CRA Weld Overlay - Influence of welding process and parameters on dilution and corrosion resistance

Paper presented at Stainless Steel World America 2010, Houston, Texas, USA, 5-7 October 2010.

V Kumar, C Lee, G Verhaeghe, S Raghunathan,
TWI Limited, Granta Park, Great Abington, Cambridge CB21 6AL, United Kingdom

Abstract

Corrosion-resistant weld overlays are used to improve the service life of components made with an otherwise corrosion-prone material. One of the major concerns in arc welding based overlays is dilution. Codes and Standards for qualification of procedures for weld overlay, such as ASME Section IX (2010), state that the heat input for the first weld layer is an essential variable, and a change in heat input over 110% of that qualified, requires requalification. The same heat input can be achieved by proportionally varying the welding current and the welding speed, but with an entirely different effect on dilution. Hence the above stipulation does not seem to be adequate in ensuring the 'chemistry' of the weld overlay and its integrity. Due to uncertainties involved in the quality of the weld overlay, a conservative approach is often taken while specifying the permissible dilution, resulting in substantial productivity losses, increase in cost, and associated issues such as distortion of components. Improved process control has been achieved with new generation arc welding equipment through digital control, giving the benefits of reduced heat input, improved arc stability and spatter-free welding, hence better consistency in quality. This investigation tried to understand the effect of welding process and major welding parameters on dilution and its effect on corrosion resistance, by manufacturing and testing Alloy 625 weld overlay with different degrees of dilution, on a C-Mn steel substrate material. The extent of dilution was measured in terms of the amount of iron (Fe) in the weld metal using semi-quantities energy dispersive X-ray (EDX) analysis and the corrosion resistance was evaluated using 'droplet cell corrosion testing' technique in a selected test environment. This investigation has shown that, in arc welded overlays, for the same heat input the dilution can vary over a wide range depending on the welding process and the welding parameters, and there was no apparent reduction in corrosion resistance for an iron content up to a certain level, and beyond this the corrosion resistance decreased drastically.

Introduction

Corrosion-resistant weld overlays are often used to improve the service life of components made with an otherwise corrosion-prone material. A major concern in an arc welding based overlay is dilution or the extent of change in the chemistry of the deposited metal by the mixing of base metal. Even though some generic information is available on the extent of dilution associated with common arc welding processes, the actual dilution with a particular process itself can vary over a wide range, based on the welding parameters employed. In most cases of overlaying, it is necessary to control the dilution within close limits as an uneven chemistry can reduce the service life. There are a number of variables which affect dilution such as the welding current, the arc voltage, current polarity, electrode diameter, electrode extension, weld-bead separation, welding speed, electrode grinding angle, welding position, shielding gas composition, etc. It is necessary to control each of these variables within limits to get the desired properties on the overlay, for which it is necessary to have a clear understanding of the influence of each of these variables on dilution. Codes and standards such as ASME Section IX for qualification of welding procedures states that heat input for the first weld layer is an essential variable, i.e., a change in heat input over 110% of that qualified requires requalification. The same heat input can be achieved by proportionally varying the welding current and the welding speed. For many processes it will have an entirely different effect on the penetration depth and hence the dilution. The extent of overlap between adjacent weld beads also is a key variable influencing the dilution, in many cases more than the heat input.

Weld overlay can be produced with a number of arc welding processes. Manual metal arc (MMA) welding, submerged arc welding (SAW) both with wire and strip, metal inert/active gas (MIG/MAG) welding, and tungsten inert gas (TIG) welding (hot wire and cold wire) processes are commonly employed. Improved process control has been achieved in new generation MIG welding equipments through digital control giving the benefits of reduced heat input, stable arc, uniform weld profile and a spatter free welding. There are electronically controlled short-circuit MIG welding process variants as well as high-deposition TIG welding variants giving the benefits of high-deposition rate and good control on dilution.

In conventional TIG welding, the arc provides the entire energy required for heating and melting the filler wire, hence the metal deposition rate is limited by the rate at which the heating and melting process can take place. The energy consumption from the arc can be reduced if the wire can be fed into the arc region at a higher temperature. That means the arc energy can be used elsewhere, for example for increasing the parent metal melting, or a higher quantity of filler material may be melted using the same arc energy as compared to conventional TIG welding. This principle is used in hot wire TIG welding process which employs preheating the wire before it enters in the arc region. This heating is accomplished by resistive heating of the wire between the feeder nozzle and the molten pool. Since the wire always need to be in contact with the weld pool, for a given wire feed rate there is a limit on the maximum preheating that can be applied.

TIG welding variant known as TOPTIG welding is a process patented by Air Liquide. Unlike conventional TIG welding, in TOPTIG the filler wire feeding nozzle is integral with the welding torch so that the wire could be introduced in the hottest region of the arc, enhancing the melting efficiency. It has been claimed that on steel sheets of up to 3mm thick, a welding speed comparable with that of MIG welding could be achieved by this process, still maintaining the quality of the TIG process. The advantages claimed for this process include comparable productivity to MIG welding, lower heat input, lesser distortion, less sensitivity to the orientation of the wire feeding with respect to the welding direction, and greater flexibility with respect to heat input and deposition.

In MIG welding the arc is formed in an inert atmosphere between a continuously fed consumable wire electrode and work piece. Unlike TIG welding, the current and deposition rate can not be controlled independently. The heat input in MIG welding depends on the metal transfer mode, which can be classified into short-circuit transfer, globular transfer, spray transfer and pulse transfer. The operating parameters such as the arc voltage, current, shielding gas, and electrode wire feed rate control the transfer mode.

Short-circuit transfer mode provides the lowest heat input in MIG welding. However, a major problem with conventional short-circuit transfer is excessive spatter. The digitally controlled power sources allow a greater degree of control over the voltage and current wave forms on a real time basis than that possible with the traditional ones employing analogue controllers. Digital technology has been used in systems such as the Lincoln Electric STT^TM (surface tension transfer) technology, Daihen Corporation CBT (controlled bridge transfer) technology, EWM-coldArc^® technology, etc, to achieve a short-circuit transfer condition without excessive spatter. The digitisation of the controllers also improved the dynamic response of the power sources, resulting in the generation of self-tuning power sources such as the ESAB Q-Set.

This investigation tried to identify the effect of welding process and major welding parameters on dilution, and the effect of an increased dilution on corrosion resistance for a selected environment, by manufacturing and testing Alloy 625 weld overlays with different degrees of dilution on a carbon steel substrate material.

Objectives

The objective of this investigation was to have a greater understanding on the influence of major welding parameters namely the heat input, the welding current, and the welding speed on dilution and corrosion resistance, for MIG and TIG welding process and their variants.

Experimental details

The following welding processes have been investigated in this project:

• Conventional TIG welding
• Hot wire TIG welding
• TIG welding variant known as TOPTIG
• MIG welding with globular transfer, pulse transfer, and spray transfer
• MIG welding with electronically controlled short circuit transfer (STT)

BSEN 10025 2004 355J2+N grade carbon steel plates of size 200x100x15mm were used as the substrate material for weld overlay experiments. The plates were clamped on a rigid fixture and mechanised/automated welding experiments were carried out in the down-hand (PA position) by moving either the welding torch or the work piece relative to the other. The welding torch was held at 90° to the substrate surface. Pure argon was used as the shielding gas. Alloy 625 filler wires of diameter 1.1mm and 1.2mm meeting AWS A5.14 ER NiCrMo-3 were used in all experiments. The welding current and the voltage were measured at a sampling frequency of 4kHz, and recorded at an interval of 1-5s depending upon the welding speed, using an AMV weld check arc monitor. The welding speed was calculated separately from the readings of the weld length and the welding duration. The heat input was calculated as the product of the average values of arc voltage, current, and arc efficiency factor (0.6 for TIG welding and 0.8 for MIG welding) divided by the welding speed. Single weld-bead and multiple weld-bead experiments were carried out, and transverse weld sections were prepared from the stable portion of each weld-bead. The transverse weld sections were ground and polished to a one micron diamond finish and etched in 2% Nital for optical examination and photography.

Weld metal composition was analysed using semi-quantitative EDX analysis to determine the fraction of iron in the weld. The analysis was carried out at locations corresponding to the central region of the deposit. Different processes were compared on the basis of weld metal dilution and heat input. Finally the effect of dilution on corrosion resistance was evaluated for selected number of overlays using 'droplet cell corrosion testing' technique in a 10w/v% NaCl at pH3 environment. Samples were polished to a 6µm diamond finish prior to testing. Anodic potentiodynamic polarisation tests were performed using a Uniscan Instruments Scanning Electrochemical Workstation Model 370 with scanning droplet cell attachment. A silver/silver chloride reference electrode and platinum wire auxiliary electrode formed a three-electrode cell, with the sample as the working electrode. The electrolyte was a continuously refreshed 10%w/v NaCl solution acidified to pH 3. Anodic polarisation tests were performed after a stabilization period of 10 minutes, commencing 300mV cathodic from the open circuit potential (OCP) and scanning to 1.2V anodic with respect to the reference electrode. A reverse scan was then performed from this potential back to 300mV cathodic with respect to OCP. A scan rate of 20mV/min was used throughout.

Results

Photomacrographs of typical transverse weld sections of single weld beads obtained by varying the welding current, welding speed are shown in Figure 1 and 2.

Figure 1. Photomacrographs of transverse sections of single-weld-bead produced by increasing the current from 140-300A in steps of 40A. Note the different magnifications in the photographs.

Figure 2. Photomacrographs of transverse sections of single-weld-bead produced by increasing the welding speed from 40, 60, 100, 120, and 140mm/min. Note the different magnifications in the photographs.

The variations of weld-bead size with increase in welding current are shown in Figure 3. The width of the weld-bead increased almost linearly with the welding current for constant values of the other welding parameters. The depth of penetration also showed an increase, whilst the weld height slightly decreased with increasing welding current. The rate of increase in bead width was significantly greater than the rate of increase or decrease in the penetration depth or the weld height.

Figure 3. Variation of weld-bead dimensions with welding current whilst other parameters remaining constant. Single weld-bead made with conventional TIG welding at 80mm/min welding speed and at a wire feeding speed of 910mm/min.

The variations in weld size with welding speed are shown in Figure 4. The weld width decreased with increasing welding speed. The penetration depth remained the same, almost independent of the welding speed. The weld height showed a decrease with increasing welding speed.

Figure 4. Variation of weld-bead dimensions with welding speed whilst other parameters remaining constant. Single weld-bead made with conventional TIG welding at 220A welding current and at 910mm/min wire feeding rate.

The dilution in terms of iron content of the weld metal with a change in welding current is shown in Figure 3. Dilutions of the first and third beads were compared. The dilution in the first bead was substantially greater than the subsequent beads. Dilution less than 15% was obtained at current levels lower than 150A, whereas the dilution was in excess of 50% for welding current in excess of 230A showing the uncertainties with weld dilution for TIG welding process.

The variations in dilution with welding speeds for TIG welding are shown in Figure 4. The dilution increased with increasing welding speed. However the rate of increase was significantly lower than that with varying welding current. Towards the higher end of the welding speeds, the dilution showed a stabilising tendency with increasing welding speed.

The variations in dilution with change in heat input for TIG welding are shown in Figure 5. The two sets of graphs correspond to the variations in the heat input obtained through varying the welding current at constant welding speed or through varying the welding speed at constant welding current. The dilution increased with increasing heat input when the increase in heat input was achieved through an increase in welding current at constant welding speed. However the dilution decreased with increasing heat input when the increase in heat input was achieved through a reduction in welding speed at constant welding current. There was significant difference in dilution, for the same heat input depending on whether the heat input was achieved through control of the welding current or the welding speed. For example at a heat input of 750J/mmm, dilution in the range 15-25% was achieved by controlling the welding current whilst 50-60% dilution was obtained by controlling the welding speed.

Figure 5. Variations in the dilution of the 1st and 3rd bead with heat input obtained by varying the welding current denoted by 'I' and welding speed denoted by 's' in the legend, with other welding parameters remaining constant.

Figure 6. Variations in the dilution of the 1st and 3rd bead with heat input obtained by varying the welding current denoted by 'I' and welding speed denoted by 's' in the legend, with other welding parameters remaining constant.(MIG welding)

Variations in dilution with heat input for MIG welding are shown in Figure 6. The variations were similar to those observed in TIG welding. The dilution increased with increasing heat input when the increase in heat input was achieved through an increase in welding current at constant welding speed. However the dilution tend to decrease with increasing heat input when the increase in heat input is achieved through a reduction in welding speed at constant welding current.

Figure 7 shows the dilution versus heat input for the first weld bead for different welding processes. Both conventional TIG and hotwire TIG produced largest scatter in heat input and dilution. Electronically controlled short circuit transfer process consistently produced a dilution less than 20% (first bead). At similar levels of heat input, the dilution in TopTIG process depended on the heat input control method. For higher values of current and welding speed, the dilution was significantly greater (~50%) than those (<15%) produced at lower values of current and welding speed, keeping the heat input same. Dilution levels with hotwire TIG was similar to that produced in conventional TIG welding process.

Figure 7. Dilution versus heat input for the first weld-bead for different welding processes.

Figures 8 and 9 show the breakdown potentials plotted against the degree of dilution represented by the %Fe in the composition, and against the pitting resistance equivalent number (PREN). The PREN is an index of the relative corrosion resistance of the material as calculated by the following formula:

PREN = %Cr + 3.3%Mo + 16%N

Figure 8 shows there was a marked reduction in the breakdown potential between 36 and 42%Fe which corresponds to a reduction of PREN from 33 to 29, in Figure 9.

Figure 8. Relationship between the breakdown potential and dilution (%Fe) in weld metal

Figure 9. Breakdown potential plotted against PREN. The PREN of commonly used wrought grades of corrosion resistant alloys (CRAs) are also shown for comparison (horizontal bar represents the alloy composition range and shows typical compositions).

Discussion

Dilution is defined as the change in composition of the weld metal caused by the mixing of the base metal or the previously deposited weld metal. In this investigation the effect of welding parameters on dilution was studied by taking conventional TIG and MIG welding processes. The dilution was measured at locations corresponding to the central region of the weld metal of each welding pass considering the fact that the agitation of the weld metal during welding would make the composition approximately uniform throughout the weld metal except for regions which are very near to the fusion boundary.

Significant differences in dilution were observed between the first weld bead and the subsequent beads. This was because, in the first bead, the arc directly strikes on the virgin base metal resulting in significantly higher melting of the parent metal. In subsequent overlapping weld passes, a part of the initial weld bead would be melted reducing the extent of parent metal melting, and a steady state would be reached after certain number of passes. Initial TIG welding experiments showed that the dilution reached approximately steady state conditions on the third bead, hence for the subsequent welding experiments; the dilutions of the first and third weld beads were compared.

Results of initial TIG welding experiments by varying the welding current and by keeping the other welding parameters constant showed that the weld width and depth increased with increasing the welding current. A corresponding decrease in the weld height was also observed. These results suggest an increased substrate melting hence a higher dilution. MIG welding also showed a similar trend.

The heat input per unit length increases with an increase in welding current, and decreases with increase in welding speed. That means the same heat input can be obtained either by using a low welding speed and a low welding current or using a high welding speed and a high welding current. For productivity benefits, the manufacturers tend to go for the latter. For the same heat input the higher current and higher welding speed produced substantially higher weld dilution than a lower current and a lower welding speed. For example, 148A at 80mm/min welding speed and 208A at 120mm/min welding speed results in approximately the same heat input. However the dilution at 148A welding current was only about 20% compared to 50-60% measured at 208A welding current. Results of MIG welding experiments also showed similar tendency even though in this case a higher welding current was associated with a higher deposition rate. These results question the very purpose of specifying only the heat input as the most critical variable by codes and standards in weld overlaying procedures.

Of the various welding processes studied in this investigation, controlled short-circuit transfer welding process produced the lowest level of weld dilution for a given heat input. However, the peculiar weld shape with large height and high ‘contact angle’ increased the chance of producing lack of fusion type defects at the weld toe when subsequent weld beads were deposited. So the positioning of the torch was extremely critical with this process. TopTIG welding process also provides very low dilution at low values of welding current and welding speed. TIG welding produced the largest scatter in weld dilution.

Current qualification procedures for weld overlay cladding involve measurement of bulk dilution and ensuring that it is below a maximum level and also passing a standard generic corrosion test such as ASTM G48. There are two drawbacks to this approach: firstly, the chemical analysis results gives a bulk average of the sample, locally, a higher or lower dilution levels may exist; secondly, the ASTM G48 test is aggressive and not representative of most service environments. Thus, it provides limited understanding of the effects of variations in welding procedure on actual service performance of the cladding.

A droplet cell technique was selected for use in this project in order to measure the corrosion resistance of local regions. This technique uses a small droplet of solution placed on the test surface, and by carrying electrochemical anodic polarization within this droplet, it is able to determine corrosion resistance of the material in a small region. This overcame the problems of measuring the corrosion resistance of a large surface area where the dilution levels may vary significantly.

The results from the droplet cell tests showed the corrosion resistance of Alloy 625 weld overlay does not decrease until dilutions levels greater than 36%Fe is reached in selected test environment. This indicates that dilutions levels up to 36%Fe may be tolerated if the service environment which is less aggressive than or similar to the test environment. Below these dilution levels, no pitting was observed.

Furthermore the results show a clear step change and not a gradual change in corrosion resistance in relation to dilution. It would be interesting to repeat these tests in a more aggressive environment, and to determine if the same step change were still observed but shifted to lower dilution levels or if the change were more gradual. Previous work, exposing 'bulk' weld overlay samples to an aggressive environment, showed that corrosion resistance decreased with increasing dilution up to 30%Fe^{[Gittos and Gooch, 1996]}. However, these data also showed significant scatter which may be an indication of the range of local dilutions in the areas sampled.

Although a dilution level is often specified for qualification procedures, corrosion resistance of CRAs is the often expressed as pitting resistance equivalent number (PREN) which takes into account all major alloying elements which affect corrosion resistance. The results show that the drop in corrosion resistance occurs at a PREN of <33, Figure 9. This corresponds to the lower end of alloy 904L and within the composition range of Alloy 825. However, it should be noted that the results are from the droplet cell test conducted on weld metal, essentially an as-cast alloy, but nevertheless, it shows that the Alloy 625 weld overlays need to reach dilution equivalent to those of the Alloy 825 composition range before a significant drop in corrosion resistance is observed in this environment.

Conclusions

Effect of welding parameters on dilution was studied by using TIG and MIG welding processes, and the performance of some of the process variants were compared by conducting weld overlay experiments on 15mm thick carbon steel material using 1.1mm diameter Alloy 625 filler wire. Following conclusions can be made on the basis of this experimental investigation.

• TIG and MIG welding processes produce weld overlay with a wide range of dilution. TIG welds produce the highest scatter in dilution; depending on the process parameters the dilution can be as low as <10% or as high as 70%. In MIG weld the dilution was generally less than 40%.
  
  
• In arc welding, the dilution generally increased with increase in welding current and welding speed.
  
  
• An increase in heat input with increase in current increased the dilution whilst an increase in heat input due to decrease in welding speed reduced the dilution.
  
  
• For a given heat input electronically controlled short-circuit transfer process provide the lowest dilution (<5%). TIG welding variant also could produce weld overlay with very low dilution (<10%) with single layer.
  
  
• The corrosion resistance of alloy 625 weld overlay exhibited a marked drop in corrosion resistance at dilution levels >36%Fe, in the selected test environment of 10w/v% NaCl at pH3.
  
  
• The dilution levels of the marked drop in corrosion resistance were equivalent to the PREN of wrought of alloy 825, in the selected test environment of 10w/v% NaCl at pH3.
  
  
• The above results suggest that restricting the heat input alone may not necessarily ensure the corrosion resistance and there appears to be opportunities for relaxing the present specified limit on dilution.

Further work

There are a number of digitally controlled MIG welding variant, and high-deposition TIG welding variants available to improve the deposition capabilities. Also, the variable polarity sub arc welding provides the capability to achieve different deposition rates and penetration depths without changing the voltage or current. Hence the required dilution limits or deposit height may be achieved with a fewer layers than that is possible with conventional processes. Also, there appears to be opportunities for relaxing the present permissible limit for %Fe in the weld overlay; a higher limit on dilution can improve the economics associated with weld overlaying significantly.

In order to address some of these issues, TWI will be launching a joint industry project (JIP) in autumn 2010. The project will address these issues through a review of current manufacturing practices, investigating improved manufacturing process and procedures, investigating alternative consumables, evaluating the local and bulk corrosion resistance, and developing reliable inspection procedures. It will also address the effect of a relaxed dilution limit on corrosion fatigue performance of the weld overlay. For further information on this project, please contact: Vinod.kumar@twi.co.uk

Acknowledgement

This work was funded by the Industrial Members of TWI, as a part of the core Research Programme. The authors acknowledge the support of Chris Hardy, Harry Froment, in carrying out the welding trials, and Sheila Steven and Ashley Spencer for their support in preparing and analysing large number of samples. The authors are also grateful to Dr Oliver D Lewis of Sheffield Hallam University for his support in carrying out the 'droplet cell corrosion' tests.

References

ASME Boiler and Pressure Vessel Code, ASME, 2010.

Gittos M F, Gooch T G, 1996: 'Effects of iron dilution on corrosion resistance of Ni-Cr-Mo alloy cladding'. British Corrosion Journal, Vol 31, No.4, pp309.