Wire Arc Additive Manufacturing Process for Topologically Optimized Aeronautical Fixtures

Abstract

Additive manufacturing (AM) technologies in metallic materials have experienced significant growth over recent decades. Concepts such as design for additive manufacturing have gained great relevance, due to their flexibility and capacity to generate complex geometries with AM technologies. These new design paradigms make it possible to save on material costs oriented toward more sustainable and green manufacturing. On the one hand, the high deposition rates of wire arc additive manufacturing (WAAM) stand out among the AM technologies, but on the other hand, WAAM is not as flexible when it comes to generating complex geometries. A methodology is presented in this study for the topological optimization of an aeronautical part and its adaptation, by means of computer aided manufacturing, for WAAM manufacturing of aeronautical tooling with the objective of producing a lighter part in a more sustainable manner.

Introduction

The reality is that production demands are increasingly stringent and there are a greater variety of products, shorter series, shorter manufacturing times, and a greater demand for quality. Moreover, the design of flexible fixtures has always stimulated great interest among industrial design and manufacturing engineers, because of their quality impact on manufactured parts. It inevitably means that the industry will need greater flexibility and automation in its manufacturing process, to achieve a competitive product with high returns, increasingly in need of a more exhaustive study of the specific standards.^1,2 In this regard, it should be noted that both the quality and the profitability of manufactured parts depend to a large extent on the tooling, the cost of which can be between 10% and 20% of the cost of the manufacturing system, whereas the use of flexible fixtures can save up to 80% of tooling costs.³

Economically, the tools have a great impact on manufacturing costs, and there are two types of situations: (1) turnkey developments of the machine system/system + tooling; and (2) developments to be used in existing machines and systems. Most of the tooling designs are in this last group, affecting 74% of tools.⁴ For this reason, it is important that the tooling should be adapted to the existing processes by means of flexible and modular solutions that reduce the economic costs of their development.

The tools are used to locate, secure, and support the parts during the manufacturing processes, and their behavior affects the result of the process under consideration (machining, welding, assembly, …) in terms of quality, cost, and performance. Likewise, the tools have a great impact both on the development of the processes and on the machine tool capabilities. They are usually designed as a subsystem, independent of such aspects as cycle times and the cutting process itself,³ although they have an important influence on these aspects, and even the tooling and the process can be considered mutually dependent.⁴

Additive manufacturing (AM) with its great freedom of design and adaptability is an interesting alternative to manufacture these types of parts. The first steps of AM, initiated 150 years ago, were taken to manufacture parts with two-dimensional (2D) layers by using topographic maps and free-form photosensitive sculptures.^5,6 The extension of its application to mechanical materials occurred in the early 2000s, although pioneering work already existed in the 1980s.⁷ The expiry of some of the patents associated with AM processes and its development by third parties has contributed to research and has helped extend its use. The adoption of the design and manufacturing paradigm by means of additive techniques is a fact and its growth appears inexorable, having consolidated itself as an alternative to traditional means of production. When paradigms for AM are mentioned, one of the main terms is the “Design for additive manufacturing” (DfAM),^8–12 which different authors have defined.^13,14 The DfAM aims at improving productivity and efficient use of resources in the design and optimization of parts/assemblies, considering the production system through additive material technologies. Therefore, the boundary conditions offered by each additive technology^15,16 must be considered, to adjust to the specific requirements of the market.¹⁷

At a first level, DfAM usually addresses the dimensional and geometric characteristics of the part (part size, and symmetrical planes that affect its integration into the assembly).¹⁸ Other studies also integrate qualities throughout the lifecycle of the piece, such as useful life, functionality, customer satisfaction, safety, reliability, etc.¹⁹ Finally, the integration of the process variable is sought, so that the designer adjusts the models to the process in use, to make it more effective and productive, including the feasibility of its use with the material to be processed⁹; therefore, computer aided design (CAD) tools have been improved, by adapting them to DfAM.²⁰

The adaption of computer aided manufacturing (CAM) toolboxes of current software to AM processes that are under development is another important area of study, adding knowledge of the process leading to the most advantageous path toward the scaling-up of the parts for manufacture.^21,22 In this sense, it is important to establish the limitations of the different technologies, such as growth per layer or overlap width in direct energy deposition (DED) techniques and best path selection.²³ Lastly, and taking advantage of the possibility of manufacturing unconventional geometries, the adaptation of topological part improvement techniques will enhance effective lean production from limited material resources.²⁴ These design concepts can, therefore, be adjusted to fixture-device manufacturing needs.

Finally, from among the different types of AM technologies, wire arc additive manufacturing (WAAM) technology is a new manufacturing process that builds parts layer by layer using an electric arc. It should be noted that other AM processes have received more attention in recent years. For this reason, although great advances have been made that demonstrate the immense potential of WAAM technology, there are still certain aspects, beyond the machine itself, that must be developed to offer a complete solution to this process: CAD/CAM software and final quality verification of the piece. In this study, some of the issues are addressed and the integration of this technology into industrial machinery is presented, offering automated component manufacturing. Among the different WAAM technologies, the gas metal arc welding (GMAW) process was chosen for the manufacture of the topologically optimized tooling in aluminum. The main problem of GMAW aluminum deposition is the appearance of pores, which is reduced through the use of alternating negative and positive intensity cycles or the alternating current (AC) mode. For example, Cong et al.²⁵ performed a comparison of different welding methods, focused on pore size and pore-size numbers.

In this study, a new methodology that is also adaptable to other DED technologies, for manufacturing topologically optimized aeronautical tools using WAAM technology is presented. Although WAAM is not the solution that obtains more precise geometry within AM technologies, this article opts for this alternative due to its advantages. Among the comparative advantages is the ability to manufacture high deposition ratios, the use of raw material, and the ease of industrial implementation of the technology. This methodology consists of establishing the operative restrictions of the part, optimizing its design limited only by the process constraints, adapting it to polynurb functional surfaces and to the process, setting the geometric and power generation parameters, and finally establishing the most suitable manufacturing strategy with CAM tools. The objective of the redesign of these parts is to arrive at a part model that only consists of the material that provides the required rigidity for the plant. The consequence will be tooling that, on the one hand, removes less material than the original (less material equals cost-effective AM) and, on the other hand, it will mean a considerably lighter version compared with the original model. Thus, the main objective is an AM technology for light, easily customizable tools that meet functional requirements.

Materials and Methods

Case study: fixing turret design

The manufacture and assembly of industrial products require tools of intermediate complexity, made up of structures of different types. It would be impossible to cover all the specific variants as far as the tools are concerned. In the present study, a common tooling method used for processing and assembly of aeronautical components was considered. These are tools mounted on a long bench consisting of clamping systems, with degrees of freedom, so that parts that will be assembled may be positioned with due precision. Figure 1 shows an example of this type of tooling:

FIG. 1.

CAD design of the assembly fixturing system. CAD, computer aided design.

The different types of elements indicated in the previous image are usually manufactured by welding, folding, and/or machining processes. The objective of this section is to evaluate the suitability of each type of element in the tooling for AM manufacture.

The structural supports are extruded bars and welded tubes, forming light yet rigid structures. They do not usually require close manufacturing tolerances and, consequently, they are usually relatively low-cost components of little added value.

The linear trolleys are mainly commercial rails to which plates or supports are added to secure the pieces. These plates may require close manufacturing tolerances, to ensure precise part positioning, but generally the 2D parts are simple to manufacture.

The turrets, shown in Figure 2a, are more complex elements that adapt to the geometry of the piece, hold it secure, and at the same time provide access to certain areas and prevent collisions with the tools used in the assembly process. Figure 2b shows the dimensions of the part to be additive manufactured. These components are of more complex geometries with very tight manufacturing tolerances. They often consist of several components screwed together, either because the object cannot be formed in one piece or because forming an integral component will require an excessive number of machining hours, the cost of which is hardly justifiable. The objective of this article is therefore the improvement, through topological optimization methodology, of the turret fixture design.

FIG. 2.

(a) CAD design of the fixing turret and (b) principal dimensions of the part.

In view of what has been stated earlier, it is evident that (AM) offers advantages over conventional manufacturing processes for these types of elements where:

• 1.Complex structures are manufactured in one part.
• 2.Assembly operations are eliminated.
• 3.Optimization of the material volumes is needed to build the part, resulting in a lighter product.

The fixturing turret in the initial assembly is made of aluminum alloy Al 5083, the mechanical properties of which are listed in Table 1 and are used for the finite element analysis (FEA).

Table 1.

Mechanical Materials Properties for Al 5083

Material	Young's modulus (GPa)	Ultimate tensile stress (MPa)	Yield stress (MPa)	Density (g/cm3)
Al 5083	71	300	145	2.66

DfAM methodology

This methodology (Fig. 3) begins by defining the topologically optimized part, which meets the functional requirements of the part in the assembly. Both the geometric and the mechanical properties of the WAAM additive material are then characterized. Information is essential in this phase for the redefinition of the part, given its processing constraints. Subsequently, the CAM that marks the trajectories of the torch for the manufacture of the part will be defined. The key features of the part with the greatest restrictions on surface and dimensional quality will have been machined, to ensure a functional and easily assembled piece. Finally, the results will be compared with topologically optimized ones and with the starting design.

FIG. 3.

Steps for WAAM process-oriented topology optimization. AM, additive manufacturing; CAM, computer aided manufacturing; CATIA, 3D CAD software; FEA, finite element analysis; PolyNURB, topological design and analysis tool; WAAM, wire arc additive manufacturing.

WAAM: material, machine, and parameters

The topologically optimized WAAM turret and test wall, manufactured to characterize the manufacturing process and its properties, consisted of Al 5356, because it is the most similar weldable aluminum to Al 5083 in terms of its composition. In both cases, the substrate was first secured to the table and prepared by cleaning and polishing before welding. Then, the pass library containing the manufacturing parameters corresponding to the parts were loaded using the customized software. After the start of the welding, the beads were built up layer by layer. The test sample and the turret were manufactured out of Al 5356, using WAAM technology based on GMAW. The aluminum was introduced in the process as a commercially produced Al 5356 wire (EN ISO 18273:S AL5356) with a diameter of 1.2 mm. The substrate consisted of 15 mm wide aluminum 5356 plates. The inclination of the torch with respect to the substrate was 90° angles with a 15 mm stick-out.

In these experiments, the Addilan WAAM machine (3 axis +2 tilt table) equipped with a Titan XQ 400 AC pulse (EWM AG) generator was operated by utilizing AC-GMAW welding mode at wire-feed speeds of between 1.2 and 12 m/min. Figure 4 shows the machine configuration utilized for the wall specimen and part production. In addition, 100% argon gas was used as the shielding gas at a flow rate of 30 L/min, through a nozzle diameter of 20 mm.

FIG. 4.

Addilan WAAM dedicated machine and torch head.

A test wall of 5356 aluminum was prepared, to characterize the manufacturing process, material, and geometry of the bead. In this case, the start and the end of the weld beads were marked in the customized software, thereby determining the length of the wall (240 mm). This wall was manufactured by overlapping three beads and superimposing 22 layers on top of the flat substrate. The bead overlap was 65% of the width (3.25 mm) with alternating layers and the controlled deposition welding during production ran in opposite directions, to compensate for the bead-weld direction and to maintain the geometric tolerances of each part. A cooling time, in this case a period of 50 s, was introduced between the consecutive bead welds, so that the thermal input was never excessive. In addition, the most appropriate set of parameters to manufacture the test wall and the topologically optimized WAAM turret was chosen, in so far as the contribution rates were as high as possible with good penetration and an adequate wetting angle, but without excessive thermal input, determining a wire-feed speed of 8 m/min and a travel speed of 168 cm/min for all the layers (Table 2). The inclination of the torch with respect to the substrate was 90° angles with a 15 mm stick-out.

Table 2.

Gas Metal Arc Welding Manufacturing Parameters of the Test Wall and Topologically Optimized Wire Arc Additive Manufacturing Turret

Number of layers	Wire-feed rate (m/min)	Deposition rate (kg/h)	Overlap (%) distance between layers (mm)	Cooling time (s)
22	8	1.44	65 (3.25 mm)	50

The specimens were extracted from this test wall to perform the tensile tests, the mechanical characterization, and the microstructural analysis. The distribution of specimens on the test wall is shown next, in Figure 5.

FIG. 5.

Sample distribution for mechanical characterization.

Results

Topological optimization of the fixing turret

An FEA of the preliminary turret design was performed, to calculate stress deformation at the working stage for later comparison with any stress deformation of the part designed for AM. Figure 6 shows the forces acting on the turret, the compression force due to the part support, and traction forces for positioning purposes. The non-design surfaces are dedicated to the part clamping, so no modification of their geometries is permitted. The turret is fixed on the base of the assembly, and the other surfaces are considered as free bounded.

FIG. 6.

Load assumption and constraints. Forces supported by the turret.

The FEA calculation is based on a simplified turret model. Small holes, radii of agreement, and holes are eliminated and the three pieces of which the set is composed are merged into a single solid. The loads shown in Figure 4 were applied to this model, and the base of the assembly was embedded. Four control points were defined on which to measure displacement. The displacement is defined as the total permissible variation that the deformed part can have from its initial spatial position (all directions) at control points 1 and 2 of the original part model. The maximum admissible displacement in each of these points is ±0.1 mm, whereas a safety factor of at least 2 must always be guaranteed for the maximum stresses that appear in the part. The displacements observed from the FEA calculation for the original turret are shown in Figure 7 and Table 3. It is a very rigid piece that hardly deforms under the applied loads. Maximum Von Misses Stress is located on the holes at the top of the part. Very low stress values (<2 MPa) appeared in the material, far removed from the elastic limit of aluminum (145 MPa).

FIG. 7.

Displacement and von Misses Stress distribution on the original fixing turret.

Table 3.

Displacements Observed at the Control Points of the Original Design

Control point	1	2
Displacement (mm)	4.268 · 10−4	4.793 · 10−4

Having analyzed the original part, in this section the turret redesign is described, upon the premise that it may be manufactured with an AM process, on the one hand, seeking compliance with specifications (displacement at the control points <0.1 mm, voltages with a safety factor >2) and, on the other, seeking to reduce the weight of the turret and, consequently, the manufacturing cost. The next step to proceed with the optimization of the turret geometry is to determine which parts of the turret are functional and must therefore remain unaltered, so that the turret can fulfill its function, and which parts may be subject to modification during the redesign process. Once the work schedule had been defined, it was possible to launch an optimization calculation for the turret design. To do so, the same forces that were applied to the original model were applied to the AM model. The same boundary conditions as in the original turret were also considered, that is, that the base was embedded (there is no displacement of the underside). The objective established for the calculation of mass reduction has been to reduce the design volume to 30% of the original volume, maximizing the rigidity of the part. The result obtained from the optimization is shown in Figure 8.

FIG. 8.

Mass reduction calculation result.

The optimization calculation established which areas of the design volume that were defined were the areas that “worked” from the structural point of view and which were not, and could therefore be eliminated. In a conventional manufacturing process such as machining, this approach of eliminating the material that hardly provides structural resistance requires additional machining hours (if the traditional approach is considered), which makes the part more expensive. So, unless it really has maximum admissible weight limitations, the process is bypassed as without interest. However, this approach makes every sense when the manufacturing process is of an additive type, because it will produce lightweight parts that are structurally strong enough for their purpose.

Once the volumetric areas of the material were designed, but were yet to be machined, the volume was worked on by using geometries called polynurbs to trace what is called an organic design of the part. In those structural areas in which material must remain, polynurbs are drawn that adapt to the calculated geometry. Once all the structural volume had been covered with polynurbs, the following turret design was achieved (Fig. 9).

FIG. 9.

Use of Polynurbs on the left-hand-side volume and resulting CAD design of the fixing turret.

The final step was to verify that the lightened structure met the turret specifications. To do so, an FEA of the redesigned component was performed (Fig. 10), to determine the magnitude of the displacements that appeared under loading. From the results of the displacement calculation, the redesigned model was slightly more flexible than the original, but it still amply complied with the specifications that establish a maximum but still amply complied with the specifications that establish a maximum deformation of 0.1mm at the control points (Table 4). Regarding stresses, the maximum calculated values (<5 MPa) were well below the material limit (145 MPa). The weight of the redesigned piece was 2.3 kg.

FIG. 10.

Displacement and von Misses Stress distribution on the topically optimized fixing turret.

Table 4.

Displacements Observed at the Control Points of Both the Optimized and the Original Design

Control point	1	2
Displacement [original design] (mm)	4.49 · 10−4	4.793 · 10−4
Displacement [optimized part] (mm)	1.46 · 10−3	1.64 · 10−3

The new turret design will require some finishing operations, before it can be mounted on the tooling and comply with the functionality of the original part, which the AM process cannot perform. One such operation will be thread machining and hole diameter adjustment operations with tight tolerances. In the case of thread machining, it might be the direct machining of threads or the placement of metal inserts. As for holes with close tolerances, the part should be inserted into a milling machine and the holes that require it should be reviewed. In the case at hand, it would be necessary to perform the following operations:

• 1.Planning the base and inserting threaded bushings for clamping guides.
• 2.Threading of holes.
• 3.Adjust H7 tolerance on 20 mm diameter holes.

Mechanical properties and microstructure characterization of added material

Figure 11 shows the wall macrostructure in the Al 5356 alloy. No manufacturing defects, lack of filling, or cracks or pores at the macroscopic level may be observed. Deposited material is shown layer-by-layer, and material diffusion may be seen. The microstructure on the two different directional planes (XZ and YZ) of grain configuration revealed microscopic pores of dimensions <100 μm. This pore formation when manufacturing aluminum parts using arc-welding technologies was addressed by Horgar et al.,²⁶ Cong et al.,²⁷ and Su et al.²⁸

FIG. 11.

Macrostructure and microstructure at the different direction of the smaller wall.

Further, six samples were extracted from the additive manufactured test wall in the vertical direction and another six in the horizontal direction, as can be seen in Figure 5. These samples were tensile tested, to characterize the material produced by GMAW. These values were agreed upon as input for the FEA calculation of the topologically optimized part oriented toward WAAM manufacturing. The values obtained (Table 5) are similar to those described in Table 1. The presence of micropores slightly reduced the ultimate tensile strength (UTS) and the plasticity of the WAAM manufactured samples, so much so that the UTS dropped by 30% with respect to the material used in the original part (300 MPa ref.). The presence of micropores can alter the fracture mode to a combined ductile and brittle fracture. This change in fracture type is similar to the one reported by Li et al.²⁹

Table 5.

Mechanical Characterization Results for Al 5356 Added Material

WAAM	Ultimate tensile stress (MPa)		Yield stress (MPa)		Elongation (%)
	H	V	H	V	H	V
Al 5356	277 ± 3	264 ± 1	111 ± 6	106 ± 3	33 ± 1	27 ± 3

WAAM, wire arc additive manufacturing.

Finally, to monitor the geometric shape of each layer, commented on earlier, they were scanned by using a geometric laser (Laser Scanner Q4 Series) placed alongside the welding torch on the tip of the torch holder. This laser scanning generates data for easy calculation of wall growth. The mean layer profile in each torch pass is shown in Figure 12.

FIG. 12.

WAAM manufacturing shape of the layer-by-layer bead geometry.

CAM definition and WAAM of the final part

CAM dedicated software Autodesk Powermill Additive was used to define the trajectories that produce the final part. The bead width (5 mm) and growth per layer (1.5 mm) values, obtained from the test wall analysis, were entered for the generation of the trajectories. Figure 13 shows the trajectories of the layers that are arranged by defining the contour paths first and then, the filling of the inner part following the oscillatory strategy. The oscillation overlap distance was determined at 3 mm, and the distance to the contour path was 2 mm. In consecutive layers, the direction of the oscillation is changed, and the oscillations become perpendicular and reverse, as can be seen in the Figure 13 layer [i] and layer [i + 1].

FIG. 13.

WAAM manufacturing torch path trajectory of the layer-by-layer sequence.

The comparison between the original and the redesigned part was only possible, once the part had been manufactured by using WAAM technology. Table 6 shows the comparison between the weights of the initial model and the lightened models, whose weight reduction was at 61%. In the WAAM-adapted model, a weight reduction of 31% was achieved, with respect to the initial piece. Both the lighter-weight parts showed similar results, under the maximum stress to which the part can be subjected with the defined load hypotheses, well over a safety factor of 23.5%. It should be noted that the safety factors for the final part were calculated from the mechanical characterization of the aluminum alloy obtained in the WAAM process. The initial piece already had a very high safety factor. The estimated displacements were also within permissible limits, satisfying the initial requirements for the use of the part in this application.

Table 6.

Comparison of Results from Original Model, Lightweight Model, the Additive Manufacturing, and the Wire Arc Additive Manufacturing Designs of the Part

	Original design	Design for AM	WAAAM design
Weight	6 kg	2.3 kg (↓61%)	4.1 kg (↓31%)
Maximum stress	1.99 MPa	4.83 MPa	4.65 MPa
Elastic limit	145 MPa	145 MPa	109 MPa
Security factor	73	30	23.5
Displacement point 1	1.49 · 10−4 mm	1.03 · 10−3 mm	2.58 · 10−3 mm
Displacement point 2	6.85 · 10−5 mm	5.11 · 10−4 mm	1.32 · 10−3 mm

The total times used in the AM of the WAAM part are shown next, in Table 7. The significant lengths of time used for both the interlayer cooling and positioning movement are added under idle time in the table. The matrix strategy can be applied to use the idle time to make another piece in parallel.

Table 7.

Wire Arc Additive Manufacturing Times for the Optimized Fixing Turret

	Additive time	Idle time
Fixing turret	2 h 40 min	1 h 46 min

Finally, the machining of the critical features for positioning the part was performed with a fixing turret. Three critical machining operations may be distinguished: (i) the top seats and a slot of the fixing turret were milled; (ii) three larger lateral holes in the lateral part; and, lastly (iii) the threaded holes for clamping the bracket and the guide to the upper part. Figure 14, given next, shows the final topologically optimized mechanized part ready to be positioned in the assembly.

FIG. 14.

Final WAAM manufactured and machined fixing turret.

Residual stresses and deformations in parts manufactured by WAAM additive are often a critical source of error, especially in thin-walled parts.³⁰ In this case, when we are faced with a geometry before the final machining, the incidence of distortions is mitigated thanks to the machining of the critical geometries to ensure the correct operation of the fixing turret in the final application, as can be seen in Figure 14.

Conclusions

The methodology that has been presented is summarized by the evolution of the part geometry shown in Figure 15. Starting with the initial part and its use requirements, a topological optimization was performed. The characteristics of the WAAM process were not considered in this first approach; therefore, a process was characterized in a test wall. The target geometry was reconstructed with the results of the process and the material characterization and entered into the CAM to complete the WAAM manufactured part on the computer numerical control machine. This part was, finally, finished by machining the reference surfaces.

FIG. 15.

Summary of the methodology for process-oriented topological optimization WAAM.

The conclusions and observations from this work can be summarized as follows:

• The topological optimization of an aeronautical fastening component for the WAAM process has produced a lighter-weight part that required 31% less material than the original application.

• A first optimized piece has been proposed by means of polynurb surfaces for possible AM that could lighten the component by 61%.

• The process parameter settings have been presented for AM GMAW of an aluminum alloy series 5XXX. These process parameters can be modified by entering the parameters for the path deviations to program the machine. The geometric characterization of the bead served as input to the dedicated CAM software, and the mechanical characterizations were used in the FEA simulation that yielded the safety factor and the maximum stresses.

• Materials with better mechanical properties, which might otherwise be disregarded due to their weight, may be included with this methodology, if the application requires so.

• The manufacturing time of the piece is 4 h and 26 min, of which a large part is idle time. This factor opens up possibilities for production-oriented optimization through component matrix manufacturing of more than one part at the same time.

Author Disclosure Statement

The authors declare no conflict of interest.

Funding Information

The authors thank the Basque Government for funding the QUALYFAM project, through the ELKARTEK 2020 (KK-2020/00042) and the ADIFIX, HAZITEK 2019 (ZL-2019/00738) programs.