1 Introduction

Mechanical components, molds, dies, pipes, gears, and other equipment are routinely exposed to adverse environmental conditions such as fatigue, friction, corrosion, UV radiation, vibration, transition from heat to cold, etc. [1], thereby losing their function or even being irreversibly damaged as a result of abrasive phenomena (scratches, cracks, pits) and/or chemical phenomena (corrosion, plastic fracture) [2]. Repair of such damaged components requires complex manual procedures or is often not even possible. Traditional remanufacturing processes include polishing, grinding, and manual welding [3]. In addition, these are often components in the aerospace or automotive industries, where it is necessary to replace the components with new ones and manual repairs are not permitted. At this time, the advanced additive manufacturing process (AM) technique comes into play, which can work with metal powder or welding wire. Metal powder can be obtained by atomizing existing parts that are damaged or discarded, and this powder can be reused in the printing process [4], for example, using selective laser melting (SLM). Welding wire can be obtained even cheaper and easier and can also be used in AM, for example, using wire arc additive manufacturing (WAAM). Thanks to these options, AM is in line with the circular economy and more environmentally friendly.

Additive manufacturing (AM), or 3D printing, involves creating a physical object using an additive process from a 3D model, most commonly in the stereolithography (STL) format. AM has found its place in modern manufacturing and medical fields thanks to the flexibility in terms of shape creation and material usage that it offers [5,6,7,8,9], successfully re-placed conventional manufacturing methods in some cases [9, 10]. AM can be classified to fused deposition modeling (FDM), selective laser melting (SLM), stereolithography (SLA), and more depending on the working principles and materials that are printed. WAAM, on the other hand, is a technology for robotic welding using an electric arc, falling under the category of Direct Energy Deposit (DED) [11,12,13].

Given that many additive methods, including WAAM, use heat sources similar to traditional welding (e.g., electric arc), the structure can be likened to the structure of a multi-layered weld created using an arc welding method. The structure of a component created in this way largely depends on the deposition of the additive material and its subsequent solidification. In some cases, the structure may exhibit lower values of yield strength and tensile strength compared to its counterpart created using traditional methods. This effect can be mitigated, for example, by subsequent heat treatment of the deposited component or by cold working individual layers [14,15,16].

Creating components using the WAAM method depends on key parameters such as welding current, welding voltage, and the welding speed. Among other parameters, factors like wire feed rate and the use of shielding gas are also significant. In the initial phase, it is essential to determine the appropriate welding current and voltage to ensure sufficient metallurgical bonding between the individual layers of the weldments. The welding speed is also critical, particularly in areas where the torch direction changes [15, 16].

This study focuses on the reverse engineering of a braking pedal from a historical motorcycle that is no longer in production, making traditional mass manufacturing methods less relevant. The goal is to evaluate the part’s printability and mechanical performance to determine the most suitable production method for replacement components. Rather than replacing an existing large-scale production process, this research explores whether additive manufacturing (AM) offers an efficient solution for fabricating legacy components.

To achieve this, the distribution and magnitude of residual stresses in braking pedals manufactured using two different 3D printing technologies, SLM and WAAM, were determined through experimental measurements. These technologies were chosen for their availability and the unique challenges they present in the reverse engineering of the braking pedal. While SLM provides higher geometric accuracy, it results in a naturally rougher surface finish that requires additional treatment. In contrast, WAAM enables faster production, though it necessitates further machining. The chosen material was SS316L, which is the most commonly used materials for both technologies under comparison. Based on the evaluation of printability and the mechanical properties of the printed samples, we can subsequently explore alternative material choices.

Additionally, numerical simulations were conducted in the case of 3D printing. Experimental measurements of residual stress were realized using a semi-destructive hole drilling method (HDM) following the methodology outlined in ASTM E837 - 20 [17], and a destructive method was also employed [18, 19]. Numerical simulation of the distribution of residual stress in the braking pedal was carried out using Simufact Additive 2020 and Ansys Workbench 2020R2 [20]. The objective of simulating the 3D printing process is to predict the deformation of the printed part and estimate the magnitude and distribution of residual stresses [21]. By understanding deformation and residual stresses, it is possible to prevent the failure of the printed part during operation, thereby reducing prototyping costs.

2 Materials and methods

The braking pedal produced using WAAM technology herein is compared to its counterpart printed using SLM technology reported in [22]. The optimal parameters for producing the parts using the two technologies were selected based on the recommendations of the machine manufacturers. In this initial study, we focus primarily on assessing printability and their subsequent parts’ properties. Further investigation into the influence of printing parameters on residual stress are the subject of a future Design of Experiments (DoE) study.

For WAAM production, the G-code for the component was generated from the 3D model. This G-code consists of a network of points that the robot welding torch must traverse during the actual fabrication process. The G-code was used in the ABB RobotStudio program (see Fig. 1) to perform a simulation of the 3D printing process using the Metal Active Gas (MAG) method. After a successful simulation, the generated program was uploaded to the control computer of the ABB IRB1660ID welding robot, and the welding phase commenced using the Lorch S5 Robomig welding power source. The welding parameters can be seen in Table 1. For welding, Esab OK Autrod 316LSi wire with a diameter of 1.0 mm was used in a shielding gas atmosphere of M13 (98% Ar + 2% CO2). The welding of the component (see Fig. 1) was carried out on a base plate made of 316L stainless steel (1.4404), which, like the wire material, belongs to the austenitic stainless-steel category.

Fig. 1
Fig. 1
Full size image

Simulation of welding in the ABB RobotStudio program (left), positioning of the first two layers of the weld (middle), and the final appearance of the welded component (right)

Table 1 Welding parameters
Full size table

2.1 Analysis of residual stresses in the braking pedal using the HDM

The semi-destructive HDM, following the procedure outlined in ASTM E837 - 20 [17], involves attaching strain gauges (rosettes) to the surface of the measured component. Measurement locations were selected to facilitate their trouble-free installation. Two measurement points were chosen for residual stress analysis, as shown in Fig. 2. The measurement points are the same for both manufacturing methods of the pedal, SLM and WAAM.

Fig. 2
Fig. 2
Full size image

The selected locations for measuring residual stress in the braking pedal [21]

The measurement and subsequent evaluation of residual stress in the braking pedal were conducted in accordance with the methodology specified in ASTM E837 - 20 for both uniform and non-uniform residual stress up to a depth of 1 mm. Strain gauges CAE- 06 - 062UL- 120 [23] were used for the experimental measurement of residual stress using the HDM. To achieve the required depth of drilling, equipment conforming to ASTM E837 - 20 was employed. Prior to attaching the strain gauges, the surface roughness of the pedal was refined through gentle manual grinding to ensure that the values of surface residual stresses were not affected by the surface roughness of the measured component [20]. The installed strain gauges at the selected measurement locations on the braking pedal are shown in Fig. 3 for both SLM and WAAM manufacturing technologies.

Fig. 3
Fig. 3
Full size image

Adhered strain gauges on the braking pedal at the selected measurement locations for SLM pedal (left) and for WAAM pedal (right) [21]

The drilling of the hole in the pedal is performed intermittently in defined steps [17, 24]. At each step, the released deformation from each strain gauge in the strain gauge rosette is recorded. Within the analysis, the distribution of both uniform and non-uniform stresses was evaluated according to the relevant ASTM E837 - 20 standard [17]. For the experimental measurement of RS using the drilling method, strain gauges CAE- 06 - 062UL- 120 [25] were utilized along with the appropriate accessories. The drilling was carried out to the desired depth using equipment compliant with the ASTM E837 - 20 standard. Before attaching the strain gauge to the desired location, the surface roughness of the pedal was carefully adjusted by fine hand grinding using sandpaper of appropriate roughness so that the RS surface values were not affected by the surface roughness of the measured component [26]. After attaching the strain gauge, the drilling setup is established.

2.2 Analysis of residual stresses in the braking pedal using the sectioning method

The sectioning method is considered a destructive method for measuring residual stresses. The measurement of residual stresses in a component is based on the deformations in the body when it is split or sections of it are cut off. Sectioning the test sample results in the release of residual stresses and deformation of the cut sections [14]. To deter-mine the magnitude of residual stresses from the deformation of the body, the finite element method can be employed. In this case, the program Ansys Workbench was used. This method is commonly used as a comparative test for quality control in material production. Furthermore, it can be used to assess residual stresses in thin-walled tubes [18, 19, 27, 28] and other mechanical components.

2.3 Simulation of residual stresses in the braking pedal using computational methods

Values and distribution of residual stresses in the printed part can be simulated using one of several computational programs. The authors are experienced with programs such as Ansys Workbench and Simufact Additive 2020. Ansys Workbench has an implemented Additive Manufacturing System module, which allows the creation of a simulation workflow for additive manufacturing. The principle of the AM module [28, 29] involves integrating thermal analysis with structural analysis. Alternatively, Ansys Additive can be used for simulating 3D printing. This software is suitable for predicting residual stresses and deformations after printing or for optimizing the printing process [30].

Another program is Simufact Additive 2020, developed for simulating metal-based AM processes. This program allows simulation of four different types of manufacturing processes: metal powder bed fusion, metal binder jetting, Geometry inspection, and machining. In Simufact Additive 2020, it is possible to simulate 3D printing technologies such as selective laser melting, selective laser sintering, and direct metal laser sintering. At the end of the 3D printing simulation, it is possible to evaluate the deformation of the printed part. Furthermore, this program enables the prediction of the distribution and magnitude of residual stresses introduced into the printed part during the 3D printing simulation [31].

In addition, experimental measurements were realized to compare the residual stresses in the real components with the simulated ones. This is essential for adjusting boundary conditions and input parameters when optimizing 3D printing and predicting residual stresses using computational methods [11].

2.4 Microstructure of parts produced with SLM and WAAM

The microstructural analysis of the parts produced by the two technologies was conducted using a light inverted metallographic microscope, Olympus IX70 (LM). The SLM sample and the WAAM sample were sectioned where three samples on each were used for further inspection. The sections S1, S2, S3 for SLM pedal and W1, W2, W3 for WAAM pedal can be observed in Fig. 4.

Fig. 4
Fig. 4
Full size image

Cutting locations for microstructure analysis of a, b SLM pedal and c, d WAAM pedal

Metallographic samples are first prepared by hot mounting of the sections in Poly-Fast resin and subsequent grinding and polishing. To reveal the microstructure, the samples are further chemically etched using etchant V2 (an aqueous solution of hydrochloric acid (47.5%) and nitric acid (7.5%)). There are two distinct directions for cutting, including the direction that is perpendicular to the layer orientation, and across the crack resulted from sectioning method. In addition, the HV0.5 hardness tests are conducted on the cut samples using the LECO AMH microhardness tester at a low load level of 4.905 N (0.5 kg).

3 Results and discussions

3.1 The resulting values of residual stresses in the braking pedal using the HDM

The HDM can determine the distribution of residual stresses within the body, up to a certain depth. Residual stresses can subsequently be evaluated as either uniform or non-uniform, depending on their nature within the body.

3.1.1 Calculation of uniform stress

From the measured deformation values, it is possible to determine the magnitude and direction of the principal stresses, as well as the equivalent stress, such as von Mises stress, for uniformly distributed residual stresses at various depths from the drilled hole for both measurement locations denoted here as numbers 1 (74 mm) and 2 (57.5 mm). Furthermore, it is possible to calculate axial stresses according to the orientation of the strain gauge. Table 2 provides the values of normal stresses σa(1) and σc(3) respectively at the locations of strain gauge number 1 and strain gauge number 3 (as per the orientation shown in Fig. 3 and norms ASTM E837 - 20), along with the magnitude of the equivalent stress σVMS for both measurement locations for the SLM pedal. Table 3 presents the same measurement results for the WAAM pedal. The meanings of individual quantities are the same as in Table 1.

Table 2 The calculated values of uniform stress at both measurement locations for the SLM pedal
Full size table
Table 3 The calculated values of uniform stress at both measurement locations for the WAAM pedal
Full size table

3.1.2 Calculation of non-uniform stress

In real bodies, non-uniform stress is commonly present. To determine non-uniform stress, a blind hole is drilled, and its depth is gradually increased in steps of 0.05 mm. Several methods can be used for calculation [17, 32, 33]. To determine non-uniform stress values from the released deformations, the procedure specified in ASTM E837 - 20 was utilized here. The resulting values for non-uniform residual stress at both measurement locations for the SLM pedal are provided in Table 4, while Table 5 presents the values of non-uniformly distributed residual stresses in the WAAM pedal.

Table 4 The calculated values of non-uniform stress at both measurement locations for the SLM pedal
Full size table
Table 5 The calculated values of non-uniform stress at both measurement locations for the WAAM pedal
Full size table

3.2 Residual stress analysis in braking pedal by the sectioning method

The second experimental method used for measuring residual stresses was the sectioning method [18]. After cutting the braking pedal, there was deformation in the cut halves due to the release of residual stresses. For illustration, Fig. 5 shows the deformation of the braking pedal caused by residual stresses after the cut was made at measuring point no. 2 (57.5 mm) on the pedal manufactured using SLM technology.

Fig. 5
Fig. 5
Full size image

Deformation of the braking pedal after cutting it at measuring point no. 2 on a SLM pedal and b WAAM pedal

It is obvious that the SLM pedal distorted more in comparison with the WAAM pedal. The calculation of RS σa was performed using the FEM in the Ansys Workbench computational program. The calculation was performed as a static structural analysis. The material parameters of the material model used for this room temperature analysis of SLM technology were taken from the literature [21]. For the static structural analysis, the material model was from the Ansys Workbench database. The calculation of RS was carried out as a reverse problem. Boundary conditions were applied to each half of the pedal arm to achieve the same opening displacement as measured after cutting. The gap between the halves in Ansys had to match the measured gap on the actual braking pedal. Subsequently, the stress σa was evaluated at both measuring points. The values of stress σa at both measuring points as a function of depth from the outer surface are presented in Table 6 for the SLM pedal and in Table 7 for the WAAM pedal.

Table 6 The values of stress σa(1) at measuring point 1 and measuring point 2 determined by the sectioning method in the Ansys Work-bench environment for the SLM pedal
Full size table
Table 7 The values of stress σa(1) at measuring point 1 and measuring point 2 determined by the sectioning method in the Ansys Work-bench environment for the WAAM pedal
Full size table

In the following graphs, the values of residual stresses obtained through experimental measurements (HDM and sectioning method) on pedals manufactured using two production technologies, SLM and WAAM, are compared at two measuring locations. The residual stresses are reported in terms of equivalent residual stress σVMS, residual normal stresses σa(1), and σc(3).

Figure 6 provides a mutual comparison of the equivalent residual stress according to the Von Mises theory (σVMS) for both measuring locations with the HDM. The stresses are evaluated as uniform and non-uniform throughout the depth according to the methodology described in ASTM E837 - 20. For both evaluation methodologies, the results obtained from the SLM pedal and the WAAM pedal, which was subsequently machined by five-axis CNC machining, are compared.

Fig. 6
Fig. 6
Full size image

Comparative analysis of the achieved results of Von Mises equivalent stress (σVMS): a at measurement location number 1 and b at measurement location number 2, for both pedals

When looking at both profiles, we can observe that the residual stresses in the SLM pedal are significant and several times higher than in the WAAM pedal. In the WAAM pedal, the residual stresses are relatively low, and they likely underwent a majority of stress relief during the pedal’s production. Both residual stresses exhibit a decreasing character from the depth of approximately 0.2 mm for both manufacturing methods. Up to the depth of 0.2 mm, there is an increase in residual stresses introduced in the surface layer of the WAAM pedal owing to the milling process.

Figure 7 reports the normal residual stresses σa(1) and σc(3) obtained from HDM at location no. 1 (74 mm) and location no. 2 (57.5 mm) for both uniform and non-uniform cases. The results show a similar trend in terms of stress reduction as the depth increases. The normal residual stress in the SLM pedal is significantly higher than in the WAAM pedal. Additionally, the results of the sectioning method indicate that the residual stress in WAAM pedal follows the results of the HDM. Normal residual stresses are presented for both calculation methodologies (uniform and non-uniform) and both pedal manufacturing technologies. The residual stress in the SLM pedal is significantly, almost an order of magnitude higher. Additionally, the results of the sectioning method show that the residual stress in both technologies corresponds with the results of the HDM.

Fig. 7
Fig. 7
Full size image

Comparison of achieved normal stress results (σa and σc) respectively at a, c measurement location number 1 and b, d measurement location number 2 for both pedal manufacturing technologies

3.3 Microstructure analysis in SLM pedal and WAAM pedal

The second experiment was to investigate the microstructures of the SLM pedal and the WAAM pedal. Sections of the SLM pedal in Fig. 8 were evaluated. Metallographic sections S1, S2, and S3 correspond to the marked sections in Fig. 4a, b. In addition to this, the hardness is compared for a better understanding of the effect of grain types on the part’s mechanical properties as shown in Figs. 9 and 10.

Fig. 8
Fig. 8
Full size image

Samples S1, S2, and S3 for microstructural study from SLM pedal. Yellow arrows indicate the normal of the layers. Red arrows indicate the directions of the hardness tests

Fig. 9
Fig. 9
Full size image

Hardness HV0.5 at the cuts of SLM pedal, sample S1, S2, and S3

Fig. 10
Fig. 10
Full size image

Microstructure at the cuts of SLM pedal, a, b sample S1, c, d sample S2, and e, f sample S3. Red arrows indicate the voids caused either by trapped gas (small, spherical pores) or lack of fusion (big, irregular shape)

Sample S1 (in the thicker part of the pedal, along the crack) has an average hardness of (232 ± 8) HV0.5, sample S2 (in the thinner part of the pedal, in normal of the sample) has an average hardness of (256 ± 8) HV0.5 and sample S3 (taken across the board, from the surface to surface) has the hardness of (213 ± 10) HV0.5. The average hardness is (234 ± 19) HV0.5 and is comparable in the three lines. The highest hardness is found in sample S2, where there are fewer defects. If voids are present in the microstructure, especially often for printed components, these are inhomogeneities in the microstructure that can affect the hardness. Even though we avoid places with voids when measuring hardness (at the point of intent with a diamond pyramid), they can be hidden just below the surface.

The microstructure is formed by fine melt pools formed during SLM and varies depending on the orientation of the samples [34]. In the transverse direction, individual weld beads with a width of about 100–200 μm can be observed, which resemble fish scales. Each scale represents the cast state of the material after solidification, which is formed by fine crystals of different orientations (Fig. 10a–d). The microstructure in the longitudinal direction (Fig. 10e, f) is different and appears as individual tracks. These differences in the microstructure without additional heat treatment will affect the anisotropy of the mechanical properties [35, 36]. In addition, the voids caused by lack of fusion (about 50 µm in size) can be observed. Those are distributed primarily on the grain boundaries, along which the crack can propagate.

The microstructure of these samples consists of bright areas of the main phase austenite, crystallizing in a face-centered cubic lattice (FCC), due to the higher nickel content (10.0–13.0 wt%). In the melt pools and along their boundaries, delta ferrite, crystallizing in a body-centered cubic lattice (BCC), occurs. Furthermore, due to the presence of a higher proportion of chromium (16.5–18.5 wt%) and molybdenum (2.0–2.5 wt%) in the microstructure of 316L steel, carbides (e.g., Cr₇C₃, Cr₂₃C₆) or molybdenum (e.g., Mo₂C or M₆C) may occur.

Similarly, sections of the WAAM pedal in Fig. 11 were evaluated in terms of hardness in Fig. 12 and microstructure in Fig. 13.

Fig. 11
Fig. 11
Full size image

Samples W1, W2, and W3 for microstructural study from WAAM pedal. Yellow arrows indicate the normal of the layers. Red arrows indicate the directions of the hardness tests

Fig. 12
Fig. 12
Full size image

Hardness HV0.5 at the cuts of WAAM pedal, sample W1, W2 and W3

Fig. 13
Fig. 13
Full size image

Microstructure at the cuts of WAAM pedal, a, b sample W1; c, d sample W2; and e, f sample W3

The macrostructure consisted of several large weld beads, whereas there were discontinuities in some places. Sample W1 (in the thicker part of the pedal, along the crack) has an average hardness of (195 ± 9) HV0.5, sample W2 (in the thinner part of the pedal, in normal of the sample) has an average hardness of (218 ± 11) HV0.5, and sample W3 (taken across the board, from the surface to surface) has the hardness of (231 ± 8) HV0.5. The average hardness is (215 ± 17) HV0.5 which is comparable in the three lines and is slightly lower than the hardness of SLM samples.

The microstructure of the weld bead is austenitic, with the presence of dendritic delta-ferrite [37]. Columnar crystals are present in places. Individual weld beads are several micrometers in size. The larger volume of the weld bead cools more slowly than in the SLM process, which leads to the growth of dendrites and the exclusion of larger delta-ferrite particles [38]. Remarkably, the microstructure does not depend on the orientation of the samples. Therefore, it implies that the manufacturing process does not have any effect on the mechanical performance of the component. However, the slower cooling rate led to a decrease in the hardness of the part produced by the WAAM process.

4 Conclusions

From the residual stress results on SLM and WAAM pedals, it can be concluded that residual stresses in WAAM-processed pedals are significantly lower than those in pedals using SLM technology, almost an order of magnitude. Residual stresses in SLM-processed pedals are tensile, and the measured values border on the material’s limit parameters for this material made with this technology from virgin powder. On the contrary, in WAAM-processed pedals, residual stresses in greater depths are mostly compressive, which generally has a positive effect on the static and fatigue strength of the component. This agreement applies to both HDM and sectioning methods. Residual stress on the surface to a depth of 0.2 mm in WAAM pedals and subsequently machined by milling is influenced by the machining process. In addition, the internal microstructure of the SLM pedal contains pores caused by either trapped gas or fusion defects, which can potentially serve as crack initiators under static and especially fatigue loading. WAAM manufacturing technology seems more favorable for reproducing components in terms of residual stress levels, overall deformation, and subsequent mechanical loading due to lower tensile residual stresses. However, this manufacturing technology requires subsequent milling to achieve the final shape of the pedal, significantly increasing its cost-effectiveness for practical applications. This reason leads to only functional surfaces being machined, and the rest of the surface remains as it was printed.