Open Access
Add to BibSonomy
Abstract
In this paper, a novel surface finishing method of cone annular beam laser re-melting (CALR) is proposed which aims at optimizing the side surface quality of AlSi10Mg parts produced by directed energy deposition. Meanwhile, a feature-based characterization method was conducted to realize the multiscale analysis of the topographic features. The results show that the developed CALR is a powerful tool which can greatly reduce the surface roughness: the final optimal aluminum alloy thin-walled parts with a roughness of 7.1 µm was obtained, which was 68.3% less than the original roughness of 22.4 µm without the implementation of CALR. The optimization mechanism reveals the roughness reduction was mainly attributed to the elimination of the stair-step effect, while the role of powder particle features was much less significant on roughness reduction of the overall surface. The CALR method together with the feature-based surface characterization provides an innovative solution for side surface quality optimization for laser additive manufacturing.
© 2021 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Additive manufacturing (AM) is regarded as one of the most promising breakthrough technologies in the manufacturing industry in recent years and is treated as an essential manufacturing development direction by countries all over the world [1]. AM is based on the theory of materials metallurgical bonding (powder, wire or liquid material, etc.) through layer-wise deposition, so as to realize the generation of solid parts from a CAD model [2]. There are several similar techniques in the field of metals AM, which are usually given different names, such as powder bed fusion (PBF) and directed energy deposition (DED) [3,4]. In PBF, powder materials are used, and there are two common used heat sources: laser or electron beam, including selective laser melting (SLM) [5] and electron beam melting (EBM) [6]. In comparison, the DED (developed by the United Technologies Research Center (UTRC) [7]) utilizes powder feed fusion technique, which is of particular interest in this paper. Generally, AM technology has rapid prototyping capabilities for complex structural parts that can greatly reduce the product design and development cycle [8]. At present, the market of AM is booming and many industrial manufacturing fields like aerospace and medical equipment have begun to gradually popularize AM technology. Other new industries, such as autonomous driving and 5G communication, are keen to implement these advanced manufacturing techniques [9]. Nevertheless, AM technology still faces numerous challenges e.g. high porosity and poor surface quality [10].
The surface quality of most PBF and DED components is poor: the roughness of surface along the build orientation often remains with high value (Ra/Sa = 20-30 µm) [11], which is significantly higher than typically machined parts (Ra=0.5-2 µm) like rough turning, milling operation. Hence, the produced parts for industrial applications usually require further steps of surface finishing treatment which will increase production time and cost [12]. For this reason, an important challenge in AM parts is how to optimize the surface roughness of the formed parts. Besides, the previous research results have shown that low surface roughness can reduce the existence of surface voids and dents, thereby improving the fatigue life of parts significantly [13,14] and slowing down corrosion behavior [15–17]. In fact, many factors can affect the surface roughness of AM parts. Among them, the coupling of gas-particle flows [18], laser beam [19], powder particle size [18] and transportation [20] are the main factors. In addition, the process parameters like energy density, scanning strategy, built direction and atmosphere environment as well as spattering can also affect the surface quality [21–25]. Usually, the methods used to improve the surface roughness mainly include forming control and post-treatment: forming control methods optimize the surface quality by adjusting process parameters, optimizing scanning strategy, adding magnetic field and ultrasonic assistance, etc. [26–28], while post-treatments use additional surface finishing methods, such as machine processing, chempolishing and shot-peening, to reduce surface roughness [29–31]. Besides, other novel methods of forming control and post-treatment were proposed to improve the surface quality as well, such as hybrid approach (combines additive manufacturing with subtractive manufacturing) [32], Abrasive Fluidized Bed (AFB) finishing [33], chempolishing and electropolishing [34] and laser polishing [35].
Re-melting finishing technology (both used in forming control and post-treatment) has been extensively studied in PBF and DED due to its excellent property modulation and surface roughness optimization ability [11,36,37]. In the literature, the vast majority reports in DED area are focusing on top surface re-melting [11]. For example, Yu et al. [38] used different energy densities to re-melt the upper surface of the DED parts in process. The results reveal that the roughness of top surface of the deposited layer and the intralayer porosity both reduced by increasing the laser re-melting energy density. An optimal laser re-melting energy density was found, under which the lowest surface roughness was obtained. The related reports of roughness optimization on side surface in DED are relatively few and they usually focus on post-treatment re-melting technology. For example, Bruzzo et al. [39] applied post-treatment of direct laser re-melting to side surface of thin-wall structures right after the DED deposition. Their results show that a reduction of 79% of Ra/Sa was achieved. However, few reports can be found that focus on re-melting the side surface with forming control using laser beam in DED process [31]. This is because the traditional parallel solid laser beams come from top and they are not able to irradiate to the side surfaces in the forming process. In this situation, the traditional methods cannot be applied to optimize the quality of side surfaces during the forming process. Besides, the previous studies with forming control mainly focused on the normal laser beam (refers to Gaussian solid beam) re-melting to optimize the surface roughness. It is expected that more studies shall be carried out and more advanced techniques shall be applied to verify the effectiveness of the re-melting techniques in DED area. For example, the attempt of applying a different innovative solution of laser beam mode with different energy distribution modes on the side surface re-melting process.
Aluminum alloy has excellent performances like high specific strength, good corrosion resistance, which are considered to be suitable for mechanical processing. It is widely used in light and heavy industries such as aerospace, automobile and building decoration. AlSi10Mg aluminum alloy is suitable for AM benefiting from the possession of hypoeutectic microstructure [40] and thus becoming one of the main research materials of AM technology due to its importance and huge demand in industry 4.0 [8,9]. However, the low melting point and the high liquidity make it to be more prone to surface deformation and powder adhesion than other materials, so that the problem of poor surface quality of aluminum alloy AM parts has not been properly solved. At present, the studies on surface roughness of AlSi10Mg PBF-parts are mature relatively [11]. As for DED, only a few studies have been published on the process parameters, material microstructure and properties [12]. Specifically, the research of roughness optimization on the side surface of AlSi10Mg DED-parts is still scarce, and there is a great need to develop innovative methods to pioneer this area.
In parallel with manufacturing process, the surface measurement and characterization are necessary to be carried out to provide useful feedback information [8,12,36,38]. The surface topography of AM parts usually exhibits high complexity and irregularity [42,46]. The surface texture mainly consists of tracks, deep recesses and sphere-like protrusions. The tracks come from the fusion and the subsequent solidification of a melt pool. Their texture directions also indicate the scanning direction of the laser beam path [41–43], and recesses may result from incomplete seams between tracks, balling phenomena or micropores at a small scale [42,44,45]. Sphere-like protrusions are formed mainly from unmolten or partially-melted powder particles and spatter particles (i.e. molten metal material ejected from the melt pool [46,47]. In previous studies, the most common used characterization technique is roughness characterization, in which the traditional roughness parameters, such as Ra/Sa, Rq/Sa, Rz/Sa etc., are evaluated [48–52]. Based on the literature review, the average roughness Ra/Sa is used in 90% of the Refs. [53]. However, the Ra/Sa is a statistical value that refers to the arithmetic mean of the absolute of the ordinate values within a definition profile/area. It cannot reveal the local information, namely the contribution of each individual feature to the overall roughness. Using feature-based characterization method can give a deep understanding of the roughness composition and can explain why is one surface more rough than another. In combination with re-melting technology, feature-based characterization method is helpful to give an insight of which exact surface features are optimized so that the overall roughness is reduced.
In this paper, an innovative surface finishing method using cone annular beam laser re-melting (CALR) is proposed for the purpose of optimizing the side surface quality of AlSi10Mg parts fabricated by DED. This method applies a unique laser beam and spot with novel energy distributions, which adopts the cone annular beam laser surface finishing with forming process and can improve roughness of side surface and top surface simultaneously. For the surface characterization, a feature-based characterization method is introduced to conduct multiscale surface analysis which enables a close view of the re-melting effects on different sub-scales. The feature-based characterization method allows a comprehensive understanding of the optimization mechanism of CALR and the formation mechanism of the surface topography. In order to verify the proposed methods, DED experiment of 316L stainless steel is carried out so that the universality and effectiveness of CALR are clarified.
2. Experimental setup, material and methodology
2.1 Feedstock materials
Aluminum alloy powder produced by the rotating electrode process was used in this work and the morphology of spherical AlSi10Mg particles is shown in Fig. 1(a). As can be seen, the morphology is spherically shaped with a spheroidization ratio of almost 100% and satellite particles (< 5.0 µm) are scarcely attached to larger particles. The diameters ranging from 75.0∼135.0 µm were selected and the particle size distribution of the powder was measured by a laser particle size analyzer (Mastersize 3000). The D10, D50 and D90 of the sieved powder are 90.2 µm, 107.1 µm and 134.6 µm respectively as shown in Fig. 1(b). The chemical compositions which were measured by an energy dispersive x-ray spectroscopy (EDX, Oxford Instruments X-Max 20, Oxford, UK) are presented in Table 1.
Table 1. Chemical composition of AlSi10Mg powder measured by EDX
2.2 DED experiment and CALR method
The robotic DED system used in this experiment is presented in Fig. 2(a). The laser source is a multimode Yb-fiber laser (IPG YLS-2000-CT Cambridge, MA, USA), which has a maximum power of 2 kW, a focused spot size of 800 µm and the wavelength λ is 1070 nm. The positioning and walking of the deposition head were carried out by a six-axis anthropomorphic robot (KUKA KR60-3F, Germany). A double-barrel powder feeder (GTV, Twin PF 2/2, Germany) and a self-designed annular beam laser DED deposition head were employed. Argon conveying was used both as carrier and shielding gas to ensure the oxygen content of molten pool is less than 500 ppm during the process.
The schematic of the DED is shown in Fig. 2(b), the Gaussian solid laser beam produced by the laser is transmitted to the deposition head by optical fiber. The collimated laser beam with a solid circular spot is obtained after the original laser beam is expanded and collimated by the collimating lens. The collimated laser beam is diffused and focused by a cone mirror and a ring mirror respectively. Consequently, the Gaussian solid laser beam is transformed into a shaped multimode annular laser beam as shown in Fig. 2(b) and Fig. 3(a). The main characteristics of the system are summarized in Table 2. Based on the experimental setup above, the annular laser beam is coupled with the powder beam sent from the central coaxial single tube. A molten pool is formed derived from their interaction. Thus, a high-precision DED forming can be realized.
Fig. 3. The energy integration along scanning direction based on the three-dimensional model of the spot energy distribution with different Df of the (a) cone annular laser beam: (b) Df = 0 mm, (c) Df = −3 mm, (d) Df = −5 mm.
Table 2. Main characteristics of the DED system
The deposition process was firstly fixed based on the setup in Table 2 to obtain similar samples before the use of CALR method. A thin-wall shape was selected as representative and it was fabricated with a height of 20 mm and a length of 58 mm on a ZL101 cast aluminum alloy piece substrate. The layer thickness was set with 0.4 mm as a constant throughout the whole process. The DED deposition was performed with the step strategy, which started with the single track deposition of the first layer, and then the constant height was set for stacking the next layer in each round to form the component. The parameters used for the deposition process are listed in Table 3, which were obtained from the previous work [54].
Table 3. Parameters of directed energy deposition process
In Fig. 3, the main principle of CALR is illustrated. The spot formed by the cone annular laser beam can evolve into three typical modes by adjusting the defocus value Df. As a result, there are three types of coupling styles: unsaturated style, saturated style and supersaturated style (see Figs. 3(b)-3(d)). The corresponding three-dimensional energy density distribution model of the laser spot and the energy integration along the scanning direction can be obtained by Eq. (1) and Eq. (2) respectively,
Unsaturated style: when Df = 0 mm (see Fig. 3(b)), the Gaussian solid laser beam with a diameter of 0.8 mm is obtained with a laser spot energy density model of central single-peak, this is typically used in traditional DED forming. Under this condition, the coupling of the laser beam and the DED part is unsaturated. The peak position of energy integral along scanning direction is coaxial with the centerline of the scanning path. The energy integral is an uneven shape with the characteristic of high in the center while low on sides (see Fig. 3(b)), which leads to an equal energy distribution of the molten pool. In this situation, the surface quality can get worse with a higher rate of powder adhesion because of the edge uneven fusion of deposited track.
Saturated style: when performing a general DED deposition, Df is set at −3 mm. In this situation, an annular beam with a 2.6 mm outer diameter and 1.1 mm inner diameter are formed for deposition. The saturated coupling of the laser beam and the top surface of the forming part are shown in Fig. 3(c): the three dimensional model of the energy density distribution of the laser spot is a ring-peak shaped with equal height and small radius. In addition, the energy integral of scanning direction is also visualized in Fig. 3(c) with energy intensity plane distribution of small radius shaped annular spot. The two peak positions of the energy integral are close to the centerline of the scanning path. As a result, the energy distribution of the molten pool is even, which can provide a preliminary surface quality optimization of the DED parts compared to the unsaturated style.
Supersaturated style: When Df = −5 mm (see Fig. 3(d)), a larger annular laser beam is formed with a 4.9 mm outer diameter and a 3.0 mm inner diameter. In this condition, the laser beam is supersaturated coupling with the top surface. The corresponding energy density distribution model exhibits a ring-peak whose radius is larger in comparison with the single-peak in Fig. 3(b). At the same time, the energy integration along the scanning direction is lower and wider, and the peak value is reduced by about 26% compared to Df = −3 mm. Because of the increased spot size, the two peak positions of energy integration move outward to a position, which is beyond the edge of thin-wall top surface 1.6 mm (see Fig. 3(d)). With the benefit of a larger cover volume of the cone annular laser beam, the high-density energy at the peak point focuses on the side surface of the thin-wall part (see Fig. 3(d)). In this way, the side re-melting processing of thin-wall parts can be realized e.g. by eliminating the stair-step effect, irregular edge fusion, powder adhesion and balling behavior.
In the experiment, the application of CALR forming control to DED deposition was conducted with following steps: the powder feeding metal deposition process using Df = −3 mm along the positive direction (see Fig. 4(a)) and the no powder feeding metal re-melting using Df = −5 mm along the negative direction (see Fig. 4(b)). With the help of the alternated implementation of the deposition in such a manner, the high-precision deposition can be realized for forming the thin-wall. The AlSi10Mg material was used in the experiment firstly with the described strategy above. Afterwards, the ordinary 316L stainless steel was used to verify the universality of the proposed method. The concrete investigation was divided into three parts: assessment of the feasibility of AlSi10Mg deposition with CALR method by adjusting the linear energy density (EL-CALR) obtained from Eq. (3)); analysis of the influences of variable speed on constant EL-CALR during the DED deposition of AlSi10Mg; verification of CALR method via orthogonal experiment using 316L stainless steel.
Fig. 4. Deposition process of annular laser beam with added CALR method: (a) metal deposition of annular laser beam; (b) CALR surface finishing of thin-wall side/top surface.
2.3 Feature-based surface characterization
In this work, surface texture was measured by a confocal laser scanning microscopy (CLSM) with the model Keyence VK-X1000, Japan. The 5X objective with the FOV (Field of View) of 2770 × 2076 µm2 was applied for the data acquisition. In Fig. 5, the acquired color image and 3D topography height map are presented. It can be seen that the surface texture is mainly comprised of three features: tracks, particle protrusions and random roughness.
Fig. 5. the color image (a) and the 3D topography (b) of the side surface measured by a confocal laser scanning microscope
It should be noted that very few grooves and valleys features are found in the formed surfaces and their discussions are out of scope in this paper. From the perspective of roughness analysis, the tracks belong to larger scale features which contribute the most to the overall roughness; the protrusions have a smaller scale which also plays a role in the roughness values, especially the value of Sz (the deviation of surface peak and surface valley); the random roughness features, which also belong to the small scale features, are usually considered to be the inherent or innate marks of the manufacturing process. For the purpose of surface quality optimization, the understanding of each feature on the contribution to the overall roughness value is crucial. Based on this knowledge, the reasonable efforts on suppressing and reducing the surface roughness can be further made. The feature-based characterization method is considered to be an appropriate method which is able to characterize the contributions of each feature to the total roughness.
At first, the form error is removed by using least-square fitting method. Because the side surface has a parabolic form, a parabolic surface model is chosen to conduct the fitting as following:
Fig. 6. Multi-scale surface decomposition: Form (a), S-F surface (b), L-F surface(c) and S-L surface (d) Fig. 7 Particle structure detection and segmentation on S-L surface: decomposed S-L surface (a); binary image after segmentation (b); labeled particle structure (c); S-L surface (without particle) (d)
In comparison with a normal Gaussian filter, the Rk filter applies a weighting function to reduce and eliminate the influences of the particle protrusion and outlier points on the mean surface extraction, which can lead to a more precise surface decomposition. In Fig. 6(c) and Fig. 6(d), L-F surface and S-L surface are visualized after implementing the Rk filter. It can be clearly seen that the L-F surface mainly contains the tracks features, while random roughness features and particles are the main features on the S-L surface.
In order to further separate the particle protrusion from the S-L surface, a threshold method is applied for particle segmentation and the threshhold value is chosen to be the RMS roughness, which means the height values above it are considered as the particle features. After segmentation, the particle can be visualized in a binary map as presented in Fig. 7(b).
Fig. 7. Particle structure detection and segmentation on S-L surface: decomposed S-L surface (a); binary image after segmentation (b); labeled particle structure (c); S-L surface (without particle) (d)
Subsequently, 8-pixel connectivity method is applied to detect the particle structure region. The function of this method is to detect and categorize which pixels belong to one structure region. In Fig. 7(c), the detected particle structures are labeled with different colors after using 8-pixel connectivity algorithm. Then, the particle structures can be further separated from the S-L surface by deleting the corresponding high values of the particle structures on the S-L topographic map as shown in Fig. 7(d).
At last, the representative roughness parameter Sa in each bandwidth is calculated and they are denoted with S-LSa, L-FSa, S-LSa and S-LSa-(without particle), respectively. In addition, the ρparticle (powder adhesion rate) is defined by the following equation,
where Aparticle denotes the particle structures’ area (the number of white pixels in Fig. 7(b)), while AFOV refers to the measurement area (total number of the pixels in Fig. 7(b)).3. Results and discussion
3.1 Assessment of feasibility and effectiveness on CALR method
Firstly, the track features derived from the stair-step effect can be clearly observed in Fig. 8(b)) which is extremely uneven because of the layer-wise deposition of the DED process with the saturated style. That is to say, the stair-step effect of the side surface of the as-deposited AlSi10Mg thin-wall sample was very severe without the CALR method. According to the coupling mechanism analysis of laser beam and as-deposited sample in Figs. 3–4, it can be seen that energy homogenization of laser spot occurred during the DED process with CALR (the supersaturated style), which alleviated the negative impact of Marangoni effect and Rayleigh instability in the molten pool effectively [57]. Thus, the CALR provided a preliminary optimization of the top surface. Meanwhile, the laser beam of CALR can directly finish the side surface by its conical irradiation (conical laser beam focused on the side surface of the deposited layer) based on the effect of re-melting and restructuring. As a result, the solidified irregular side surface becomes regular, the stair-step effect was eliminated (see Fig. 8(a)). In other words, the amplitude of the waviness features reduced on the side surface, thus a layer-wise surface finishing was realized and the side surface quality was improved. However, due to the limited processing window and energy density susceptibility of AlSi10Mg, the temperature field of the sample can be affected easily during the deposition process, which could cause edge collapsing, deposit deformation, and even failure of the forming.
Fig. 8. Experimental results of DED deposition with different EL-CALR: (a) principle of surface finishing and stair-step effect elimination (b) without CALR (0 J/mm); (c) low EL-CALR of 76.47 J/mm; (d) suitable EL-CALR of 94.12 J/mm; (e) high EL-CALR of 111.77 J/mm.
Secondly, an EL-CALR value 94.12 J/mm based on Eq. (3) was obtained as the initial energy density in this experiment which was equivalent of the original EL-DED. The purpose was to relieve the negative impact of the CALR on the temperature field and heat accumulation. Then, a higher EL-CALR (111.77 J/mm) and a lower EL-CALR (76.47 J/mm) were chosen with a step of 20 J/mm for investigating the impact of different EL-CALR values on the effect of the CALR. Here, this criterion of EL-CALR selection could be used as a universal method to obtain appropriate CALR parameters. The typical formed AlSi10Mg alloy thin-wall parts with the applied EL-CALR are presented in Figs. 8(c)-8(e). The experiment shows that the thin-wall can be formed under a lower EL-CALR (76.47 J/mm). However, it must be noted that the absorption rate of bulk aluminum is lower than that of the powder due to the existence of Brewster's angle effect (surfaces with non-uniform features tend to cause absorption events at high angles of incidence) [58] and the refractive re-absorption effect of the powder [59]. It would reduce irradiation energy of EL-CALR and the re-melting would be insufficient, hence the expected CALR surface finishing effect could not be achieved (see Fig. 8(c)). In contrast, when adopting a higher EL-CALR (111.77 J/mm), the unacceptable edge collapse and crack failure appeared on as-deposited thin-wall parts (see Fig. 8(e)) due to the intense heat accumulation effect, which eventually lead to a complete failure of forming. The experiment shows that only an appropriate process window of EL-CALR (e.g. 94.12 J/mm as shown in Fig. 8(d)) can achieve surface finishing optimization.
At last, the surface topographies of the S-F surfaces of the as-deposited aluminum alloys are presented in Figs. 8(b)-8(e) after conducting the feature-based characterization method. The S-F surface gives a first impression of how the surface texture was modified after using CALR method. The results indicate that the thin-wall produced by traditional DED (see Fig. 8(b)) has severe surface defects of raised stripes and slags, and the evaluated S-FSa value was larger than 22.4 µm. After finishing by CALR in forming process, the side surface S-FSa value has reduced to 22.2 µm, 8.1 µm and 17.1 µm respectively (see Figs. 8(c)-8(e)). This is mainly attributed to the side irradiation of the CALR laser (see Fig. 8(a)). In Fig. 8(d), it can be clearly seen that the best result was obtained under the suitable EL-CALR 94.12 J/mm, and there are no convex large particles or deep grooves, and the S-FSa value was reduced to 8.1 µm which was 63.8% lower than the original value of 22.4 µm without CALR.
3.2 Optimization strategy of CALR method
At first, the detailed results of the feature-based characterization in relation to CALR’s setup are presented in Fig. 9 which allows a fully understanding of the optimization mechanism of the CALR method. In Fig. 9(a), the scale-dependent Sa values of S-F surface, L-F surface, S-L surface and S-L surface (without particle) in correlation with different setups of CALR (high EL-CALR, low EL-CALR and suitable EL-CALR) are presented, while the ρparticle in relation with them is shown in Fig. 9(b). As previously stated, the S-FSa is considered as the overall roughness of the surface. After surface decomposition, the L-F surface, S-L surface and S-L surface (without particle) represent the sub-bands and each contains distinctive surface features. For instance, the component L-F surface (waviness band) has larger scale features like tracks, while S-L surface (roughness band) contains smaller scale surface like powder particle.
Fig. 9. Details of the characterization results in correlation with typical EL_CALR: (a) surface roughness Sa of different bandwidths (b) ρparticle under various EL-CALR
Secondly, it can be observed from the Fig. 9(a) that the S-FSa and L-FSa values tend to fall firstly and then rise. This trend can be explained with the help of EL-CALR: as the EL-CALR value increased, the finish effect of the CALR was more pronounced which led to a decreased value of S-FSa and L-FSa. However, when EL-CALR grows continually, excessive melting occurred which caused the collapse of the previously solidified layer that lost its support for the deposition of the next layer. In this situation, the deposition process failed. Therefore, a higher S-FSa and L-FSa were obtained. In addition, the S-LSa and S-LSa-(without particle) values are negatively correlated with EL-CALR (see Fig. 9(a)). This is because as the EL-CALR value increases, the molten pool surface tension reduced during the re-melting process. As a result, the wettability of the AlSi10Mg was optimized which strengthened the elimination effect of the balling behavior during the molten pool solidifying and also smoothed the small ripples on the surface.
Thirdly, it can be observed from the Fig. 9(b) that the ρparticle also has a negative correlation with EL-CALR. This can be explained that the adhered powder fusion ability of the CALR increased with the rise of the EL-CALR. It can be further observed that a decreased ρparticle in Fig. 9(b) can be reflected by a decreased deviation between S-LSa and S-LSa-(without particle) in Fig. 9(a).
At last, it can be clearly seen that S-LSa and L-FSa are highly correlated, and L-FSa component dominated the overall surface roughness in comparison with S-LSa and S-LSa-(without particle). As previously discussed, the best optimization of AlSi10Mg parts occurred when a suitable EL-CALR value (EL-CALR=94.12 J/mm in Fig. 9(a)) was chosen. In this condition, the decreased magnitudes of S-FSa and L-FSa are much larger than the corresponding decreased magnitudes of S-LSa and S-LSa-(without particle). In other words, the reduction of the surface roughness was mainly caused by the decreased L-FSa values. Remember that the L-F surface mainly contains the track transition features which were caused by stair-step effect. This result reveals the optimization mechanism using CALR method was mainly realized by eliminating the stair-step effect, while the optimization via reducing the powder microstructures and random roughness features on the surface was less significant.
3.3 Influence of CALR scanning speed
AlSi10Mg alloy was further deposited with CALR method under different scanning speed with the suitable EL-CALR 94.12 J/mm, in order to investigate the influences of the speed on surface finish. The scanning speed was set from 11 mm/s to 19 mm/s with a step of 2 mm/s and the suitable EL-CALR 94.12 J/mm was selected based on section 3.1 of this paper. The results are presented in Fig. 10–11. In Fig. 10, the 2D color images and surface topographies of different scales are visualized as the scanning speed varies. In Fig. 11, the corresponding Sa values and the ρparticle of different surfaces against scanning speed are presented. As can be seen from Fig. 11, the representative roughness parameter L-FSa climbs to the peak firstly and then falls as the scanning speed rises. When scanning speed was set at 11 mm/s, a relatively high L-FSa value was obtained. This is because the absorbed energy was partially dispersed by the heat conduction for the high thermal conductivity of aluminum alloy (155 W/(m·k)). The CALR molten pool has a lower surface temperature, and thus leading to a high surface tension that weakened the ability of the stair-step effect elimination. When scanning speed grows from 11 mm/s to 15 mm/s continually, the L-FSa value decreases. This is because the constant EL-CALR 94.12 J/mm adjusted the corresponding power from 1 kW to 1.4 kW synchronously based on Eq. (3). Consequently, the surface temperature went up, aluminum can be fused better for the reduction of surface tension. Thus, the stair-step effect can be eliminated effectively. When the scanning speed rises further from 15 mm/s to 19 mm/s continually, the outer metal of AlSi10Mg thin-wall part solidified before it can be re-constructed and re-fused together. In this situation, the surface finishing using CALR was limited, which caused the ascent of roughness. As previously stated, the L-FSa value indicates track transition features of the stair-step effect. The change of the L-FSa value along with the scanning speed can also be reflected by the surface morphology in the 2D color images in Fig. 10.
Fig. 10. Surface roughness Sa of different CALR scanning speed: (a) 0 mm/s (without CALR); (b) 11 mm/s; (c) 15 mm/s; (d) 19 mm/s.
Fig. 11. Variation of surface roughness Sa with different scanning speed of CALR finishing: (a) average height of surface profile Sa; (b) powder adhesion rate ρparticle.
On the contrary, ρparticle and S-LSa reach to the peak firstly and then decline as scanning speed grows from 11 mm/s to 19 mm/s gradually (see Fig. 11(a)). This is because surface tension decreased within the scanning speed range from 11 mm/s to 15 mm/s, which lead to an enhanced effect of stair-step effect elimination. In spite of it, the decreased tension also reduced the energy required for powder particle adhesion simultaneously, and this strengthened the trend of powder adhesion during the next layer DED deposition as can be seen with the ρparticle curve in Fig. 11(b). This phenomenon can also be observed in the 2D color image in Fig. 10(c): the stair-step effect was eliminated but the ρparticle increased. When the scanning speed is within the range from 15 mm/s to 19 mm/s, higher temperature was obtained and the particles on AlSi10Mg thin-wall part were re-melted, which finally caused the fall of ρparticle. The corresponding evidence with surface topographies can also be found in Fig. 10 with S-L surface and S-L surface (without particle). Besides, the Fig. 11(a) once again gives a clue that the L-FSa value dominates the overall surface roughness S-FSa. Although, the increased ρparticle led to the rise of S-LSa. The stair-step effect was eliminated simultaneously which caused the decline of L-FSa. As a result, the overall roughness S-FSa also decreased because of the dominance of L-FSa.
In a short summary, the results demonstrate that the scanning speed has a significant effect on the finishing effect of the CALR. An appropriate scanning speed of 15 mm/s can promote the finishing effect. A thin-wall side surface roughness S-FSa of 7.1 µm was achieved, which is 68.3% less than the original value of 22.4 µm without CALR, also less than the results by PBF [12]. The top surface S-FSa = 1.3 µm, which is slightly lower than the result of optimal AlSi10Mg top surface roughness produced using PBF techniques [23]. To sum up, the top surface and the side surface were both optimized because of the simultaneous radiation ability of CALR. High quality aluminum alloy thin-wall surface is achieved by the developed method.
3.4 Verification of 316L stainless steel
At last, the 316L stainless steel was used with the proposed CALR method to verify its universality. Similarly, the parameters for 316L were also obtained from the previous work [60]. Then, the initial EL-CALR (100.0 J/mm) was obtained based on Eq. (3), a higher EL-CALR (123.08 J/mm) and a lower EL-CALR (62.50 J/mm) were chosen with a step of 25 J/mm. The results are shown in Fig. 12–13. As can be seen that the Sa value in each scale-dependent surface falls continually with the increase of EL-CALR. In other words, the amplitude of the waviness features caused by track transition and the ρparticle of S-L bandwidth reduced simultaneously. Consequently, the two main roughness ingredients have been reduced greatly, the side surface quality was improved. This is attributed to the good processability and the wide processing window of 316L stainless steel. That is to say, the iron-based material 316L stainless has a higher melting point and lower energy density susceptibility, which has the characteristics of a low possibility for collapsing caused by heat accumulation. Thus, these characteristics allowed the optimization under a higher EL-CALR.
Fig. 12. Typical results of CALR method verification on 316L stainless steel: (a) no CALR finishing treatment (0 J/mm); (b) low EL-CALR of 62.50 J/mm; (c) suitable EL-CALR of 100.00 J/mm; (d) high EL-CALR of 123.08 J/mm
Fig. 13. The curve of the surface roughness Sa with different EL-CALR: (a) average height of surface profile Sa; (b) surface adhesion rate ρparticle.
In detail, as the EL-CALR increases from 0 J/mm to 123.08 J/mm, the L-FSa decreases continually (see Fig. 13) because of the stair-step effect elimination process of surface re-melting reconstruction. At the same time, the fusion of adhesion powder and the elimination of balling behavior cause the decline of ρparticle (see Fig. 13(b)), which eventually lead to the decline of S-LSa value. As a result, the overall roughness S-FSa decreases from 15.1 µm to 2.9 µm as the EL-CALR grows (see Fig. 13(a)), the optimal EL-CALR of 123.08 J/mm was obtained. Finally, the final optimized side surface roughness S-FSa is 2.9 µm, which is 80.8% less than the original value of 15.1 µm by using DED without CARL. The top surface S-FSa = 0.4 µm, which is less than the optimal top surface roughness produced by DED or PBF [20,32].
The verification experiments above imply that the CALR method can be successfully applied to the DED deposition of 316L stainless steel and it has the potential to be applied to other materials and forming processes to become a universal method.
4. Conclusion and outlook
In this work, a novel surface finishing method of CALR has been presented to realize the quality optimization of AlSi10Mg thin-wall side surface. Meanwhile, a feature-based characterization method was proposed. With the help of it, the optimization mechanism of the CALR method was revealed. The main conclusions are summarized as follows:
- 1. The DED deposition process with the CALR method was designed based on the spot controllability of the cone annular laser beam. The DED deposited AlSi10Mg thin-wall part with CALR method was realized. The spot mode, energy density distribution model and energy integration along scanning direction were calculated and analyzed with a variable Df.
- 2. By adopting the feature-based surface characterization method, scale-dependent surfaces were obtained. With the help of feature segmentation technique, the powder particles can be extracted from the S-L surface and analyzed separately. Together with the CALR method, it can be concluded that improved surface roughness via CALR method is mainly attributed to the optimization of stair-step effect, where the amplitude of the track transition features of the L-F surface is damped. The effect of damping the amplitude of the powder particle features of S-L surface is however less significant.
- 3. With the increase of EL-CALR, the finishing effect of CALR for AlSi10Mg thin-wall parts was becoming more obvious and even led to failure forming of DED finally. Under the optimal EL-CALR (94.12 J/mm), the side surface with S-FSa = 8.1 µm was achieved, which is 63.8% lower than the original value of 22.4 µm without CALR. The scanning speed has a significant influence on the surface finishing of the CALR. Under the optimal scanning speed (15 mm/s), the side surface quality is further improved with S-FSa = 7.1 µm, which is 68.3% less than the original value of 22.4 µm. The top surface quality is also improved with S-FSa = 1.6 µm because of the simultaneous radiation and the optimization ability of CALR.
- 4. The CALR method can be well applied to 316L stainless and has the potential to be universally applicable to other materials and forming processes. In comparison with AlSi10Mg, the 316L stainless allows the optimization under a higher EL-CALR, under which the L-FSa and the ρparticle reduced simultaneously. The optimized side surface was obtained with L-FSa = 2.9 µm, which is 80.8% less than the original value of 15.1 µm, and the top surface remains high quality with S-FSa = 0.4 µm.
In the further, the research might be expanded to other materials such as Titanium, Nickel and techniques like wire + arc additive manufacturing. It is expected that the developed methods in this work could eventually make contribution to the progress of AM technology.
Funding
Natural Science Research of Jiangsu Higher Education Institutions of China (21KJB460025); National Natural Science Foundation of China (52105197).
Acknowledgments
The authors would like to thank the National Natural Science Foundation of China (52105197) and Natural Science Research of Jiangsu Higher Education Institutions of China (21KJB460025) for their financial support.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. D. Herzog, V. Seyda, E. Wycisk, and C. Emmelmann, “Additive manufacturing of metals,” Acta Mater. 117, 371–392 (2016). [CrossRef]
2. T. DebRoy, H. L. Wei, J. S. Zuback, T. Mukherjee, J. W. Elmer, J. O. Milewski, A. M. Beese, A. Wilson-Heid, A. De, and W. Zhang, “Additive manufacturing of metallic components – Process, structure and properties,” Prog. Mater. Sci. 92, 112–224 (2018). [CrossRef]
3. P. A. Colegrove, H. E. Coules, J. Fairman, F. Martina, T. Kashoob, H. Mamash, and L. D. Cozzolino, “Microstructure and residual stress improvement in wire and arc additively manufactured parts through high-pressure rolling,” J. Mater. Process. Technol. 213(10), 1782–1791 (2013). [CrossRef]
4. P. Kazanas, P. Deherkar, P. Almeida, H. Lockett, and S. Williams, “Fabrication of geometrical features using wire and arc additive manufacture,” Proc. Inst. Mech. Eng., Part B 226(6), 1042–1051 (2012). [CrossRef]
5. J. Wu, X. Q. Wang, W. Wang, M. M. Attallah, and M. H. Loretto, “Microstructure and strength of selectively laser melted AlSi10Mg,” Acta Mater. 117, 311–320 (2016). [CrossRef]
6. P. Wang, X. Tan, M. L. S. Nai, S. B. Tor, and J. Wei, “Spatial and geometrical-based characterization of microstructure and microhardness for an electron beam melted Ti-6Al-4V component,” Mater. Des. 95, 287–295 (2016). [CrossRef]
7. F. Weng, C. Chen, and H. Yu, “Research status of laser cladding on titanium and its alloys: A review,” Mater. Des. 58, 412–425 (2014). [CrossRef]
8. M. Vaezi, S. Chianrabutra, B. Mellor, and S. Yang, “Multiple material additive manufacturing - Part 1: A review: This review paper covers a decade of research on multiple material additive manufacturing technologies which can produce complex geometry parts with different materials,” Virtual Phys. Prototyp. 8(1), 19–50 (2013). [CrossRef]
9. M. Seabra, J. Azevedo, A. Araújo, L. Reis, E. Pinto, N. Alves, R. Santos, and J. Pedro Mortágua, “Selective laser melting (SLM) and topology optimization for lighter aerospace componentes,” Procedia Struct. Integr. 1, 289–296 (2016). [CrossRef]
10. R. Hague, S. Mansour, and N. Saleh, “Material and design considerations for Rapid Manufacturing,” Int. J. Prod. Res. 42(22), 4691–4708 (2004). [CrossRef]
11. J. Zhang, B. Song, Q. Wei, D. Bourell, and Y. Shi, “A review of selective laser melting of aluminum alloys: Processing, microstructure, property and developing trends,” J. Mater. Sci. Technol. 35(2), 270–284 (2019). [CrossRef]
12. I. M. Kusoglu, B. Gökce, and S. Barcikowski, “Research trends in laser powder bed fusion of Al alloys within the last decade,” Addit. Manuf. 36, 101489 (2020). [CrossRef]
13. P. Fathi, M. Rafieazad, X. Duan, M. Mohammadi, and A. M. Nasiri, “On microstructure and corrosion behaviour of AlSi10Mg alloy with low surface roughness fabricated by direct metal laser sintering,” Corros. Sci. 157, 126–145 (2019). [CrossRef]
14. A. Fatemi, R. Molaei, S. Sharifimehr, N. Phan, and N. Shamsaei, “Multiaxial fatigue behavior of wrought and additive manufactured Ti-6Al-4V including surface finish effect,” Int. J. Fatigue 100, 347–366 (2017). [CrossRef]
15. S. Bagehorn, J. Wehr, and H. J. Maier, “Application of mechanical surface finishing processes for roughness reduction and fatigue improvement of additively manufactured Ti-6Al-4V parts,” Int. J. Fatigue 102, 135–142 (2017). [CrossRef]
16. S. Thuketana, C. Taute, H. Möller, and A. du Plessis, “Characterization of surface roughness and subsurface pores and their effect on corrosion in 3D-printed AlSi10Mg,” J. South. Afr. Inst. Min. Metall. 120(6), 369–376 (2020). [CrossRef]
17. A. Lerebours, P. Vigneron, S. Bouvier, A. Rassineux, M. Bigerelle, and C. Egles, “Additive manufacturing process creates local surface roughness modifications leading to variation in cell adhesion on multifaceted TiAl6V4 samples,” Bioprinting 16, e00054 (2019). [CrossRef]
18. B. He, D. Li, A. Zhang, J. Ge, X. Do, H. Xie, and H. Yang, “Influence of powder flow on sidewall quality of solid parts in laser metal direct forming,” Int. J. Adv. Manuf. Technol. 68(9-12), 2703–2711 (2013). [CrossRef]
19. G. Zhu, D. Li, A. Zhang, G. Pi, and Y. Tang, “The influence of laser and powder defocusing characteristics on the surface quality in laser direct metal deposition,” Opt. Laser Technol. 44(2), 349–356 (2012). [CrossRef]
20. A. B. Spierings, N. Herres, and G. Levy, “Influence of the particle size distribution on surface quality and mechanical properties in AM steel parts,” Rapid Prototyp. J. 17(3), 195–202 (2011). [CrossRef]
21. D. Wang, Y. Yang, X. Su, and Y. Chen, “Study on energy input and its influences on single-track,multi-track, and multi-layer in SLM,” Int. J. Adv. Manuf. Technol. 58(9-12), 1189–1199 (2012). [CrossRef]
22. M. Mohammadi and H. Asgari, “Achieving low surface roughness AlSi10Mg_200C parts using direct metal laser sintering,” Addit. Manuf. 20, 23–32 (2018). [CrossRef]
23. X. Han, H. Zhu, X. Nie, G. Wang, and X. Zeng, “Investigation on Selective Laser Melting AlSi10Mg Cellular Lattice Strut: Molten Pool Morphology, Surface Roughness and Dimensional Accuracy,” Materials 11(3), 392 (2018). [CrossRef]
24. S. R. Ch, A. Raja, P. Nadig, R. Jayaganthan, and N. J. Vasa, “Influence of working environment and built orientation on the tensile properties of selective laser melted AlSi10Mg alloy,” Mater. Sci. Eng., A 750, 141–151 (2019). [CrossRef]
25. U. Ali, R. Esmaeilizadeh, F. Ahmed, D. Sarker, W. Muhammad, A. Keshavarzkermani, Y. Mahmoodkhani, E. Marzbanrad, and E. Toyserkani, “Identification and characterization of spatter particles and their effect on surface roughness, density and mechanical response of 17-4 PH stainless steel laser powder-bed fusion parts,” Mater. Sci. Eng., A 756, 98–107 (2019). [CrossRef]
26. M. Gharbi, P. Peyre, C. Gorny, M. Carin, S. Morville, P. Le Masson, D. Carron, and R. Fabbro, “Influence of various process conditions on surface finishes induced by the direct metal deposition laser technique on a Ti-6Al-4V alloy,” J. Mater. Process. Technol. 213(5), 791–800 (2013). [CrossRef]
27. Y. Zhao, T. Yu, J. Sun, W. Xi, and X. Bi, “Effect of laser cladding on forming qualities of YCF101 alloy powder in the different lap joint modes,” Int. J. Adv. Manuf. Technol. 96(5-8), 1991–2001 (2018). [CrossRef]
28. K. L. Tan and S. H. Yeo, “Surface modification of additive manufactured components by ultrasonic cavitation abrasive finishing,” Wear 378-379, 90–95 (2017). [CrossRef]
29. U. Ali, H. Fayazfar, F. Ahmed, and E. Toyserkani, “Internal surface roughness enhancement of parts made by laser powder-bed fusion additive manufacturing,” Vacuum 177, 109314 (2020). [CrossRef]
30. E. Maleki, S. Bagherifard, M. Bandini, and M. Guagliano, “Surface post-treatments for metal additive manufacturing: Progress, challenges, and opportunities,” Addit. Manuf. 37, 101619 (2021). [CrossRef]
31. N. N. Kumbhar and A. V. Mulay, “Post Processing Methods used to Improve Surface Finish of Products which are Manufactured by Additive Manufacturing Technologies: A Review,” J. Inst. Eng. Ser. C. 99(4), 481–487 (2018). [CrossRef]
32. P. Wüst, A. Edelmann, and R. Hellmann, “Areal surface roughness optimization of maraging steel parts produced by hybrid additive manufacturing,” Materials 13(2), 418 (2020). [CrossRef]
33. E. Atzeni, M. Barletta, F. Calignano, L. Iuliano, G. Rubino, and V. Tagliaferri, “Abrasive Fluidized Bed (AFB) finishing of AlSi10Mg substrates manufactured by Direct Metal Laser Sintering (DMLS),” Addit. Manuf. 10, 15–23 (2016). [CrossRef]
34. P. Tyagi, T. Goulet, C. Riso, R. Stephenson, N. Chuenprateep, J. Schlitzer, C. Benton, and F. Garcia-Moreno, “Reducing the roughness of internal surface of an additive manufacturing produced 316 steel component by chempolishing and electropolishing,” Addit. Manuf. 25, 32–38 (2019). [CrossRef]
35. D. Bhaduri, P. Penchev, A. Batal, S. Dimov, S. L. Soo, S. Sten, U. Harrysson, Z. Zhang, and H. Dong, “Laser polishing of 3D printed mesoscale components,” Appl. Surf. Sci. 405, 29–46 (2017). [CrossRef]
36. B. Xin, X. Zhou, G. Cheng, J. Yao, and Y. Gong, “Microstructure and mechanical properties of thin-wall structure by hybrid laser metal deposition and laser remelting process,” Opt. Laser Technol. 127, 106087 (2020). [CrossRef]
37. X. hui Chen, B. Chen, X. Cheng, G. chao Li, and Z. Huang, “Microstructure and properties of hybrid additive manufacturing 316L component by directed energy deposition and laser remelting,” J. Iron Steel Res. Int. 27(7), 842–848 (2020). [CrossRef]
38. Z. Yu, Y. Zheng, J. Chen, C. Wu, J. Xu, H. Lu, and C. Yu, “Effect of laser remelting processing on microstructure and mechanical properties of 17-4 PH stainless steel during laser direct metal deposition,” J. Mater. Process. Technol. 284, 116738 (2020). [CrossRef]
39. F. Bruzzo, G. Catalano, A. G. Demir, and B. Previtali, “Surface finishing by laser re-melting applied to robotized laser metal deposition,” Opt. Lasers Eng. 137, 106391 (2021). [CrossRef]
40. X. P. Li, X. J. Wang, M. Saunders, A. Suvorova, L. C. Zhang, Y. J. Liu, M. H. Fang, Z. H. Huang, and T. B. Sercombe, “A selective laser melting and solution heat treatment refined Al-12Si alloy with a controllable ultrafine eutectic microstructure and 25% tensile ductility,” Acta Mater. 95, 74–82 (2015). [CrossRef]
41. J. Mazumder, L. Aided, and M. Processing, “Overview of melt dynamics in laser processing,” Opt. Eng. 30(8), 1208–1219 (1991). [CrossRef]
42. N. Senin, A. Thompson, and R. K. Leach, “Characterisation of the topography of metal additive surface features with different measurement technologies,” Meas. Sci. Technol. 28(9), 095003 (2017). [CrossRef]
43. T. Özel, A. Altay, A. Donmez, and R. Leach, “Surface topography investigations on nickel alloy 625 fabricated via laser powder bed fusion,” Int. J. Adv. Manuf. Technol. 94(9-12), 4451–4458 (2018). [CrossRef]
44. H. Gong, H. Gu, K. Zeng, J. J. S. Dilip, D. Pal, B. Stucker, D. Christiansen, J. Beuth, and J. J. Lewandowski, Melt pool characterization for selective laser melting of Ti-6Al-4V pre-alloyed powder, 25th Annu. Int. Solid Free. Fabr. Symp. An Addit. Manuf. Conf. SFF 2014. 256–267 (2014).
45. N. K. Tolochko, S. E. Mozzharov, I. A. Yadroitsev, T. Laoui, L. Froyen, V. I. Titov, and M. B. Ignatiev, “Balling processes during selective laser treatment of powders,” Rapid Prototyp. J. 10(2), 78–87 (2004). [CrossRef]
46. R. K. Leach, D. Bourell, S. Carmignato, A. Donmez, N. Senin, and W. Dewulf, “Geometrical metrology for metal additive manufacturing,” CIRP Ann. 68(2), 677–700 (2019). [CrossRef]
47. B. Rosa, A. Brient, S. Samper, and J. Hascoët, “Influence of additive laser manufacturing parameters on surface using density of partially melted particles,” Surf. Topogr.: Metrol. Prop. 4(4), 045002 (2016). [CrossRef]
48. D. Ahn, J. Kweon, S. Kwon, J. Song, and S. Lee, “Representation of surface roughness in fused deposition modeling,” J. Mater. Process. Technol. 209(15-16), 5593–5600 (2009). [CrossRef]
49. R. I. Campbell, M. Martorelli, and H. S. Lee, “Surface roughness visualisation for rapid prototyping models,” CAD Comput. Aided Des. 34(10), 717–725 (2002). [CrossRef]
50. F. Calignano, D. Manfredi, E. P. Ambrosio, L. Iuliano, and P. Fino, “Influence of process parameters on surface roughness of aluminum parts produced by DMLS,” Int. J. Adv. Manuf. Technol. 67(9-12), 2743–2751 (2013). [CrossRef]
51. M. Król, L. A. Dobrzański, L. Reimann, and I. Czaja, “Surface quality in selective laser melting of metal powders,” Arch. Mater. Sci. Eng. 60, 87–92 (2013).
52. K. Mumtaz and N. Hopkinson, “Top surface and side roughness of Inconel 625 parts processed using selective laser melting,” Rapid Prototyp. J. 15(2), 96–103 (2009). [CrossRef]
53. A. Townsend, N. Senin, L. Blunt, R. K. Leach, and J. S. Taylor, “Surface texture metrology for metal additive manufacturing: a review,” Precis. Eng. 46, 34–47 (2016). [CrossRef]
54. L. Wan, S. Shi, Z. Xia, T. Shi, Y. Zou, K. Li, and X. Chen, “Directed energy deposition of CNTs / AlSi10Mg nanocomposites: Powder preparation, temperature field, forming, and properties,” Opt. Laser Technol. 139, 106984 (2021). [CrossRef]
55. B. Muralikrishnan and J. Raja, Computational surface and roundness metrology, springer, London, 2009. https://doi.org/10.1007/978-1-84800-297-5.
56. J. Seewig, “Linear and robust Gaussian regression filters,” J. Phys.: Conf. Ser. 13, 254–257 (2005). [CrossRef]
57. R. Liu, Q. S. Liu, and S. C. Zhao, “Influence of Rayleigh effect combined with Marangoni effect on the onset of convection in a liquid layer overlying a porous layer,” Int. J. Heat Mass Transfer 51(25-26), 6328–6331 (2008). [CrossRef]
58. L. K. Ang, Y. Y. Lau, R. M. Gilgenbach, and H. L. Spindler, “Analysis of laser absorption on a rough metal surface,” Appl. Phys. Lett. 70(6), 696–698 (1997). [CrossRef]
59. Y. Yang, D. Gu, D. Dai, and C. Ma, “Laser energy absorption behavior of powder particles using ray tracing method during selective laser melting additive manufacturing of aluminum alloy,” Mater. Des. 143, 12–19 (2018). [CrossRef]
60. T. Shi, J. Shi, Z. Xia, B. Lu, S. Shi, and G. Fu, “Precise control of variable-height laser metal deposition using a height memory strategy,” J. Manuf. Process. 57, 222–232 (2020). [CrossRef]
