Effect of additive and subtractive hybrid manufacturing process on the surface quality of 18Ni300 maraging steel

The effect of process parameters on the surface quality of 18Ni300 maraging steel formed by selective laser melting (SLM) was investigated. Surface modification of SLM specimens was performed using milling as a subtraction method to investigate the effect of milling process parameters on the surface quality of SLM specimens. Comparing and analyzing the surface quality after additive and subtractive processes, the results show that the increase of laser power during SLM can improve the surface morphology, but there is always a balling effect. The surface quality deteriorates when the scanning speed increases. When the laser power increases or the scanning speed decreases, the microhardness increases and the error decreases. The residual stress does not vary linearly with the change of laser power or scanning speed, and the scanning speed has a greater effect on the residual stress than the laser power. The best surface quality was achieved with a laser power of 180 W and a scanning speed of 300 mm s−1. The laser power and scanning speed did not significantly affect the microstructure of the SLM-formed specimens. In the milling process, an increase in the feed rate will make the surface quality worse, and an increase in the cutting speed will make the surface quality better. The best surface quality was obtained with a cutting speed of 10 m min−1 and a feed rate of 36 mm min−1. The grain refinement effect is weakened when the feed rate is increased, and the grain refinement effect is enhanced when the cutting speed is increased. The surface quality of SLM-formed maraging steel specimens improved somewhat after milling.


Introduction
The mold is known as the 'mother of industry', and its accuracy and quality directly determine the accuracy and quality of molded parts. 18Ni300 is a typical maraging steel with the advantages of high strength, good toughness, and low thermal distortion, which is widely used in the field of precision tooling. Maraging steel is expensive and often used for the parts with small quantities, complex structures and high mechanical property requirements, such as the mold with conformal cooling channels, which is difficult to manufacture using conventional processes. The selective laser melting process has the advantages of a high degree of individualization, freedom of design and manufacturing, so 18Ni300 maraging steel is suitable for laser additive forming.
Selective laser melting (SLM) is an additive manufacturing process that uses laser as the energy source to melt powder materials layer by layer according to three-dimensional model data, which can realize the highperformance precision manufacturing of metal parts with complex structural. The mechanical properties of selective laser melted parts are comparable to those of traditional forgings, so they are widely used in aerospace, medical, injection mold, automotive and other industries [1]. At present, the research objects of SLM mainly focus on four metals, namely, iron [2], nickel [3], aluminum [4], and titanium [5], and the main research content is the influence of process parameter optimization, heat treatment, and powder composition on SLM forming.
For SLM forming 18Ni300 maraging steel, most of them also focus on optimizing process parameters and heat treatment. For example, Yuchao Bai's team investigated the effect of SLM process parameters on the relative density of laser formed maraging steel [6], and their results showed that low laser power, high scanning speed or scanning spacing can reduce the relative density. In addition, Yuchao Bai's team investigated the changes in structure and mechanical properties [7], and the changes in machinability [8] of SLM-formed maraging steel after heat treatment. The study of Gan Li et al [9] showed that the relative density of SLM-formed maraging steel increases rapidly and then remains almost constant when the laser energy density increases, and the hardness increases as a whole with increasing relative density. Kempen et al [10] found that SLM can produce almost completely dense maraging steel parts with mechanical properties comparable to conventionally produced maraging steel, and that both higher powder layer thickness and scanning speed will lead to a reduction in density and macrohardness. Ferreira et al [11] pointed out that energy density could not fully explain the properties of SLM-formed 18Ni300, so a prediction model for density and microhardness was developed using the response surface method. Yanan et al [12] found that the laser energy is not as high as possible, high laser power and low scanning speed will make the melt pool instability due to overheating and cause defects such as pores and cracks in the formed parts, while low laser power and high scanning speed will make the formed parts have unfused defects due to insufficient energy. Li et al [13] investigated the mechanism and mechanical properties of SLM forming 18Ni300 maraging steel and found that the specimens of 18Ni300 formed by this process were dominated by columnar and equiaxed crystals, and the fracture mode of tensile specimens was mixed failure (brittle and ductile fracture). Zhenjiang et al [14] investigated the effect of the texture of SLMformed 18Ni300 maraging steel on the mechanical properties, and found that the diverse crystal texture could coordinate the plastic deformation of the specimens more adequately than the single crystal texture.. The present study of SLM-formed 18Ni300 maraging steel has seen little analysis of surface morphology and residual stresses, which are important components of surface integrity and are therefore necessary for investigation.
Even though additive manufacturing has the advantages of free design and free forming, the surface quality and dimensional accuracy of its formed parts cannot be compared to those of traditional machined subtractive manufacturing, and thus the applications of additive manufacturing techniques will be limited. In view of the respective advantages and disadvantages of additive manufacturing and machining subtractive manufacturing, scholars combine the two to form the additive and subtractive hybrid manufacturing technology (ASHM). The additive and subtractive hybrid manufacturing technology not only has the advantage of high degree of freedom of forming, but also overcomes the defects of poor surface quality and large size error by means of machining [15]. As an advanced manufacturing technology, the additive and subtractive manufacturing technology can rapidly form products and put them into service, which has a very broad application prospect. Shuai Zhang et al [16] studied the effects of wire arc additive manufacture (WAAM) and milling composite processing on residual stresses and tensile anisotropy of aluminum alloy, and found that milling can significantly reduce the residual stresses left by wire arc additive manufacturing, and make the tensile anisotropy increase with the increase of milling thickness. Thomas Feldhausen et al [17] investigated the mechanical properties of 316 L stainless steel manufactured by additive and subtractive composite, and the results showed that the average relative porosity was reduced by 83% and the average relative elongation at break was increased by 71% for additive and subtractive composite manufacturing compared with wire and laser additive manufacturing. Rangasayee Kannan et al [18] investigated the deformation mechanism of 316 L stainless steel fabricated by laser wire deposition and additive/subtractive hybrid manufacturing, and found that the main deformation mechanism of 316 L was twinning in both additive and additive/subtractive hybrid manufacturing, while the twinning ratio obtained by hybrid manufacturing was relatively low and therefore the elongation at fracture was increased. Yanhua et al [19] studied the stress coupling mechanism of laser additive and milling subtractive hybrid manufacturing for FeCr alloy, and the results showed that milling can generate residual compressive stresses on the near surface of FeCr alloy manufactured by laser additive, and the magnitude of compressive stresses is the result of the combined effect of mechanical load and thermal load caused by milling. So far, there is little literature on the application of additive and subtractive hybrid manufacturing technology to 18Ni300 maraging steel.
Although the SLM technology has a high degree of freedom in forming, its formed parts are not destined for direct use as finished products due to their poor surface quality. 18Ni300 maraging steel as a mold steel, after laser additive forming into the mold with conformal cooling channels, need to be machined to improve the surface integrity, such as reducing surface roughness, increasing surface hardness, improving the distribution of residual stress, etc. Surface integrity has a large impact on the use of molds, and the existing literature rarely deals with the effect of machining on the surface integrity of laser additive forming mold steel, so this paper is dedicated to investigate the application of additive and subtractive hybrid manufacturing process to 18Ni300 maraging steel. The laser selective melting process was used to form 18Ni300 maraging steel specimens to study the effect of process parameters on the surface quality of laser formed specimens, and to obtain the optimal process parameters for laser selective melting of maraging steel. The surface modification treatment of SLM specimens was performed by milling as a subtractive method, and the effect of milling process parameters on the surface quality was investigated to derive the optimal process parameters for milling the SLM specimens. This research can provide ideas for the process research of die steel by additive and subtractive hybrid manufacturing.

Power and SLM equipment
The 18Ni300 maraging steel powders with a normal distribution of 15-45 μm was prepared by gas atomization method. The powder morphology is approximately spherical as shown in figure 1, and the chemical composition is shown in table 1. The SLM process was accomplished by Mlab 200 R selective laser melting equipment, whose appearance and principle of which are shown in figure 2. The equipment consists of a 200 W Yb fiber laser (wavelength: 1070 nm; spot diameter: 70-100 μm), scanning system and an f-θ lens. Argon gas was used to protect the material from oxidation during the manufacturing process.

SLM forming scheme
The substrate used in the SLM process is 45 # steel, with each layer rotated 90 degrees during the printing process, a powder thickness of 40 μm, and a scan spacing of 100 μm. The scanning method is the island exposure scanning strategy, which is illustrated in figure 3 in principle. The island exposure scanning strategy was developed from a checkerboard scanning strategy, the core of which is to divide each layer of powder into a finite number of small 'islands', and then randomly expose the 'islands' at different locations. The Island exposure scanning strategy can reduce the energy accumulation caused by long scanning lines, effectively reduce the thermal stresses created during the melting and solidification of the metal, and prevent warpage and distortion of the printed part. SLM process parameters are shown in table 2. The sample size was 10 mm * 10 mm * 5 mm (height 5 mm), and one sample was printed for parameters 1, 2, 4, 5, 6, 7 and 8 respectively for the testing after SLM. Six samples were printed for parameter 3, one of which is used for the testing after SLM, and the other five were used for subsequent milling experiments.

Milling experiment
The specimens prepared by the third group of SLM process parameters were used as the objects of milling experiments. The milling process parameters are given in table 3. The SLM specimens were milled on an XL5036A vertical lifting table milling machine with a 16 mm diameter HSS three-edge end milling cutter, and no   cutting fluid was used in the milling process. The milling method was down milling with a machining allowance of 1 mm and a back cutting depth of 0.5 mm. Macroscopic photographs of the milling equipment and specimens before and after milling are shown in figure 4.

Microstructure and property characterization
The two-dimensional morphology and microstructure of the specimens were photographed with the VHX-5000 ultra-depth-of-field three-dimensional observation microscope system. Before observing the microstructure, the specimens were ground with 400#, 800#, 1200#, 2000# sandpaper and polished to mirror effect, and then the specimens were etched with diluted aqua regia (H2O:HCl:HNO3 = 6:3:1). The threedimensional morphology and roughness of the specimen were obtained by Contour Elite K three-dimensional optical microscope host. The microhardness of the specimens was measured with the HXD-1000TMC microhardness tester at a load of 500gf and a residence time of 15 s. Ten microhardness values were measured for each specimen. The residual stress on the specimen surface was obtained using the Proto LXRD X residual stress diffractometer.

Results and discussion
3.1. The effect of additive and subtractive hybrid manufacturing on the surface morphology of 18Ni300 maraging steel (1) The influence of additive process parameters   Figure 5 reveals the variation of 2D and 3D morphologies on the upper surface of SLM samples with laser power. According to figures 5(a)-(f), when the laser power is low (100 W, 120 W), the surface of SLM sample has many defects such as voids and gaps. This is due to the low laser power, the metal powder is not melted enough, the wettability of molten pool is poor and the size of molten pool is not enough, so the melt tracks cannot be perfectly overlapped, thus there are holes and gaps. When the laser power increased to 140 W, the surface quality had been significantly improved, as shown in figures 5(g)-(i). Due to the increase of laser power and the increase of energy absorbed by the powder, the size of molten pool became larger, so the lap state of melt tracks was improved and the surface defects were reduced. As can be seen in figures 5(j)-(o), when the laser power increased to 160 W, the powder can basically melt completely due to the high laser energy, and the melt tracks bonding quality was better, so the surface was flatter. The laser power continues to increase to 180 W, and the surface quality is better because the higher temperature gradient will improve the fluidity of the molten pool, as shown in figures 5(m)-(o). However, the metal powders of 18 W and 160 W are fully melted, and the lap state of the melt tracks is not significantly different, so the surface quality of 180 W is not much better than that of 160 W. When the laser power is high (160 W, 180 W), the surface can also become flatter to some extent under the recoil pressure caused by metal vapor [20].
In addition, spheroidization appeared on the surface at low laser power (100 W, 120 W), as shown in figures 5(a) and (d). The spheroidizing effect is due to the fact that the energy irradiated by the laser on the powder is not enough to melt the powder fully, and the molten pool does not have enough wettability to flow and forms a sphere under the action of surface tension [21]. The phenomenon of spheroidization gradually decreases with the increase of laser power. But at this time, there are droplets and metal powder carried by metal vapor, as shown in figures 5(j), (m), which is not conducive to the full melting of the powder, so there is always spheroidization in the SLM process. The splash in the SLM process will become the crack source in the tensile process [22,23], which is not good for the performance of the parts.
The top surface morphology of the specimen when the scanning speed increased from 300 mm s −1 to 750 mm s −1 is shown in figure 6. It can be seen from figures 6(a)-(c) that when the scanning speed is 300 mm s −1 , the laser energy is sufficient and the powder can be fully melted, so the lap state of the melt tracks is good and the surface quality is the best. Figures 6(d)-(f) shows the morphology when the scanning speed increases to 450 mm s −1 . The melt pool size becomes smaller and the lap state of the melt tracks becomes worse due to the deterioration of powder melting state. As shown in figures 6(g)-(l), when the scanning speed increased to 600 mm s −1 and 750 mm s −1 , large holes and humping appeared on the surface, the gaps between melt tracks further became larger, and the melt tracks became irregular and discontinuous. The formation of humping is closely related to the complex dynamics of the melt pool [24] and is the result of a combination of driving forces such as surface tension, recoil pressure and thermal gradient at high scanning velocity. The irregularity and discontinuity of melt track indicate that the melt track is unstable and can be explained by the Platform-Rayleigh instability [25] effect, the liquid column will decompose into several spherical droplets under surface tension when the length-to-diameter ratio of liquid column is large. The increase of scanning speed will increase the length-diameter ratio of the melt track and decrease the wettability of the melt pool [26], so it is easier to trigger the Platform-Rayleigh instability effect [27].
As can be seen from figures 6(h) and (k), the increase of scanning speed will increase the bulge of the edge of the melt tracks. This is because the increase of scanning speed reduces the energy absorbed by the metal powder, and the melting state of the powder becomes worse, so the melt pool will move to one side of the insufficient melted powder under the action of Marangoni to form a bulge. Another reason is that during SLM, it is inevitable to remelt the incomplete melting powder on the melt tracks of previous layer [28], but it cannot be completely remelted due to the limitation of laser spot, as shown in the dotted box and arrow in figure 6(j). Insufficient remelting will lead to poor melting state of the newly formed melt pool, so the new melt pool will shift and form a bulge under the action of Marangoni effect.
From the above analysis, it is clear that increasing the laser power and decreasing the scanning speed will both improve the metallurgical bonding of the powder within the range of additive process parameters chosen in this study, which will result in a flattened surface topography.
(2) The influence of subtractive process parameters The surface quality of SLM specimens is too poor to obtain roughness, which can be significantly improved after milling. Figure 7(a) shows the variation curve of roughness with feed rate when the cutting speed is 14 mm min −1 , and it can be seen that the surface roughness increases with the increase of feed rate. This is due to the increase in material removal rate per unit time due to the increase in feed rate, which in turn leads to an increase in cutting forces and tool vibration during milling. Therefore, the residual height and roughness of sample surface increase, and the surface quality deteriorates. Figure 8 shows the 2D and 3D topographic of the upper surface with varying feed rate, which also confirms this. As the feed rate increases, the spacing between the textures becomes larger, the maximum distance between the pit and the convex peak becomes larger, and the surface quality becomes worse. Figure 7(b) shows the variation curve of surface roughness with cutting speed when the feed rate is 63 mm min −1 , it can be seen that the surface roughness increases and then decreases when the cutting speed increases. This is because when the cutting speed is initially increased, some of the metal material will bond to the tool, forming chip tumor under the action of temperature and pressure. While the chip tumor can protect the tool, it can damage the machined surface and increase roughness. As the cutting speed continues to increase, the chip tumor grows to a certain height and breaks due to the impact and vibration of the cut, thus improving the surface quality. Increasing the cutting speed also increases the cutting temperature. The thermal softening of the material will make the cutting process easier and thus improve the quality of the surface to some extent. Figure 9 shows the 2D and 3D surface topography of the upper surface as a function of cutting speed, which also shows the above variation. As the cutting speed increases, the texture spacing decreases, and the maximum distance between the convex peak and the pit first increases and then decreases. When the cutting speed was 10 m min −1 and the feed rate was 36 mm min −1 , the roughness was the lowest (2.1 μm), and the maximum distance between the convex peak and the pit was the minimum (52 μm). In addition, it can be seen from figures 8 and 9 that the SLM specimens exposed the hole defects of SLM after milling.
From the above analysis, it can be seen that the roughness of SLM specimens decreases after milling subtractive, down to a minimum of about 2.1 μm, and the surface morphology changed from the irregular melt tracks of laser additive to flat machining texture, which can meet the requirements of some surface of the mold with conformal cooling channels. Increasing the feed rate will result in a higher material removal rate per unit time and a worse surface quality over the chosen range of subtractive process parameters. While increasing the cutting speed will cause the surface quality to deteriorate first and then improve under the action of chip tumor and the cutting heat.    Figures 10 and 11 show the low magnification metallographs of the SLM specimens at different laser power and scanning speed, respectively, which show the melt tracks and melt pool boundaries as well as the direction of additive forming. The overlapping of circular tracks can be seen in the metallographic picture of the longitudinal section, which is a typical SLM melting-solidification mode, and the circular tracks come from the Gaussian distribution of the laser energy [29]. As can be seen from figure 10, the melt tracks on the top surface tended to be continuous when the laser power increased, which indicated that the increase of laser power makes the lap between melt tracks better and better, and the metallurgical bonding was improved, while no obvious pattern was observed in the tracks change of the longitudinal section. Figure 11 shows that when the scanning speed increases, the melt tracks on the top surface become increasingly discontinuous and holes appear. The melt tracks on the side are getting more and more irregular and some of them have been twisted, as shown in the dashed box in figures 11(d-2). The reason is that the laser energy is insufficient and some of the powders melts incompletely, which causes the height of the melt tracks to vary. Figure 12 shows the high magnification metallographic image of the longitudinal section of SLM specimens at a magnification of 2000X. The microstructure is dominated by equiaxed and columnar structure, but some coarse grains also exist due to the complex heat conduction and thermal gradient. The maximum cooling rate during SLM process is located at the boundary of the molten pool in the direction indicated by the black arrow  [30] in figure 12. During the solidification process, columnar structure grows along the direction of maximum cooling rate and extend outward perpendicular to the melt pool boundary to form epitaxial structure [31]. There is no obvious pattern in the effect of changing process parameters on the microstructure.
As can be seen from the above, increasing the laser power or decreasing the scanning speed will improve the melt overlap in the range of process parameters chosen in this study, but the effect on the microstructure is less pronounced.
(2) The influence of subtractive process parameters Figures 13 and 14 show the microstructure of the longitudinal section of the SLM specimens after milling, with a magnification of 2000X. Figure 13 shows the microstructure when the feed rate changes, and it can be seen that the depth of the plastic deformation layer increases with the increase of feed rate under the action of cutting tool. The reason is that when the feed rate increases, more material is removed per unit time, the cutting force   increases, so the degree of plastic deformation increases. In addition, an increase in feed rate will cause the cutting temperature to rise and thermal softening of the specimen surface to occur, reducing the cutting difficulty and therefore increasing the depth of plastic deformation. However, due to the increase of the material removed per unit time, the shear and tensile effect of the cutting tool on the unit volume of material will become weaker, which is not conducive to the distortion and deformation of the grain, so the larger the feed, the weaker the grain refining effect. Figure 14 shows the microstructure when the cutting speed varies, and it can be seen that the increase of cutting speed also increases the depth of the plastic deformation layer. The reason for this is that increasing the cutting speed increases the cutting temperature, thermal softening of the material occurs, and cutting becomes easier, thus increasing the depth of the plastic deformation layer. In addition, 18Ni300 maraging steel is a twophase (α + γ) alloy, whose α and γ phases are body-centered cubic structure and face-centered cubic structure, respectively. The number of slip systems for the α phase is 12, which is the same as the number of slip systems for the γ phase. Because the γ phase has a greater number of slip directions than the α phase, and the atomic density on the slip surface is higher than that of the α phase, plastic deformation is more likely to occur. When the cutting speed increases, the temperature rises and phase transition occurs on the surface, and part of the α phase is transformed into the γ phase, which exacerbates the plastic deformation of the cutting layer to a certain extent. Therefore, the increase of cutting speed will cause thermal softening and phase transition on the surface of specimens, which in turn will intensify the plastic deformation of material and the twisting and stretching of grain to achieve grain refinement.
From the above, it can be seen that milling can refine the surface grain of SLM specimens to some extent.
3.3. The effect of additive and subtractive hybrid manufacturing on the microhardness of 18Ni300 maraging steel (1) The influence of additive process parameters Figure 15 shows the microhardness of the top surface of SLM-formed maraging steel specimens. The two long horizontal lines in figures 15(a) and (b) show the microhardness of maraging steel castings. Compared with the maraging steel castings, the microhardness of the SLM specimens has increased significantly, by about 60 HV. This is because SLM specimens have finer grain, and even if there is residual austenite in the forming process, the effect of fine grain strengthening is greater than the degree of tissue softening. As shown in figure 15, the increase of laser power will increase the microhardness of SLM sample, since as the increase of laser power, the powder can be fully melted and the melt tracks overlap can be improved, thus the surface quality will be improved. Increasing the scanning speed decreases the microhardness due to the fact that too fast scanning speed does not allow the powder to melt sufficiently and the lap of melt track becomes poor, resulting in poor surface quality. The microhardness value was 364 HV at a laser power of 180 W and a scanning speed of 300 mm s −1 .
As can also be seen in figure 15, the lower the laser energy (such as low laser power and high scanning speed), the greater the error of microhardness, which indicates that the lower the laser energy, the more unstable the SLM forming quality. When the laser power increases or the scanning speed decreases, the powder can be sufficiently melted to form a good metallurgical bond, and the current scanning layer has the effect of post heat treatment on the previous layer, so the microhardness gradually stabilizes.
(2) The influence of subtractive process parameters The microhardness of the top surface on the milled specimens is shown in figure 16. The two long horizontal lines in figures 16(a) and (b) show the microhardness of SLM specimens before milling, and it can be seen that there is a significant overall increase in the microhardness of the SLM specimens after milling, which indicates that milling causes work hardening of the specimens [32]. Figure 16(a) shows that with the increase of feed rate, the plastic deformation effect of the sample increases, but the grain refinement effect decreases and the surface work hardening decreases, so the microhardness decreases gradually and fluctuates greatly. As can be seen from figure 16(b), the greater the cutting speed, the greater the microhardness and the smaller the fluctuation. This is because, when the cutting speed increases, the plastic deformation and strain rate of the specimen surface also increases, and the grain refinement effect is enhanced. The maximum microhardness value of 464.9 HV was measured at a cutting speed of 10 m min −1 and a feed rate of 36 mm min −1 .
As can be seen from the above analysis, the surface of SLM sample shows a milling-induced work hardening, with a significant increase in its microhardness up to about 110 HV.
3.4. The effect of additive and subtractive hybrid manufacturing on the surface residual stress of 18Ni300 maraging steel (1) The influence of additive process parameters The surface residual stresses of SLM-formed 18Ni300 maraging steel are shown in figure 17. The test point is located in the middle of the upper surface. From figures 17(a) and (b), it can be seen that the residual stress does not change linearly with the change of laser power or scanning speed, which indicates that the physical state of the melt pool is extremely complex during SLM. In addition, defects such as holes and spheroidizing effect will appear at certain process parameters, which will undoubtedly increase the disorder in the residual stress distribution. Compared with figures 17(a) and (b), the influence of scanning speed on residual stress is greater than that of laser power.
(2) The influence of subtractive process parameters Figure 18 shows the effect of milling process parameters on surface residual stress. It can be seen that the residual stresses on the surface of milling specimens are all tensile stresses, and the sensitivity of residual stress to the cutting speed is higher than that to the feed rate. Under the condition of dry milling, the variation of surface residual stresses depends on mechanical effect and thermal effect, with mechanical effect generally forming compressive stress and thermal effect generally forming tensile stress. In this study, the residual stresses of the milling samples are all tensile stresses, indicating that the thermal effect is always greater than the mechanical effect during milling. Figure 18(a) shows that the residual tensile stress decreases when the feed rate is increased. With the increase of feed rate, the cutting force increases due to the increase of material removed per unit time, the enhancement effect of mechanical effect is greater than that of thermal effect, resulting in the increase of compressive residual stress is greater than that of tensile residual stress. Therefore, the surface residual tensile stress decreases under the combined effect of the two. Figure 18(b) shows that the residual tensile stress increases with the increase of cutting speed. The reason is that the increase of cutting speed leads to the increase of cutting temperature, the enhancement effect of thermal effect is greater than that of mechanical effect, resulting in the increase of residual tensile stress is greater than that of residual compressive stress, so the residual tensile stress on the surface increases.
Therefore, after milling SLM samples, the residual compressive stress beneficial to the surface was not produced, but the surface residual stress became regular. Within the range of subtractive process parameters chosen in this study, increasing the feed rate or decreasing the cutting speed will reduce the residual tensile stress under the action of mechanical and thermal effects.

Conclusions
The effect of laser power and scanning speed on the surface integrity of 18Ni300 maraging steel by selective laser melting has been studied in this paper. Subsequently, the SLM specimens were milled to study the impact of cutting speed and feed rate on the surface integrity of milling machining. The main conclusions are as follows: (1)During the SLM process, the surface morphology is improved and the metallurgical bonding is progressively better with increasing laser power, but there is a spheroidizing effect regardless of the laser power variation. When the scan speed is increased, the surface quality deteriorates and the gaps between the melt tracks become larger, while large holes and bumps appear and the melt tracks become irregular and discontinuous.
(2)The measured micro-hardness increases with increasing laser power or decreasing scanning speed and the error decreases. The effect of the scan speed on the residual stress is larger than the effect of the laser power.