Study on the deposition accuracy of omni-directional GTAW-based additive manufacturing

https://doi.org/10.1016/j.jmatprotec.2020.116649Get rights and content

Abstract

In gas tungsten arc welding (GTAW) based additive manufacturing (AM), omni-directional deposition with side feeding is common when depositing complex parts, which is different from the gas metal arc welding (GMAW). While side feeding may lead to unstable deposition process and deposition deviation. In this paper, a wire melting simulation model was established to analyse the behaviour of the wire in the arc column. An index of weld bead offset tolerance capacity is proposed to quantitatively analyse the sensitivity of the weld bead offset to the wire feed speed. Single-layer experiments were conducted to analyse the relationships between the deposition parameters and the weld melting/bead offset. A multi-layer sample with an actual usable area ratio of 95.11 % was deposited by using the proposed model and the optimized deposition parameters. The experimental results show that the control of the weld melting offset is the key factor in realizing the stability and accuracy of omni-directional GTAW-based AM.

Introduction

Compared with laser additive manufacturing, wire and arc additive manufacturing (WAAM) has higher deposition efficiency and lower cost and has attracted increasing attention and research, as investigated by Williams et al., 2016. WAAM is based on the fusion-based welding technologies, the research by Oliveira et al., 2020 shown that the experience in welding can be used to better understand the additive manufacturing process and the implications on the microstructural features of the produced parts. However, these experiences present difficulties in guiding geometric deposition capabilities of WAAM due to the complex deposition path.

It is known that WAAM deposition process is affected by many factors, such as welding heat input, wire feed speed, inter-layer temperature, deposition path, etc. Xiong and his co-workers (Xiong and Zhang, 2014; Xiong et al., 2015, 2016) studied forming characteristics, such as deposit velocity and heat input, of multi-layer single-bead parts. Variable layer width and height were obtained by an adaptive control system using a passive vision sensor. As for multi-layer multi-bead parts, the centre distance between the adjacent weld beads is critical for flat surfaces. Cao et al., 2011 used a sine function to fit the bead section and pointed out that the optimal centre distance is 0.6366 bead width. Nevertheless, Ding et al., 2015 argued that 0.738 bead width performs better using a parabola function. Based on this, Li et al., 2018 considered the phenomenon of bead spreading, and a smoother surface is obtained by compensating for the centre deviation of the adjacent weld bead. However, these studies are all based on gas metal arc welding (GMAW), which has a different wire feeding behaviour with gas tungsten arc welding (GTAW).

GMAW-based AM with coaxial wire feed is easier to realize the deposition of complex parts. Kazanas et al., 2012 investigated the production of geometrical features using GMAW-based AM, inclined and horizontal wall features had been achieved using an inclined torch position. However, omni-directional GTAW-based AM with side feeding has difficulties in achieving the complex parts, where: omni-direction means any torch moving direction with fixed wire feeding direction. The typical samples that can be deposited by GTAW-based AM are simple "walled" and "rotary" shaped samples. Ouyang and his co-workers (Ouyang et al., 2002; Wang et al., 2004) achieved hollow cylindrical parts with very low surface roughness by monitoring the arc length, and adjusting the arc current during the deposition process. In this case, the substrate was fixed on a rotating axis. Liu and his co-workers (Wu et al., 2017b; Guo et al., 2016) investigated low heat input deposition using pulse current, and horizontal rotary structures without supports were fabricated. During the deposition process of these samples, the angle between the feeding direction and the direction of motion remain unchanged. However, this angle will change during the process of omni-directional GTAW-based AM. Wu et al., 2017a revealed that the change of this angle may lead to deposition failure. An omni-directional deposition method was proposed by holding the welding torch at an angle to the workpiece, which helps realize smooth wire feeding in the omni-directional GTAW-based AM. Our previous research (Wang et al., 2019b) also studied the process stability of GTAW-based AM. The results shown that the optimized wire feed geometry parameters and arc length control system can effectively guarantee the stability of the omni-directional GTAW-based AM process. What’s more, This method also shown excellent stability in complex deposition paths, as investigated by our previous research (Wang et al., 2019a).

However, these methods mentioned above do not take into account the accuracy of the deposition process. Omni-directional GTAW-based AM experiments shown a deposition deviation phenomenon when side feeding, as show in Fig. 1. Deposition deviation exists when the molten wire is not evenly distributed in the weld pool before solidification. Although this uneven filling does not affect the continuation of the deposition process, it will result in a stable weld bead dimensional deviation. This problem is not easy to be detected because the welding torch always moves along the established path and the deposition process is stable.

Geng et al., 2017 studied the effect of wire feed behaviour on the deposition accuracy at the arc striking position. It is proposed that the dimensional deviation at the arc striking position can be eliminated by obtaining the wire bridging transfer and compensating at the arc striking position. However, this method is only for the case in which the wire feed direction is always in front of the welding arc and is therefore not applicable to omni-directional deposition. To the best of our knowledge, little attention has been focused on the deposition deviation analyses of GTAW-based AM.

Hence, in this study, the deposition accuracy of omni-directional GTAW-based AM is investigated. In Section 2, an efficient wire melting simulation model is established, which combines wire feed geometry parameters, wire feed speed, welding current and voltage parameters to analyse the wire melting offset. In Section 3, the experimental equipment and experimental procedures are introduced, and the experimental data of wire melting offset and weld bead offset are summarized. In Section 4, the effects of wire feed speed, welding current, wire feed geometry parameters on the wire melting offset and the weld bead offset are analysed in detail. The weld bead offsets in the multi-layer experiments is also analysed. Finally, in Section 5, the proposed omni-directional GTAW-based AM method is summarized.

Section snippets

Establishment of the wire melting simulation model

Most of the existing welding simulations use numerical methods that involve the arc plasma properties and its physical processes. By solving the Navier-Stokes equation or the magnetic fluid dynamics equation, the transient or steady-state temperature field, current density distribution and velocity field are obtained. The advantage of this method is that it simulates the actual welding process more accurately. However, the solution process of this method is complex, and the large amount of

Experimental system

The experiments were carried out by a GTAW-based AM system modified by a three-axis computer numerical control (CNC) system, as shown in Fig. 5. The welding machine was a Miller 350 L and works in the 120 Hz AC welding mode. The diameter of the tungsten electrode was 2.4 mm, and the shielding gas was 99.999 % high-purity argon gas with a 12 L/min flow rate. Both the wire and substrate material were 2024 aluminium alloy. The wire diameter was 1.2 mm. The substrate size was 300 × 300 × 10 mm, and

Results and discussion

Since the size of the arc column boundary is related to the welding current, it is necessary to obtain the R value in the Gaussian curve of the arc column boundary first. The set of data with a wire feed angle of 30° and a wire feeding height offset of 2 mm can better reflect the movement of the wire in the arc column. The R value was calculated using the data in which the wire feed speed was 200 cm/min: the calculated R value was 9.7 mm. The R value is used in each of the following simulation

Conclusions

To achieve omni-directional GTAW-based AM with high deposition accuracy. The relationship between the deposition parameters and the weld melting/bead offset on single-layer and multi-layer samples was analysed by simulations and experiments. The following conclusions can be obtained from this work.

(1) A simplified wire melting model is proposed for analysing the behaviour of the feeding wire in the arc column. A set of experimental data is used to calibrate the arc column boundary, and the

CRediT authorship contribution statement

Xiaolong Wang: Methodology, Software, Investigation, Data curation, Writing - original draft, Writing - review & editing. Aimin Wang: Conceptualization, Validation, Formal analysis, Visualization, Supervision. Yuebo Li: Validation, Formal analysis, Investigation, Resources, Data curation, Writing - review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work supported by the China National Defense Basic Research Program (grant number JCKY2018606B001),and China Aerospace Science & Industry Corp Foundation (grant number HTKG2017ZJT20181).

Cited by (18)

View all citing articles on Scopus
View full text