Laser peening: A tool for additive manufacturing post-processing

Accepted Manuscript Title: Laser Peening: A Tool for Additive Manufacturing Post-processing Authors: Lloyd Hackel, Jon R. Rankin, Alexander Rubenchik, Wayne E. King, Manyalibo Matthews PII: DOI: Reference:

S2214-8604(18)30339-7 https://doi.org/10.1016/j.addma.2018.09.013 ADDMA 502

To appear in: Received date: Revised date: Accepted date:

17-5-2018 6-9-2018 9-9-2018

Please cite this article as: Hackel L, Rankin JR, Rubenchik A, King WE, Matthews M, Laser Peening: A Tool for Additive Manufacturing Post-processing, Additive Manufacturing (2018), https://doi.org/10.1016/j.addma.2018.09.013 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Laser Peening: A Tool for Additive Manufacturing Post-processing

SC RI PT

Lloyd Hackel1,* , Jon R. Rankin1, Alexander Rubenchik2, Wayne E. King3, and Manyalibo Matthews3

1

Curtiss Wright Surface Technologies - Metal Improvement Company, 2NIF and Photon Sciences Directorate, Lawrence Livermore National Laboratory, 3Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory,

1

N

U

Corresponding author: Tel: +1- 925-960-1090 Email address: [email protected] (Lloyd Hackel)

A

Abstract

M

Additive manufacturing (AM) is rapidly moving from research to commercial applications due to its ability

D

to produce geometric features difficult or impossible to generate by conventional machining. Fielded

TE

components need to endure fatigue loadings over long operational lifetimes. This work evaluates the ability of shot and laser peening to enhance the fatigue lifetime and strength of AM parts. As previously

EP

shown, peening processes induce beneficial microstructure and residual stress enhancement; this work takes a step to demonstrate the fatigue enhancement of peening including for the case of geometric stress

CC

risers as expected for fielded AM components. We present AM sample fatigue results with and without a stress riser using untreated baseline samples and shot and laser peening surface treatments. Laser

A

peening is clearly shown to provide superior fatigue life and strength. We also investigated the ability of analysis to select laser peening parameters and coverage that can shape and/or correctively reshape AM components to a high degree of precision. We demonstrated this potential by shaping and shape correction using our finite element based predictive modeling and highly controlled laser peening..

Keywords: Additive manufacturing, selective laser melting, laser powder bed fusion, fatigue, shot peening, laser peening, laser forming, laser form correction.

SC RI PT

1 Introduction 1.1 Residual stresses in additive manufacturing

Additively manufactured (AM) parts have the potential to make a dramatic impact on how advanced systems are designed and perform. A comprehensive review of metal AM is given in [1] .This technology enables the realization of parts designed to meet specialty requirements through the inclusion of

U

geometric features that are difficult or impossible to produce by conventional subtractive machining and

N

by part consolidation. While there are many potential positive aspects of the process, the large amount

A

of energy deposited during the build process can produce detrimental effects including voids, residual

M

stresses and resulting distortions. To improve part quality, surface polishing, shot peening, and thermal post processing have been used to relieve the residual stresses and even enhance microstructure. [2-6]

D

But, simply polishing and relieving residual stresses at the near surface may be insufficient to meet

TE

performance requirements and in fact, relieving local stresses during layering may actually lead to inducing detrimental deeper internal stresses and part distortion because the process of compressing a

EP

local layer necessarily results in development of compensating tensile stress elsewhere. The previously

CC

cited results by AlMangour et al. and Calignaro et al. in steel [3-5] evaluate the effect of shot peening on hardness, microstructure, and residual stress but did not evaluate the all-important real-world issue of

A

fatigue performance. The scatter in the AM Ti-6Al-4V results of Wycisk et al. . [6] does not constitute an actual endurance limit as they suggest since in each case at the lowest stress levels tested several samples lasted to runout but other samples failed at much lower cycle count. The lack of actual endurance limit for both the polished and shot peened samples means that they did not find useful fatigue benefit for any treatment in the AM Ti-6Al-4V material they tested. In contrast our herein work demonstrates reasonable

SN behavior in all tests and importantly shows beneficial fatigue results for shot peening and even better fatigue enhancement generated by laser peening. Parts placed into real systems are subject to design constraints and operating conditions, such as

SC RI PT

geometric requirements of fillet radii or off-normal occurrences such as foreign object damage (FOD). Each of these generates local stress risers during operational loading and consequently reduced fatigue performance. These areas of high loading stress are typically generated at and near surfaces and become points where cracks initiate and subsequently begin propagation, resulting in part failure. Surface treatments such as shot and more recently laser peening have been well accepted as a means to

U

compressively pre-stress a local area thereby reducing net stress during loading and thus improving the

N

cracking resistance and the fatigue lifetime. [7, 8] Reducing local stress can also be used as a means of

A

reducing the overall peak stress load thereby effectively increasing the fatigue strength of a part or

M

enabling reduction in part cross section thereby enabling lighter weight and more efficient design.

1.2 Laser peening and residual stresses

D

Peening is a process that plastically compresses material normal to a surface resulting in transverse

TE

(Poisson) expansion. When peening a thicker or otherwise constrained component it’s ability to resist the transverse straining results in a local buildup of compressive stress. For thinner components, the peening

EP

results in strain and shape change. Such is the case for all types of compressive surface treatments

CC

including shot, laser, and ultrasonic peening and processes such as deep cold rolling. Figure 1 illustrates how the process works for laser peening keeping in mind that the concept of plastic compression and

A

transverse expansion is common to all treatments. Laser peening (LP) is an important post processing method for metal parts. It is now extensively used to enhance the fatigue lifetime of jet engine fan and compressor blades and more recently in aircraft structures and nuclear spent fuel storage canisters. [9-11] It has been applied to improve surface properties in additively manufactured maraging steel. [12] The technology is also used to apply curvature

and stretch to thick sections of aircraft wing panels providing precise aerodynamic shaping. In the LP process, short intensive laser pulses create a plasma in a confined geometry (Figure 1) and thereby generate pressure pulses that create local plastic deformation. An ablative layer can be used in the process

SC RI PT

or as in this work, omitted resulting in only a very shallow (10 to 20 µm thick) layer of recast material that can be left on the surface or easily polished off. Use of a water tamper increases the generating pressure by an order of magnitude thus making the process more efficient. [13] Depending on material and geometry, existing residual compressive stresses, desired strains and microstructure, modifications to stress state and/or shape can be precisely generated in parts in a spot-by-spot manner. Laser peened

U

materials typically demonstrate higher cracking and corrosion resistance and are becoming widely used

EP

TE

D

M

A

N

in manufacturing.

CC

Figure 1 Laser peening plastically compresses material normal to a surface generating a transverse compressive stress field.

Laser peening is known for creating very small amounts of cold work, typically 3% to 5%, typically leaving

A

the phase, hardness, and yield strength of the treated material unchanged. Shot peening requires multiple impacts estimated for example at 13 impacts for 100% coverage, and due to the spherical nature of the impacts, the shot generates transverse as well as normal forces and plastic deformation. This working of the surface increases hardness and generates cold work. Laser peening with a square or rectangular beam

as used here, in contrast generates 100% coverage in only one impact per beam spot. The impact angle determined by the plasma pressure on the surface and not the laser light incident angle, is totally normal to the surface thus generating little hardening or cold work. Additionally, the large footprint of the laser,

SC RI PT

typically 3 mm to 10 mm on a side and the steady nature of the shock result in a very deep (multiple mm) plastic deformation of the material before the shock drops below the yielding limit. Prevey et al. provide a detailed study comparing cold work generated by shot, gravity and laser peening as inferred from the measured angular dispersion in x-ray diffraction. [14] This deep strong shock inserts dislocations equally deep into materials which helps resist crack initiation and growth thereby supporting enhancement in

U

fatigue strength and lifetime of treated components. The process creates deep compressive stresses

N

resisting the advance of cracks as well as providing superior resistance to stress corrosion cracking in

A

susceptible materials. [11] By selectively and compressively pre-stressing high tensile stress areas of

M

components, the laser peening enables higher levels of tensile fatigue loading before the fatigue limit is

A

CC

EP

TE

D

reached.

Figure 2 Residual stress measured in 316L stainless steel laser peened at 8 GW/cm2 irradiance, 21 ns pulse duration and 3 layers of peening with aluminum tape ablative layer. The solid triangles represent the measured residual stress determined by the Slitting Method. The linear dashed line is a fit to the slope of the bending stress profile resulting from the bending of the sample. The open triangles represent the summation of measured stress and resultant bending stress giving the equivalent stress that would have resulted from an infinitely thick sample. This represents the actual depth of plastic response to the peening. [15]

Higher energy laser peening such as used in this work, employing energy output in the range of 20 J/pulse, enables use of a relatively large spot size (1 cm2) and is to be contrasted to low energy laser peening of 1 J/pulse or less and corresponding smaller footprint. The large spot size enables deep penetration and

SC RI PT

consequently deep level of compressive stress. To characterize the deep level of compressive stress generated by laser peening in 316L we show in Figure 2 peening previously done using our technology and measurements made with a slitting technique. The slitting and analysis was done by DeWald. [15] In this example, a 30-mm thick sample of normally annealed 316L stainless steel was laser peened with parameters of 8 GW/cm2 irradiance, 21 ns pulse duration and three layers of peening using a thin layer of

U

adhesive-backed aluminum tape as the ablative layer. Dewald calculated the residual stress, as shown by

N

the dark diamonds, by systematically EDM cutting and measuring strain release vs. depth thereby

A

mapping strain and thus able to compute residual stress as a function of depth in the sample. His data

M

shows a stress zero crossing of approximately 4 mm depth in the sample. However due to the finite 30 mm sample thickness and the deep plastic impact of the laser peening, the sample responded by a slight

D

bending into a spherical shape thereby generating a bending stress profile represented by the dashed line

TE

in Figure 2.[16] Extrapolating the thickness of the block to near infinity where the stiffness becomes equally large enables calculation of the depth of plastic deformation, represented as stress by the open

EP

diamonds. Thus, the laser peening is seen to have penetrated to approximately 11 mm. This deepest

CC

penetration is the realm of the higher energy laser peening and specifically for its application in nuclear spent fuel storage where deep stress is critically important to extend beyond corrosion pitting depths and

A

thus prevent chlorine induced stress corrosion cracking. [17, 18] As can be seen, the depth of stress generated by peening of any component is thus the resultant of the plastic response generated by the process and the response of the component geometry. We use measured values of plastic deformation generated by the process to normalize our finite element analysis (FEA) model in specific component geometry for predicting the stress and strain generated by the laser peening.

1.3 Fatigue in additively manufactured 316L

M

    

SC RI PT

 

U

  

N



One of the most significant barriers to the adoption of additive manufacturing to produce critical parts is the threat that defects that may be introduced by the feed material or by voids induced in the processing put into the part during the additive manufacturing process will be a crack initiation source causing the part to fail prematurely due to metal fatigue. [19] There have been several studies on fatigue in 316L with the following, sometimes contradictory, observations: Fatigue life at lower stresses is lower than conventional material but similar at higher stress levels. [20] Fatigue behavior depends on build direction. [21] Fatigue behavior is comparable to conventional material. [22] While residual stress does not seem to affect fatigue, the additively manufactured microstructure does. [23] Laser shock peening is a promising process to improve fatigue behavior. [24, 25] The ductility of additively manufactured 316L reduces the effects of residual stress and surface roughness on fatigue. [26] Post heat treatment does not significantly benefit 316L fatigue. [27] Ti-6Al-4V was found to have a low sensitivity to blunt V-notches. [28] The notch fatigue strength of heat treated Ti6Al4V was found to depend on build direction. [29] Defects in the as-built structure adversely affect the fatigue behavior of 316L. [30] Support structures (thermal connection to base plate) were found to significantly affect fatigue strength of 316L. [31]

A



1.4 Laser peening to shape parts

D

Since laser peening plastically strains material at a deep and controlled depth into the treated surface and

TE

because the application is highly localized with individually controlled impacts, the process can be

EP

engineered on a spot-by-spot basis to form or correctively shape even thick parts with high precision. This technology has been applied since 2008 to put curvature and stretch into 30 m long aluminum panels. In

CC

30 mm thick 7050 Al alloy, 500 cm radius of curvature and stretch of several percent is generated to shape aluminum wing skins for wide-body aircraft. [9] In this processing, pre-stress is applied in the direction of

A

desired curvature or stretch and the peening irradiance adjusted to generate plastic yielding in the prestress direction while only reaching elastic levels in the non-stressed direction. The laser processing leaves a very acceptable surface finish of about 1 nm eliminating the need for sanding of the wing panels prior to painting and installation on to the aircraft.

1.5 Outline In the present paper, we consider two desired results of LP treatments for AM materials: improved fatigue performance and controlled straining. First, we demonstrate the beneficial effect of SP and then the even greater benefit of LP on material fatigue life and fatigue strengths. We report work in which we designed

SC RI PT

and fabricated, using laser powder bed fusion additive manufacturing, 316L 4-point bend fatigue test samples with a geometry-induced stress-intensity factor to simulate real-world components. Specifically, we treated and tested the fatigue life of un-peened, shot peened, and laser peened samples as well as a normally wrought sample. The results show the benefit of peening for improving the fatigue life of AM parts. In the second test set, we used LP to correct part distortion such as that potentially generated

U

during the layered manufacturing. Specifically, we demonstrated the ability to dramatically distort a test

A

N

strip of AM material and then straighten it to a measurement limit of flatness.

2.1 Sample and test geometry

M

2 Experimental

D

A fatigue coupon having four (4) loading points was selected for testing because this loading provides a

TE

central region of uniform tensile stress. The samples were 120 mm in length by 12 mm wide at the base and 21 mm tall. The gauge section tapered to 5 mm width with a 2.5 mm radius of curvature. As compared,

EP

for example, to a compact tension (CT) specimen where the high stress area is typically a hole with exterior circular geometry and interior concave surface, the chosen 4-point bend sample geometry provides easy-

CC

to-access convex geometry allowing uniform peening and stress loading and thereby enables testing of the material performance with and without peening rather than complications introduced by stress risers

A

due to geometry. Figure 3 shows AM test samples with the three surface treatments used in this work.

SC RI PT U N A

M

Figure 3. AM Fatigue test coupons of 4-point bend design enables uniform stress loading over the central gauge section. Dark shaded region of the laser peened sample is the small treated area processed with conditions 4 GW/cm2 irradiance-18 ns pulse duration and 2 layers of peening without tape ablative layer. A notch was applied at the center of the gauge section in some samples as shown to generate a controlled stress riser.

D

2.2 Fabrication of Additive Manufactured Test Samples

TE

Eight 316L stainless steel fatigue samples were built in a Concept Laser M2 metal additive manufacturing machine using the parameters shown in Table 1. All samples were fabricated with the build-direction

EP

normal to the loading-direction. The samples were arranged such that their long axis was at an angle of

CC

45° to the direction of the recoater. All AM samples were tested in the as-built condition. As discussed, AM material properties are sensitive to the build process. Thus, we started from the

A

measurements and comparison of the test results for our AM material and one wrought 316L sample manufactured to the same dimensions, notched, and tested at the lower 400 MPa load to provide a normalized level of performance.

Although limited to a test of one wrought sample, the large difference in fatigue lifes between sample types shown in Figure 5 suggests that our AM material has better fatigue strength performance than the conventional, wrought sample.

Setting

Laser power

180W

Laser speed

600 mm/s

Laser power for contours

160 W

Laser speed for contours

1600 mm/s

Beam size (D4)

54 µm

Trace width

150 µm

Layer thickness

30 µm

N

Chess

A

strategy:

M

Scan A1 A2 A33

U

Parameter

SC RI PT

Table 1. Process parameters for AM fabrication of fatigue test samples.

1

0.7 0.15 0.15

TE

D

2

EP

In real applications, geometric requirements of parts, such as fillet radii and passage holes, create stress

CC

risers that concentrate stress and become potential failure points or load limiters. To simulate this condition, after fabrication of the AM samples a notch was electro discharge machined at the centerline

A

of the curved top edge of selected samples to a width of 250 microns and to a depth of 350 microns. Stress analysis indicated that the notch generated a 3x local notch stress concentration factor; a stress riser considered similar to design geometry features typically required in an actual fielded part. It is expected

1

scan spacing between successive scan tracks expressed as a fraction of the trace width Island overlap factor 3 Island offset for elongation of scan tracks in the scanning direction 2

that the EDM process leaves a recast layer in the notch with a tensile stress intensity of about 80% of yield stress and a depth of about 100 m. The notch width was small enough that the peening shot did not reach the notch tip. So, the shot peening performance of samples may have been biased by this limitation.

SC RI PT

The laser peening however is actually generated by the plasma pressure which reaches into and peens the notched surfaces. It is a recognized attribute that when laser peening with linear p-polarization surfaces can be fully treated with laser light incidence up to 70o off of normal incidence. So, with a limited number of AM samples available, the EDM notch provides for a controlled localized stress riser that allows us to test the inherent AM material performance somewhat independent of random material inclusions,

U

voids and especially surface condition. This is especially important to AM material susceptible during the

N

build process to the formation of local defects. This was proven out in our initial testing of un-notched

A

samples where the fatigue cracking initiated randomly along the test gauge length. In comparison, all

M

failures in the notched samples occurred at this stress riser. Thus, the importance of the EDM notch is to help remove surface condition as a significant variable in the testing.

D

Finally, a wrought 316L sample was manufactured to the same dimensions, notched, and tested at the

TE

lower 400 MPa load to provide a normalized level of performance.

2.3 Laser and Shot Peening Processes

EP

The focus of this work was to investigate the benefit of surface peening treatments for AM samples. In

CC

shot peening, a well-established surface treatment, the local surface of the part is impacted with hard steel or ceramic balls. Shot of 1 mm diameter typically creates compressive stress of 0.2 to 0.4 mm deep

A

and the Hertzian nature of the spherical-impacts cold works the surface in the range of 40% to 50%. We selected un-notched samples for shot peening with 230R (0.5 mm diameter) shot of hardness RC 45 to 52. Other samples were EDM notched and then shot peened or laser peened. The shot peening was performed to an Almen intensity of 10A and with 200% coverage. [32]

In the laser peening process, laser light (irradiance of 0.2 to 10 GW/cm2) incident on a surface generates and heats a plasma that is tamped by a thin (typically 1 mm) water layer. The irradiance is adjusted to create a plasma pressure of 1 to 2 times the Hugoniot elastic limit (dynamic yield stress of the material

SC RI PT

being peened). The pressure wave thus generated propagates into the material plastically compressing it to depths ranging from 1 to 10 mm depending on selected process parameters. The compressed material transversely expands pushing against surrounding material. If the geometry of the part has sufficient stiffness, it mostly resists the expansion and thus builds a local compressive field. If the part has a lower local moment of inertia, the peened surface will undergo a greater amount of transverse strain leaving

U

less local compressive stress.

N

Selected samples in the test were laser peened with irradiance of 4 GW/cm2, 18 ns laser pulse duration

A

and 2 layers of coverage using no ablative layer. A square laser spot profile with linear spot size of 4.7 mm

M

and an overlap of 3% was used. Because of the much larger footprint used in the process in comparison with shot peening or low energy laser peening, the plastic compression penetrated deeper before the

D

natural diffraction of the shock wave reduced the pressure below the yield stress of the material. Peening

TE

was performed over the semicircular top area of the gauge section down 5 mm along the side. The process parameters used create a plastic deformation to depths of 3 to 4 mm shallower than the heavier peening

EP

shown in Figure 2 but much deeper than the 0.2 mm depth expected of the shot peening. This penetration

CC

was chosen to create a near through compressive stress in the local thickness of the part and thus was tailored to fully treat but not over-peen and deform the part. Deeper compressive stress plays a most

A

beneficial role in and generally provides greater fatigue strength and lifetime enhancement.

2.4 Fatigue Sample Testing Following surface treatment, test samples were fatigue cycled in 4-point bend loading to a lower R=0.1 limit until failure was detected at 10% amplitude compliance. Stress loading ranged from 400 MPa to 575 MPa using a 90 kN (20 kip) Instron test rig as shown in Figure 4.

SC RI PT U

A

2.5 Precision forming of AM Material

N

Figure 4. AM sample loaded on Instron fatigue rig tested with R=0.1 loading. 4-point bend testing generates uniform tensile stress loading on the lower portion of the sample over the separation distance of the upper rollers.

M

To implement a precision shaping application, we begin by generating a laser interferometer scan of the selected part or by obtaining other precision information of the starting shape. In parallel we obtain a

D

computer solid model of the part which represents the desired shape and then compute the difference

TE

between the actual shape and the desired shape. We then perform finite element stress/strain analysis

EP

using simulated mechanical loading to “virtually” laser-peen strain the part into the desired shape iterating the design to obtain a precision match between the desired and “virtual” shape. The resultant

CC

prediction of strains needed in specific areas of the part then defines where and how much peening needs to be applied to correctly shape the part. Measurements are then made of the strain response of the

A

specific material to selected laser peening parameters and these used as the code input. We iterate with our FEA model to compute the local peening parameters required to obtain a best fit to the required strain. This code, normalized with strain measurements, allows virtual trials on a spot by spot basis to search for a peening process (peening intensity and coverage) that best achieves the desired shape. The best virtual solution is then applied selectively to the defined areas of the part resulting in transverse

elongation and/or curvature of that surface. This approach was developed and now successfully used as the first step in our approach to evaluating and implementing forming and/or fatigue benefit to essentially all new laser peening applications.

SC RI PT

If a plate is uniformly peened over a surface, the plate will attain primarily a spherical shape. If that surface is pre-stressed in tension in one of the axial directions, the peening can be adjusted to plastically strain in that direction only allowing generation of a cylindrical shape. Other techniques that preferentially apply local tensile pre-stressing and appropriately adjusted peening are used to generate torus, dihedral, and other complex shapes of interest. Warped or misshapen parts can be locally corrected by the appropriate

U

application of the process. The technology is limited to stretching material as peening does not generate

N

negative strain; so forming needs to be planned with that in mind.

A

As a demonstration of applicability for AM material processing, we used laser peening to shape a 213 mm

M

x 21.6 mm x 3 mm thick bar of the 316L AM material. We successively peened the sample with a first, second, and third layer of peening and measured the arc height and radius of curvature generated with

D

each peening application. We then reversed the arc height to zero by peening the opposite surface. We

TE

used a laser irradiance of 4 GW/cm2, a pulse duration of 18 ns, square spot size of 4.7 mm, 3% spot-to-

EP

spot overlap with 100% coverage for each layer.

3 Results

CC

3.1 Fatigue

3.1.1 Mechanical tests

A

Fatigue tests of samples were run on a 90 kN Instron fatigue test rig comparing performance with and without notches and with and without shot peening and laser peening surface treatments. Figure 5 shows baseline fatigue test results for two sample types tested with no EDM notch: AM samples without and with shot peening. The residual compressive stress added by the shot peening clearly increases lifetime before crack initiation and growth to failure as defined by 10% sample deflection compliance during the

sinusoidal-loading fatigue testing. The blue diamonds in the figure represent the lifetimes of the unnotched AM samples loaded between 359 MPa to 574 MPa. The red squares represent the un-notched shot peened samples. Noteworthy is that the shot peened smooth samples did not fail and were declared

SC RI PT

runouts4 at both 400 MPa and 476 MPa loading. In this smooth condition, the shot peening provided

CC

EP

TE

D

M

A

N

U

essentially a 46% fatigue strength improvement to the AM 316L material.

Figure 5. Fatigue lifetime of additive manufactured sample and additive manufactured sample with shot peening applied. Surface treatment clearly enhances fatigue life of the smooth sample.

A

The presence of some form of stress riser, for example a fillet radius or cross section narrowing, required of structural geometry is more representative of a component in an actual system. For this reason, we introduced a notch into the testing schedule. Figure 6 shows the degradation in performance of the AM

4

A runout is where can cycle infinitely without failure or where the time to failure exceeds that available for the test.

samples following introduction of the 0.35 mm-deep notch in the top center of the gauge section. The stress riser of a factor of three induced by this notch is nominally indicative of the stress enhancement generated by a fillet radius or other geometric feature required in a fielded component. As can be seen,

SC RI PT

the fatigue lifetime decreased by about a factor of three to five and the fatigue strength decreased by about 45% in the case of the notch. For this work, we also fabricated to the same geometry a sample of wrought 316L and surprisingly found that un-notched AM samples tested out at approximately 50% greater fatigue strength and roughly 9 times longer lifetime. Although again the sample size is limited and this test was run at 2.3 times the 170 MPa yield strength of wrought 316L, it is somewhat surprising that

U

the AM material with notch actually slightly outperformed the wrought material in both fatigue lifetime

A

CC

EP

TE

D

M

A

N

and in fatigue strength.

Figure 6 Un-peened samples were tested with an EDM notch of 0.35mm” depth to represent a stress concentration factor of 3 typically created by sample geometry requirements. Note the significantly improved performance of the AM material compared to the wrought 316.

We anticipated that the deep level of plastic deformation and thus deep level of residual stress generated by the laser peening would be especially effective in improving the AM fatigue life in the presence of the notch stress riser. Test results comparing notched AM material with no peening, with shot peening, with

D

M

A

N

U

SC RI PT

laser peening, and finally with the notched wrought material are shown in Figure 7.

TE

Figure 7 Fatigue lifetime test results vs. stress loading for AM 316L stainless steel comparing AM without notch against un-peened, shot peened and laser peened material with notch applied. Samples were notched to 0.35 mm depth to provide a controlled crack initiation point. Also tested for normalization is a sample of wrought material with notch.

EP

At the 400 MPa stress level the AM notched material (Kt factor5 of 3) outperforms the notched wrought material. The shot peening provided a 2.7X lifetime improvement. Most notable is that at this stress

CC

loading the laser peened sample did not fail, exhibiting a runout with testing halted at 2,513 k- cycles which is greater than 20 times the lifetime of the untreated notched AM sample. Extrapolating the data,

A

it can be estimated that the laser peening provides a fatigue strength improvement in the range of 60% for the AM notched treated material and approximately 80% above the wrought material with no stress riser.

5

stress concentration factor

At the higher stress loading of 501 MPa the shot peening provided 1.9X lifetime improvement and the laser peening provided 6.8X improvement beyond the notched AM material. Also significant, at this high stress loading the laser peening effectively eliminated the fatigue debit of the 3x stress riser of the notched

SC RI PT

AM material.

3.1.2 Fracture surfaces

Figure 8 shows the backscattered electron images of the fracture surfaces of the (a) laser peened sample, the (b) shot peened sample, the (c) as-built sample, and the (d) wrought + annealed sample. The crack initiated at the top, curved surface and propagated toward the bottom of the image. Backscattered

U

electron images are particularly useful for revealing the topography of a surface. In the images, it is possible to discern differences in the fracture surface topography. To make the comparison quantitative,

A

N

a square region (indicated in the figure), with dimension 1.64 mm ✕ 1.64 mm2 immediately below the top

M

surface was isolated. This region includes the 3 mm to 4 mm region affected by laser peening. The 2D fast Fourier transform (FFT) is a useful tool to reveal spatial frequencies and anisotropy in images. The FFT of

D

each of the selected areas was computed and the result is displayed in Figure 8a-d. The FFTs of all samples

TE

exhibit anisotropy, with more intensity along the horizontal direction compared with the vertical direction (the directions in the FFT align with the images). This indicates a common directionality to the topography

EP

of all surfaces that runs perpendicular to the horizontal direction and has spatial frequency along this

CC

direction. There is significant similarity between the shot peened FFT and the as-built FFT, as might be expected, since the shot peening only affects a region that is 400 µm from the top surface. Both of these

A

have significant intensity peaks in the horizontal direction indicating a range of spatial frequencies in the surface structure. This long wavelength structure can be observed in the images. The laser peened FFT and the wrought + annealed FFT exhibit differences from the shot peened FFT and the as-built FFT. Both the laser peened FFT and the wrought + annealed FFT have four intensity peaks immediately next to the central bright spot. These are very long wavelength features that are associated with the size of the image

and should be neglected. The remaining intensity peaks in the laser peened FFT and the wrought + annealed FFT are lower in intensity compared with the shot peened FFT and the as-built FFT. This indicates that there is less directionality in the laser peened FFT and the wrought + annealed FFT compared with

SC RI PT

the shot peened FFT and the as-built FFT. We measured the similarity of these patterns by computing the sum of the absolute value of the difference between the various FFTs. Using this metric, we found that the most similar structures were the shot-peened and the as-built structures. The laser peened structure was only slightly different (4%) from the shot-peened and the as-built structures. The laser peened surface structure, the shot peened

U

surface structure, and the as-built surface structure, were significantly (20%) different from the wrought

N

+ annealed surface structure.

A

CC

EP

TE

D

M

vertical lines that intersect the center of the FFT.

A

For a more quantitative comparison, we extracted plots of the intensity of the FFT on horizontal and

Figure 9 shows the horizontal intensities as a function of distance along a line containing the center spot. Both the shot peened and the as-built sample exhibit higher intensities in the 4-8 mm-1 region.

D

M

A

N

U

SC RI PT

These are associated with long wavelength features.

TE

Figure 9b shows the vertical intensities as a function of distance from the center spot. Peaks in the shot peened sample and the as-built sample are observed at 2.0 mm-1. Peaks in the laser peened sample and

A

CC

similarity.

EP

the wrought + annealed sample are observed at 2.5 mm-1. Beyond this point, the four FFTs have some

We further compare the FFTs by plotting the intensity in the FFT integrated at a given inverse distance

M

A

N

U

SC RI PT

as a function of inverse distance. The result is shown in

D

Figure 9c. We note here that in the long wavelength (short inverse distance) regime, there are

TE

differences among the FFTs. First, there is significantly more intensity in the range of 4-8 mm-1 in the shot peened surface compared with the laser peened, as-built, and wrought + annealed surfaces. This

EP

indicates the presence of more frequencies in this range. Second, the laser peened and the as-built

A

CC

surfaces are similar. Finally, the wrought + annealed intensity is generally lower than the laser peened

A

N

U

SC RI PT

intensity. But, there is a cross over at about 10 mm-1, see

M

Figure 9c, after which, the wrought + annealed intensity exceeds the laser peened intensity. This is due

D

to the much higher frequency (short wavelength) structure that appears in the wrought + annealed

A

CC

EP

TE

surface visible in Figure 8. None of the additively manufactured surfaces exhibited the very fine scale

A

N

U

SC RI PT

(high frequency) topography observed in the wrought + annealed surface.

M

Figure 9d shows the normalized integrated intensity as a function of angle from the horizontal. It shows

D

that relative to the other fracture surfaces, the laser peened surface has significant intensity at angles

A

CC

EP

TE

close to 90°.

SC RI PT U N A M D TE EP

A

CC

Figure 8 Fracture surfaces and Fourier transforms for the laser peened, shot peened, as-built, and wrought + annealed fracture surfaces.

SC RI PT U N

A

Figure 9 Intensity as a function of inverse distance for a horizontal line through the center of the FFTs.

D

EP

  

The shot-peened structure stands out from the other surfaces due to its low inverse distance peaks All structures are anisotropic Wrought + annealed is less anisotropic than shot-peened with high frequency features Laser peened is less anisotropic than shot-peened or as -built but not as anisotropic as wrought + annealed

TE



M

We conclude that

A logical conclusion is that the laser peening induced an effect in the microstructure that caused the

CC

microstructure to deform and fracture under fatigue loading inducing low frequency features in the fracture surface that were less directional than in the case of the shot peened surface or the as-built

A

surface.

3.2 Precision forming of AM Material Initially the AM bar used for the forming demonstration had an arc height less than 0.1 mm over the 213 mm length as measured with a Leica laser interferometer. With each successive layer application of

peening the arc height and shape were again measured. Figure 10 shows the shape of the bar with arc

U

SC RI PT

height of 5 mm after the third peening layer.

N

Figure 10 Bar of AM 316L after 3 layers of peening attained arc height of 5 mm. Peening on the reverse side with three identical layers flattened it to essentially a measurement limit

A

To demonstrate the potential of corrective forming, we then sequentially peened three layers of peening

M

on the reverse side measuring the shape of the bar between layer applications. Figure 11 shows the arc

D

height across the bar for each peening layer with the corresponding radius of curvature reaching 112 cm.

TE

As can be seen, the arc height reversed almost precisely sequentially with each subsequent layer of peening. Finally, the bar was flattened to a measurement limited arc height of 250 µm, essentially

EP

correcting all of the previously applied “distortion”. The peening pattern for an individual layer was comprised of 480 laser impact spots. The high granularity of this processing indicated that if even finer

CC

precision were required, a small number of additional impacts could be strategically applied over specific

A

areas of the bar to further refine to a desired shape.

SC RI PT

U

Figure 11. Arc height of 316L AM bar reached 5 mm after 3 layers of peening were applied to the top side. The bar then was flattened back to zero with 3 layers applied on the back side.

N

4 Discussion and Conclusions

A

The tests we performed show the fatigue strength of the AM material produced with the described

M

equipment and process to be superior to that of wrought material, a result that lends well for use of the

D

additively manufactured 316L material in real applications. Both shot and laser peening further enhanced

TE

the fatigue performance. We speculate that the improvement made by the shot peening of the smooth samples (Kt=1) is a case of added compressive stress and increased hardness due to the cold work of the

EP

shot peening. The benefit of the laser peening is due more to imparting a deeper level of residual stress as the laser peening is known to produce very little cold work and thus minimally change material

CC

hardness and yield strength. Improvements observed for the case of the notched (Kt=3) samples indicate that laser peening would be especially beneficial for applications where geometry requirements create

A

areas of increased stress such as in fillets and notched areas leading to local stress risers.

5 Discussion It should be noted that this work involved a limited data set, so conclusions must be made with care. However, the performance enhancements and limited test variances appear to be well beyond expected

statistical variation. Also, the use of the notched samples aided in generating failures that were locally stress related and not as dependent on the random location and size of flaws and inclusions in the material. The general consistency of the test results does suggest that imperfections were of a size smaller

SC RI PT

than that which would have initiated a 3X increased stress riser. We also demonstrated that AM material can be formed and the shape can be modified by the laser peening. The forming demonstration highlights the benefit of laser peening to also correct distortions in AM panels and components. The road map to apply the method for processing of the specific part is presented. It was already tested and implemented for forming aerospace components. The laser

U

technology can precisely correct component shapes thereby minimizing the need to mechanically load

N

and consequently stress sections during assembly such as with pull down clamps to obtain proper fit as

A

the straining needed to fit up components during assembly of parts can often be a major factor in

M

increasing tensile stress and consequently limiting fatigue lifetime. The laser peening process is highly controlled and the application discrete allowing precise shape corrections by means of adding beneficial

D

compressive stress. Consequently, peening technology can be a highly beneficial contributor to the

TE

adoption of AM materials in critical applications.

EP

Laser peening is a more expensive surface treatment than other forms of peening if applied over large areas. However, the process has proven to be highly cost effective and a solution to problems for which

CC

other surface treatments have failed. When applied strategically to localized high stress areas, its deep level of plastic deformation provides superior fatigue enhancement. In forming operations, the process

A

generates degrees of strain providing solutions that other processes are unable to achieve. The highly deterministic nature of applying laser peening via individually controlled shots, the ability to selectively apply it over small defined areas and most importantly the very deep levels of residual stress have made it especially effective in the aerospace, power generation and nuclear industries.

In the field of AM components, there has been consideration of online, layer by layer peening. [25] If using this approach, in addition to the need for additional processing time, the need for a tamping layer presents a cost and process complication. An even more serious consideration is the potential to develop

SC RI PT

compensating tensile stress beneath the individual layers being processed, stresses that could lead to distortions and internally induced fatigue failures. With a well-controlled AM process we believe it will be very effective to build components and engineer and induce laser peening post-AM build to add stress

A

CC

EP

TE

D

M

A

N

U

and strain to those selected areas needing treatment.

References:

A

CC

EP

TE

D

M

A

N

U

SC RI PT

[1] T. DebRoy, H.L. Wei, J.S. Zuback, T. Mukherjee, J.W. Elmer, J.O. Milewski, A.M. Beese, A. Wilson-Heid, A. De, W. Zhang, Additive manufacturing of metallic components – Process, structure and properties, Prog. Mater. Sci. 92 (2018) 112-224. [2] B. Vrancken, Study of Residual Stresses in Selective Laser Melting, Engineering Science, KU Leuven, Leuven, 2016. [3] B. AlMangour, J.-M. Yang, Improving the surface quality and mechanical properties by shot-peening of 17-4 stainless steel fabricated by additive manufacturing, Materials & Design 110 (2016) 914-924. [4] B. AlMangour, J.-M. Yang, Integration of Heat Treatment with Shot Peening of 17-4 Stainless Steel Fabricated by Direct Metal Laser Sintering, JOM 69(11) (2017) 2309-2313. [5] F. Calignano, D. Manfredi, E.P. Ambrosio, L. Iuliano, P. Fino, Influence of process parameters on surface roughness of aluminum parts produced by DMLS, The International Journal of Advanced Manufacturing Technology 67(9-12) (2013) 2743-2751. [6] E. Wycisk, C. Emmelmann, S. Siddique, F. Walther, High Cycle Fatigue (HCF) Performance of Ti-6Al-4V Alloy Processed by Selective Laser Melting, Advanced Materials Research 816-817 (2013) 134-139. [7] R.I. Stephens, H.O. Fuchs, Metal fatigue in engineering, Wiley, New York, 2001. [8] B.P. Fairand, A.H. Clauer, Effect of water and paint coatings on the magnitude of laser-generated shocks, Optics Communications 18(4) (1976) 588-591. [9] Anonymous, Boeing Awards MIC Laser Peening Contract to Form 747-8 Wing Sections, Curtiss-Wright, 2008. [10] Manufacturing Group, Curtiss-Wright Awarded Rolls-Royce Contract, 2013. http://www.aerospacemanufacturinganddesign.com/article/uk-to-fund-development-of-jet-rockethybrid-engine-reaction-sabre-071713/. (Accessed September 26, 2017 2017). [11] Anonymous, Laser Peening by CWST Contributes to Successful Nuclear Canister Storage Program, The Shot Peener magazine 32(2) (2018). [12] K. Raja, M. Nathan M, T. Patil Balram, C.D. Naiju, Study of Surface Integrity and Effect of Laser Peening on Maraging Steel Produced by Lasercusing Technique, SAE Technical Paper Series, SAE International, 2018, pp. 2018-28-0094. [13] N.C. Anderholm, LASER‐GENERATED STRESS WAVES, Applied Physics Letters 16(3) (1970) 113-115. [14] P.S. Prevey, D.J. Hornbach, P.W. Mason, Thermal Residual Stress Relaxation and Distortion in Surface enhanced Gas Turbine Engine Components. , in: D. Milam, J. Dale A. Poteet, G.D. Pfaffman, V. Rudnev, A. Muehlbauer, W.B. Albert (Eds.) Heat Treating 1997: Proceedings of the 17th Conference, Indianapolis, IN., 1997. [15] A. Dewald, Measurement and modeling of laser peening residual stresses in geometrically complex specimens, University of California, Davis, 2005. [16] W. Cheng, I. Finnie, KII solutions for an edge-cracked strip, Engineering Fracture Mechanics 36(2) (1990) 355-360. [17] Z.Y. Chen, R.G. Kelly, Computational Modeling of Bounding Conditions for Pit Size on Stainless Steel in Atmospheric Environments, Journal of The Electrochemical Society 157(2) (2010) C69-C78. [18] M.T. Woldemedhin, M.E. Shedd, R.G. Kelly, Evaluation of the Maximum Pit Size Model on Stainless Steels under Thin Film Electrolyte Conditions, Journal of The Electrochemical Society 161(8) (2014) E3216E3224. [19] Anonymous, Standardization Roadmap for Additive Manufacturing, America Makes & ANSI Additive Manufacturing Standardization Collaborative (AMSC), 2017. [20] A.B. Spierings, T.L. Starr, K. Wegener, Fatigue performance of additive manufactured metallic parts, Rapid Prototyping Journal 19(2) (2013) 88-94.

A

CC

EP

TE

D

M

A

N

U

SC RI PT

[21] R. Shrestha, u. Simsiriwong, N. Shamsaei, S.M. Thompson, L. Bian, EFFECT OF BUILD ORIENTATION ON THE FATIGUE BEHAVIOR OF STAINLESS STEEL 316L MANUFACTURED VIA A LASER-POWDER BED FUSION PROCESS, 26th Annual International Solid Freeform Fabrication Symposium – An Additive Manufacturing Conference, Austin, TX, 2016, pp. 605-616. [22] A. Riemer, S. Leuders, M. Thone, H.A. Richard, T. Troster, T. Niendorf, On the fatigue crack growth behavior in 316L stainless steel manufactured by selective laser melting, Engineering Fracture Mechanics 120 (2014) 15-25. [23] A. Riemer, H.A. Richard, J.‐P. Brüggemann, J.-N. Wesendahl, Fatigue crack growth in additive manufactured products, Frattura ed Integrità Strutturale 34 (2015) 437-446. [24] N. Kalentics, E. Boillat, P. Peyre, S. Ćirić-Kostić, N. Bogojević, R.E. Logé, Tailoring residual stress profile of Selective Laser Melted parts by Laser Shock Peening, Additive Manufacturing 16 (2017) 90-97. [25] N. Kalentics, E. Boillat, P. Peyre, C. Gorny, C. Kenel, C. Leinenbach, J. Jhabvala, R.E. Logé, 3D Laser Shock Peening – A new method for the 3D control of residual stresses in Selective Laser Melting, Materials & Design 130 (2017) 350-356. [26] M. ZHANG, H. LI, X. ZHANG, D. HARDACRE, REVIEW OF THE FATIGUE PERFORMANCE OF STAINLESS STEEL 316L PARTS MANUFACTURED BY SELECTIVE LASER MELTING, Proc. of the 2nd Intl. Conf. on Progress in Additive Manufacturing, Research Publishing, Singapore, Singapore, 2016, pp. 563-568. [27] S. Leuders, T. Lieneke, S. Lammers, T. Tröster, T. Niendorf, On the fatigue properties of metals manufactured by selective laser melting – The role of ductility, J. Mater. Res. 29(17) (2014) 1911-1919. [28] S.M.J. Razavi, P. Ferro, F. Berto, J. Torgersen, Fatigue strength of blunt V-notched specimens produced by selective laser melting of Ti-6Al-4V, Theoretical and Applied Fracture Mechanics (2017). [29] G. Nicoletto, Directional and notch effects on the fatigue behavior of as-built DMLS Ti6Al4V, International Journal of Fatigue 106(Supplement C) (2018) 124-131. [30] M. Zhang, C.-N. Sun, X. Zhang, P.C. Goh, J. Wei, H. Li, D. Hardacre, Elucidating the Relations Between Monotonic and Fatigue Properties of Laser Powder Bed Fusion Stainless Steel 316L, JOM 70(3) (2017) 390– 395 [31] Y. Kajima, A. Takaichi, T. Nakamoto, T. Kimura, N. Kittikundecha, Y. Tsutsumi, N. Nomura, A. Kawasaki, H. Takahashi, T. Hanawa, N. Wakabayashi, Effect of adding support structures for overhanging part on fatigue strength in selective laser melting, Journal of the Mechanical Behavior of Biomedical Materials 78 (2018) 1-9. [32] J.O. Almen, Shot blasting test, United States Patent 5230440, June 6, 1944.