Laser shock forging—a novel in situ method designed towards controlling residual stresses in laser metal deposition

This paper presents a novel hybrid in situ additive manufacturing (AM) method–laser shock forging (LSF), which combines laser shock peening (LSP) with laser melting deposition (LMD). Based on the classical bar-frame model and inherent strain theory, the mechanisms of the effects of pretreatment and posttreatment on AM process have been elaborated for the first time. Towards controlling tensile residual stress (TRS) in the as-built (AB) state of AM parts which has a detrimental effect on their fatigue life, we then developed LSF to introduce tensile inherent strains in LMD built parts in an in situ manner, which will convert TRS to compressive residual stress (CRS). The laser beam used for shock peening can be adjusted to move synchronously with the laser beam used for metal deposition and keep a certain distance, ensuring the laser shock peening to act on the region where the material temperature cools down to the forging temperature range. Then, experimental works have been conducted on 316L stainless steel; residual stress distributions of the AB, LSP, and LSF treated specimens were compared; results show that LSF increases both the magnitude and depth of CRS compared with conventional LSP treatment, thus providing a promising application in enhancing fatigue life in AM process.


Introduction
In recent years, additive manufacturing (AM) [1] has attracted much attention due to its capabilities of fabricating complex geometry parts without dies, especially for those with internal features. It is realized by slicing the CAD model of a part into many thin layers, and then "print" each layer successively in a bottom-up manner, that is why it is also called "3D printing" [2]. The powder bed fusion (PBF) process is one of the most commonly used AM process for manufacturing complex metal parts; selective laser melting / Published online: 20 January 2023 The International Journal of Advanced Manufacturing Technology (2023) 125: [2289][2290][2291][2292][2293][2294][2295][2296][2297][2298][2299][2300][2301][2302][2303][2304] (SLM) [3][4][5] and electron beam melting (EBM) [6] are classified to this category. On the other hand, directed energy deposition (DED) is more suitable for making repairs or adding features to an existing part, although it can also be used to build intact components. Laser melting deposition (LMD) [7,8] and laser engineered net shaping (LENS) [9] are classified to this category. Compared with PBF, the size of the powder used by DED is larger and needs higher laser energy density; it has a faster building velocity but brings about lower surface quality which may need extra processing.
At the macroscopic scale, all these different metal AM processes that concerning about a moving high-energy heating source basically have similar physical processes which involves melting and solidification of metal [10]. Therefore, the large temperature gradient and high cooling rate occur due to intensive heat input, and large local energy density will lead to accumulation of detrimental tensile residual stress (TRS) during the process [11][12][13]. This TRS distribution accounts for the geometrical inaccuracy, cracking, warpage, layer delamination, and mechanical strength reduction of the AM built part, which leads to reduction of fatigue life and has to be eliminated as much as possible.
Many researches have been done to control and reduce residual stresses in AM [14][15][16][17][18][19][20]. Preheating the substrate is a commonly used in situ method [21]. It is found that adjusting laser scanning strategies have significant effect on the final residual stress distribution of AM built parts [22,23]. For post treatment, annealing is also a widely used method and can reduce 70% of residual stress in some cases. However, the methods mentioned above cannot fully convert TRS to compressive residual stress (CRS) which has been proved to capable of improving fatigue life of the built part. Moreover, in some cases where in situ preheating or optimized scanning strategies cannot be successfully implemented, it will meet difficulty in building parts by AM.
In recent years, laser shock peening (LSP) as a novel surface treatment method has attracted much attention in its capability of enhancing fatigue life of AM built parts [24][25][26][27]. It can introduce CRS in the near surface layer of the material and deal with intricate parts with efficiency and accuracy. Compared with shot peening and ultrasonic shot peening [28][29][30], the CRS introduced by LSP has greater magnitude and deeper depth, thus has become a powerful post treatment to improve the fatigue life of the AM built parts.
However, it is still a post treatment and cannot introduce the bulk accumulation of high TRS during the AM building process. Kalentics et al. [31][32][33][34][35] proposed a novel hybrid AM process which they described as "3D LSP" to successfully allow the 3D control of residual stress in SLM built parts. The method realized this by applying the LSP treatment every few SLM layers, which can obtain accumulated CRS in any critical zone in the bulk of the part. However, this method still needs careful realignment when the built parts come to the rebuilding phase after LSP treatment, which is not an in situ method and sacrifice manufacturing efficiency.
The inherent strain (also called eigenstrain) theory was first presented by Ueda et al. [36,37] in late 1970s and early 1980s to predict the residual distortion and stress in metal welding process. In recent years, inherent strain theory has also been applied to analyze metal AM processes as they share similar characteristics with metal welding processes in which the materials will experience a cycle of heating and cooling and left with inherent strains [38,39]. In the AM built parts and heat-affected zone of welding pass, when the metal cools down to ambient temperature, there will be in-plane compressive inherent strains (usually inherent strain equals to plastic strain in this case if phase transformation is not considered) left, which are formed due to the constraint of the surrounding area during contraction.
To counterbalance these compressive inherent strains, tensile residual elastic strains are formed, leading to TRS distribution in the as-built (AB) parts. As compressive inherent strains and its corresponding TRS distribution in the manufactured parts are not favorable for fatigue life, all the pretreatment or posttreatment are aiming at optimizing compressive inherent strain distribution, or converting them into positive. Obviously, preheating or optimizing scanning strategies belong to the former while LSP or 3D LSP treatment belong to the latter.
In this paper, based on the classical bar-frame model and inherent strain theory, the mechanisms of the effects of pretreatment and posttreatment on the manufacturing process concerning a cycle of heating and cooling have been elaborated for the first time. Towards introducing tensile in-plane inherent strains into the parts during the AM process, we then developed a novel hybrid method called "laser shock forging" (LSF) to introduce tensile inherent strains in LMD built parts in an in situ manner; the method has been patented by the Laboratory of Sino-US joint Laser Shot Peening at Guangdong University of Technology (GDUT) [40]. By careful calibration of robotic arms, the laser beam used for laser shock peening can be adjusted to move synchronously with the laser beam used for metal deposition and keep a certain distance. Then experimental works have been conducted on an austenitic 316L stainless steel; residual stress distributions of the AB-, LSP-, and LSFtreated specimens were compared; results have shown that LSF increases both the magnitude and depth of CRS compared with conventional LSP treatment, thus providing a promising application in enhancing fatigue life in AM process.

Inherent strain theory and LSP treatment
The conventional LSP introduces tensile plastic strain into the surface layer of the metal which induces compressive residual stress field due to elastic-plastic stress balance [41], as shown in Fig. 1. The introduced plastic strain in LSP is the source of inherent strain, which is any kind of permanent strain due to inelastic process such as plastic deformation, crystallographic transformation, and thermal expansion mismatch between different parts of assembly which causes the incompatibility in the material, resulting in the existence of residual stress distribution. The inherent strain can be denoted as [38]: where * represents inherent strain, p is plastic strain, and th is thermal strain; it is noted that in this paper, strain due to phase transformation has not been considered.
Warm laser shock peening (WLSP) [42] or laser shock peening without coatings (LSPwC) [43] are both designed to introduce larger tensile plastic strain (inherent strain) in the surface layer of the material, as higher temperature of the peening material or peening without ablative coatings will both cause the material easy to yield, thus leading to larger tensile inherent strain after LSP treatment.

Residual stress formation mechanism in LMD
The LMD process is a complicated manufacturing process in which the material experiences a cycle of heating and cooling, like the process of welding. As shown in Fig. 2, during the deposition, the material in front of the heat source (along the scanning direction) experiences thermal expansion due to the high energy input of laser beam, and the expansion is constrained by the surrounding relatively less hot material, resulting in compressive stress in this region. After the heat source passes, the material cools down rapidly and begins to contract; then, tensile residual stress field will be generated due to the constraint of the surrounding cold material. The similar kind of expansion and contraction cycle is also typical in the welding process, which is not favorable in the manufactured parts.
(1) * = p + th It is noted that tensile residual stress field is due to the compressive plastic strain formed in the heating and cooling cycles, which is the source of incompatibility in the material; in the inherent strain theory, if the material temperature cools down to ambient temperature, the inherent strain in the material is then denoted as [39]: Therefore, to reduce or eliminate the effect of tensile residual stress generated during the cycle of heating and cooling process, the key lies in reducing the compressive plastic strain formed in process.

LSF designed for reducing tensile residual stress
For the research concerning using LSP to treat SLM or LMD built parts, the researchers aim to introduce tensile plastic strain in the built parts to counterbalance the compressive plastic strain generated by the cycle of heating and cooling [44]. In order to make the LSP treatment more significant, Kalentics et al. [31][32][33][34][35] proposed the concept of 3D LSP, in which laser shock peening is conducted after n layers of SLM parts are built, the experimental results have shown that after 3D LSP treatment a significant increase in the magnitude, and depth of CRS has been observed when compared with as-built (AB) SLM parts and traditionally LSP parts.
As shown in Fig. 3, the primary difference between the proposed LSF approach in this paper and the 3D LSP approach proposed by Kalentics et al. is that the laser beam used for shock peening moves synchronously with the laser beam used for metal deposition, and by careful calibration of robotic arms, they can be kept at a certain distance. The laser beam used for shock peening can be adjusted to the region where the material temperature cools down to the forging temperature range (typically 800-1250 °C), so that the laser shock peening can acts like a hammer forging on the metal just deposited by the scanning laser beam, which is why we present this novel process as laser shock forging.

Inherent mechanism for reduction of tensile residual stress
In order to explain the inherent mechanism for the reduction of tensile residual stress in LMD process, the classical bar-frame model which has been used to illustrate the residual stress formation mechanism in the research of welding process is adopted [45]. As shown in Fig. 4, three rods of equal length L are connected the rigid wall and the rigid body, the rigid wall is fixed while the rigid body can only move in one dimension. Assume that the materials of the rods are the same and they can only expand or contract in the direction of length; for simplification purpose, the rods are considered as perfectly elastic-plastic model, which means no strain hardening after yielding. Von Mises yield function and the associated flow rule are adopted; during the loading and unloading process, the Bauschinger effect is not considered. The rod in the middle is uniformly heated to a temperature T max ( T max < T m ,T m is the melting point) and then cools down to ambient temperature. There are maybe many possible stress-strain evolution cycles according to the magnitude of T max ; we only take the most typical one [46] in which 2T Y < T max < T m , where T Y represents the temperature at which the uniaxial stress of the heated rod reaches the yield point and the stress state of the material turns from elastic to plastic.
As shown in Fig. 5, T represents current temperature and T 0 represents ambient temperature; the stress and strain evolve at different stages of heating and cooling; they can be characterized as follows [45]: As the temperature of the rod increases, the thermal strain can be calculated as: where represents the thermal expansion coefficient.  The rod has the tendency to expand in the length direction; however, its two ends are constrained by the rigid wall and rigid body, so uniaxial compressive strain and stress will be induced in the heated rod to remain stress equilibrium; the stress and strain will increase linearly before yielding as it is in elastic state.
: As the material begins to yield, it enters into plastic state, because it is an perfectly elastic-plastic material, the yield stress will not increase any more, and the material undergoes neutral loading while the total plastic strain will accumulate due to the presence of the plastic strain increment; this state is described in the following formulations: where von is the von Mises stress and ′ is the deviatoric stress tensor.
The yield function is denoted as: where Y is the yield stress of the material. When the yielding occurs, it follows that: By the associated flow rule, the strain increment can be calculated as: where d is a positive multiplier.
Then, the total accumulated plastic strain can be calculated as: As the temperature increases, the plastic strain keeps accumulating due to neutral loading, and attains at its maximum Δ p1 = (T max − T Y ) when T = T max . It has to be noted that in the heating phase, the accumulated plastic strain is compressive with a minus sign; the total accumulated plastic strain in the heating stage can be calculated as: As the temperature reaches T max , it begins to cool down, the corresponding thermal strain gradually decreases, and the rod enters from plastic state into elastic state to start elastic unloading; this stage proceeds until = 0 ; at that point, elastic strain vanishes, the compressive accumulated plastic strain and the thermal strain due to thermal expansion reach a balance to remain deformation compatibility.
As the temperature continues to decrease, the thermal strain also decreases, to remain deformation compatibility, the vanishing part of the thermal strain will then be compensated by the tensile elastic strain, thus the rod experiences a reverse elastic loading until yielding again, at that point After reverse yielding, the material will experience neutral loading again, which means the yield stress will remain at = Y , and the accumulated plastic strain changes according to the plastic strain increment. As in the cooling phase, the plastic strain increment is tensile with a plus sign; it is noted that the accumulated plastic strain will decrease in magnitude due to reverse loading; the total accumulated plastic strain in this stage can be calculated as: Finally, when the rod cools down to ambient temperature, the plastic strain (inherent strain) left in the rod can be calculated as: It is shown that the inherent strain formed due to heating and cooling cycle is compressive; to remain deformation compatibility, the rod will be subjected to the same Fig. 4 The classical bar-frame model magnitude of tensile elastic strain, by Hooke's Law; this leads to tensile residual stress.
Taking a look at Eq. (11), the inherent strain consists of two components Δ p1 and Δ p2 which are formed in the heating and cooling phase, respectively. Δ p1 is compressive while Δ p2 is tensile; as the magnitude of Δ p1 is larger than that of Δ p2 , the final inherent strain is compressive. Therefore, the difference of their absolute value determines the final magnitude of residual stress. To reduce the residual stress due to the heating and cooling cycle, there are two routes designed for eliminating their difference: decrease the magnitude of Δ p1 or increase the magnitude of Δ p2 .
For the first route, preheating is a favorable and commonly used method in welding and additive manufacturing [7,21,47]. It is explained that by preheating the material from ambient temperature to a higher temperature, the temperature difference between T Y and T max will be narrowed, thus reducing the magnitude of Δ p1 , while in the cooling phase, the material will cool down to ambient temperature, and the magnitude of Δ p2 remains unchanged; therefore, inherent strain and its induced residual stress will be reduced, as shown on the left side in Fig. 6.
For the second route, a hybrid deposition and microrolling (HDMR) process proposed by Zhang et al. [48] has been used in AM process to reduce AM residual stress, the inherent reasons of which can be explained by Fig. 7. If in the elastic loading stage the rod is imposed by an external excitation which causes the material to yield before cooling down to temperature T max − 2T Y (assume it is T ′ Y in this case), and then cools down to ambient temperature T 0 , the magnitude of Δ p2 will increase Δ � p due to this excitation, while the magnitude of Δ p1 remains unchanged; therefore, inherent strain and its induced residual stress will be reduced.
Actually, LSP treatment [24] or 3D LSP treatment [32] of the SLM built parts also follows the second route; they introduce tensile plastic strain into the surface layer of the SLM built parts to alleviate the compressive plastic strain induced by heating and cooling cycle, as shown on the right side in Fig. 6; this kind of treatment also increases the magnitude of Δ p2 , but it has to be noted that the additional tensile plastic strain is introduced at ambient temperature and not in an in situ manner.
Actually, the process of welding or AM is far more complicated than the proposed bar-frame model; for instance, phase transformation [5] and work hardening [49] should both be considered in welding and AM. The case is more complicated for AM as the remelting of the previous layers and the surrounding material when scanning the next layer or pass should be considered [20]; moreover, the constraint condition for the whole model is altering all the time due to its layer-by-layer building characteristic [15][16][17]. However, the bar-frame model can shed some light on the routes designed for reducing residual stress induced in these thermal-related processes.
Based on the above analysis, we have developed laser shock forging (LSF) method to come up with a way of controlling residual stress in LMD built parts. In this case, the laser beam used for shock peening acts as the "external excitation" which increases the magnitude of Δ p2 in an in situ manner. By careful calibration of robotic arms, the laser beam used for shock peening can be adjusted to move synchronously with the laser beam used for metal deposition and keep a certain distance. The advantage of this method over LSP or 3D LSP is that by adjusting distance between the two beams of laser, the laser shock peening can act on the region where the material temperature cools down to the forging temperature range (usually 800-1250 °C), like a hammer forging on the metal just deposited by the scanning laser beam. In the forging temperature range, the material is easy to yield and larger tensile plastic strain can be induced. Besides, it is an in situ and noncontact type of strengthening method; by comparison, LSP or 3D LSP is not in situ and needs careful realignment of the built parts, while in HDMR method, the micro-roller needs to contact the AM pass to strengthen it.

Material and laser metal deposition
The material used in the experiment is the widely used 316L austenitic stainless steel, with an ultimate tensile strength (UTS) of 625 MPa. The powder used for LMD was provided by Guangdong Lei Ben Co. Ltd., China. Table 1 shows the chemical composition of the material, and Fig. 8 shows its macro and micro morphology. Laser metal deposition was performed with a RC-LDM-2000-R (manufactured by Nanjing Zhongke Yihuan Laser Technology Co. Ltd., China) equipped with a fiber laser operated in continuous mode at a wave length 1064 nm and a spot size of 75 ~ 100 μm, as shown in Fig. 9. The specimen geometry is a 30 × 20 × 10 mm 3 cuboid, and the base plate geometry is 140 × 140 × 8 mm 3 sheet with its material 1Cr13 mild steel. The chosen LMD processing parameters are shown in Table 2. A bidirectional scanning strategy parallel to the part edge was used to avoid creating large residual stress when changing the scanning direction between layers. Processing was performed under Ar atmosphere to avoid oxidation of the powder throughout the process.

Laser shock forging
Laser shock forging (LSF) experiments were done using the laser shock peening system "PROCUDO200" manufactured by LSPT Co. Ltd., USA, as shown in Fig. 10. The laser source was a Nd/YLF with a pulse duration adjustable between 8 and 18 ns and in this case is 18 ns. The beam spatial energy distribution is "top-hat," and the pulse shape is near-Gaussian. The shape and size of laser spot is also adjustable and in this case is round with a diameter of 3 mm. The ratio of spot size and energy per pulse was adjusted carefully to keep a constant power density of 8 GW/cm 2 . The pressure P created at the surface of the part was estimated to 4.95 GPa using the following empirical equation: where I 0 is the power density of the laser beam. Laser repetition frequency was 10 Hz, and the overlapping rate was 50% without a protective ablative layer. Detailed LSF processing parameters are shown in Table 3.
The key of the LSF process is to adjust the distance between the two laser beams to ensure the LSF acts on the deposited material whose temperature cools down to the forging temperature range. Due to the rapid solidification of the 316L stainless steel, the distance needs careful calibration. For the processing parameters used in this experiment, by many trial experiments, we finally determine that at a scanning speed of 5 mm/s of the laser beam for LMD, the appropriate distance between the two laser beams is 0.40 mm.

Residual stress measurement using the X-ray diffraction method
Residual stress measurement was conducted using the X-ray diffraction method with a XL-640X X-ray stress tester manufactured by Shanghai Puchuang Technology Co. Ltd., China, as shown in Fig. 11. In-plane residual stress on the surface layer of the specimens was measured first; then, the specimens are eroded layer by layer using an electro-polishing machine to measure the residual stress in the depth direction. A constant depth increment of 0.1 mm for electro-polishing erosion was applied, and a maximum depth of 1 mm was measured for all the AB, LSP, and LSF specimens. The basic principle of residual stress measurement using X-ray diffraction is that when there is residual stress in the specimen, the crystal plane spacing will change and the diffraction peak will move when Bragg diffraction occurs, and the magnitude of the moving distance is related to the magnitude of the stress. Assume d 0 is the lattice spacing of an unstressed specimen and 0 is the corresponding diffraction  The determined strain is an average elastic lattice strain over a sampled gauge volume defined by slits and formed by the intersection of the incident and diffracted beams.
Then by Hooke's Law for the linear elastic properties of the material, the orthogonal in-plane stresses (assumed to be the principal stresses) can be calculated by the following equation [51]: where E hkl and hkl are the elastic modulus and Poisson's ratio of a specific crystallographic plane, respectively.

Results and discussion
To compare residual stress profiles of specimens due to different treatments, nine times of experiments were divided into three groups: Numbers 1 ~ 3 are the specimens built by LMD; numbers 4-6 are firstly built by LMD and then treated by LSP; and numbers 7-9 are the specimens built by LMD combined with LSF. To eliminate the effect of variable processing parameters, the laser parameters in LMD and LSP remain the same. After experiment, the specimens were cut from the base plate by electric discharge machining (EDM) and polished for residual stress measurement, and each measurement point of the specimen is measured three times to obtain average data; then, the data for the three specimens in each group will also be averaged to eliminate random errors.

As-built state
The as-built 316L LMD samples are shown in Fig. 12, and their averaged magnitude of residual stress is shown in

LSP-treated state
After LMD was done, the LMD samples were polished and covered with an ablative tape to be treated by LSP; Fig. 13 shows the specimens built by LMD before and after LSP. The measured residual stress value of LSP treated samples is also shown in Table 4, and the corresponding stress profiles are shown in Fig. 15. From Table 4, it can be concluded that after LSP treatment, the initial TRS near the surface layer of the LMD parts has converted to CRS. The maximum magnitude of CRS is 320 MPa which occurs at the depth of 0.4 mm, and CRS state remains at the depth of 0.65 mm.   Figure 14 shows the specimens built by LMD combined with LSF, the comparisons of the measured residual stress values of the AB, LSP-treated, and LSF-treated samples are shown in Table 4, and the corresponding stress profiles are shown in Fig. 15. It can be seen that LSF-treated samples have gained greater magnitude of CRS and deeper depth of CRS state, compared with the LSP-treated samples; an increase of 57% of the maximum magnitude of CRS has been presented. During the LMD process, the laser shock peening acts on the region where the temperature lies in the forging temperature range (usually 800-1250 °C for 316L austenitic stainless steel), which acts like a hammer forging on the metal just deposited by the scanning laser beam, which is why we present this novel process as laser shock forging.

Mechanisms of the obtained results
Based on the analysis in Section 2, the stress and strain variations for a specific material point for the three cases (AB, LSP, and LSF) are presented in Fig. 16. It has to be noted that the process of AM is far more complicated than the proposed bar-frame model, but the inherent strain formation mechanisms due to the cycle of heating and cooling are similar with each other. To simplify the problem, the following assumptions have to be made: The material point to be studied locates in the region where the scanning heat source has become stable; the edge effects in AM process are ignored. The material is perfectly elastoplastic with no work hardening, and its constitutive equation does not vary with temperature. In fact, if we take work hardening into consideration and assume that the constitutive equation varies with temperature, the problem will become nonlinear and much more complex, so these two assumptions are of vital importance to simplify the problem. Phase transformation is not considered. The material point to be studied experiences only one cycle of heating and cooling, which means the effects of heat source when scanning the next layer or pass are not considered. The constraint condition for the whole model which changes during AM is not considered.
It is shown that the heating stages for the three cases are the same, the generated plastic strain is compressive due to thermal expansion and constraint of surrounding materials, and by Eqs. (3)-(8), the total accumulated plastic strain in the heating stage can be calculated as: For the as-built material during cooling stage, it will yield at point C and undergo neutral loading, while tensile plastic strain will accumulate due to the presence of plastic strain increments. When the temperature cools down to ambient temperature, the only component of the inherent strain is the plastic strain, which is compressive, and can be calculated as: For the material treated by LSP, LSP treatment will introduce tensile plastic strain into the surface layer of the material at ambient temperature, which will counterbalance the already existence of compressive plastic strain to make the final inherent strain tensile. Therefore, it can be calculated as:  It has to be noted that, as the material has no work hardening, path D to D′ only represents the plastic strain of material because LSP treatment follows this path, and the stress of the material still remains at yield stress Y , as it is subjected to neutral loading; for comparison purpose, we draw this strain variation on the same figure, as shown in Fig. 16.
For the material treated by LSF, as shown in Fig. 16, LSF treatment will introduce tensile plastic strain into the surface layer of the material at the forging temperature; as the material is easier to yield at high temperature, larger tensile plastic strain will be generated compared with LSP treatment at ambient temperature; thus, the value of final tensile inherent strain will increase, which can be calculated as: It is noted that path Q′ to E does not introduce plastic strain as the material is still in elastic state, LSF treatment advances the material to the yielding state, and the accumulated plastic strain from path E to C represents the additional tensile plastic strain due to LSF treatment. Figure 17 shows the inherent strain accumulation at different heating and cooling stages for the three cases; it can be concluded that compared with LSP, LSF takes the advantage of higher forming temperature, in which larger tensile inherent strain can be introduced. However, LSF requires a high degree of synergy between the two laser beams; if they are too close, the laser shock wave will act on the molten pool, which cannot play the role of forging hammer; if they are too far, it acts in a similar way as warm laser shock peening (WLSP). It is also noted that the optimal distance between the two beams of laser varies due to change of material or laser scanning parameters; more experiments should be conducted to determine the relationship between the optimal distance and corresponding processing parameters, which is also our concentration on the future works.

Conclusions and future works
A novel in situ method to control the residual stress state of LMD built parts by virtue of laser shock peening, which we called laser shock forging, has been proposed. The mechanism of the conversion from TRS to CRS has been explained by the classical bar-frame model and inherent strain theory. The laser beam used for LSF can be seen as the "external excitation" which increases the magnitude of compressive strain in an in situ manner during the cooling phase. By careful calibration of robotic arms, the laser beam used for shock peening can be adjusted to move synchronously with the laser beam used for metal deposition and keep a certain distance.
The following conclusions can be drawn: 1. In the AB state, 316L stainless steel exhibits TRS up to the measured depth of 1 mm. 2. After LSP treatment, the initial TRS near the surface layer of the LMD parts can be converted to CRS. The maximum The maximum value of CRS is 320 MPa, and CRS state remains at the depth of 1 mm, which presents 57% increase of the maximum value of CRS and 53.8% increase of CRS depth. 4. LSF acts on the region where the material temperature cools down to the forging temperature range (usually 800-1250 °C). As the material is more prone to yield at high temperature, higher plastic deformation can be induced, and melting and solidification of successive layers will not lead to full stress relaxation; thus, CRS can accumulate in the whole bulk part during the building process.
Preliminary experimental results in our laboratory have shown that during LSF treatment, the microstructure of the material will also change compared with AB and LSP state; future work will concentrate on the microstructure evolution and assessment of fatigue life of LSF-treated specimens. The effects of the processing parameters including laser scanning speed, powder density, overlapping rate, and laser spot size on the optimal distance between the two laser beams will also be investigated.