant, c is the light velocity, N is the population density of the element and Z is the partition function. The slope of the representation on the left hand side of equation (3) versus E_m is inversely proportional to T_e and allows an estimation of the validity of the LTE hypothesis. Nevertheless, when the obtained temperature is too low (under 5000 K), LTE might not exist, and what we get with this method is only excitation temperature.

Figure 7 shows the average temperature calculated for different welding conditions (TIG, A-TIG with commercial flux and MgF₂).

OPEN IN VIEWER

Figure 7

Arc temperature during TIG and A-TIG (commercial flux and MgF₂) welding of 304L at 80 A, 3 mm of arc length, 15 cm min⁻¹ of welding speed near cathode, middle of plasma and near welded material

We can observe a decrease in the arc temperature from the cathode (9000 K in TIG) to the welded material (2700 K in TIG). This phenomenon is explained by the spatial repartition of the elements in the arc (in this work, argon). The results do not present an elevation of arc temperature during A-TIG welding with commercial flux at different heights (3065 K near the cathode and 5765 K in the middle of arc). The activating effect of commercial flux comes from the inversion of Marangoni convection in the melting pool and of the arc constriction. Figure 8 presents an elevation of arc temperature during A-TIG welding with MgF₂ at different heights of plasma (11 800 K in A-TIG for MgF₂ and only 9000 K in TIG near the cathode). This clearly proves that the activating effect of magnesium fluoride comes from an elevation of arc temperature and its constriction combined with a decrease in the arc length.

OPEN IN VIEWER

Figure 8

Arc temperature during TIG and A-TIG (commercial flux and MgF₂) welding of 304L at 80 A, 3 mm of arc length, 15 cm min⁻¹ of welding speed, at 80 A, 5 mm of arc length, 15 cm min⁻¹ of welding speed and at 120 A, 3 mm of arc length, 15 cm min⁻¹ of welding speed near welded material

In order to study the influence of welding current and arc length on the arc temperature, the experiments were repeated in TIG and A-TIG welding with commercial flux and MgF₂ at 80 A and 5 mm of arc length and at 120 A and 3 mm of arc length. The last results obtained at 80 A and 3 mm are used as reference. The average temperatures of plasma near the welded material for each welding condition are reported in Fig. 8. Some researches have shown the relationship between arc tension and arc length. The more that the arc length is higher, the more that the arc tension is elevated. Other researches have estimated an elevation of the arc tension while passing from TIG to A-TIG welding.11^– 13

Results from Fig. 8 present an elevation of the arc temperature with an increase in the arc length or an increase in the welding current. Indeed, during TIG welding, the arc temperature increases from 2700 to 4590 K while passing from 3 to 5 mm of arc length and in A-TIG welding with MgF₂ from 3300 to 4220 K. We can observe that an increase in welding current leads to an elevation of arc temperature. In A-TIG welding with MgF₂, the arc temperature is 3300 K at 80 A and 3 mm of arc length, whereas it is 4080 K at 120 A and 3 mm of arc length. The penetration and width of the weld bead are more important, while the welding current increases because the arc temperature increases too, leading to a surface tension gradient elevation.