Content uploaded by Adam Thompson
Author content
All content in this area was uploaded by Adam Thompson on Feb 03, 2021
Content may be subject to copyright.
Joint Special Interest Group meeting between euspen and ASPE
Dimensional Accuracy and Surface Finish in Additive Manufacturing,
KU Leuven, BE, October 2017
www.euspen.eu
Defect evolution in laser powder bed fusion additive manufactured components
during thermo-mechanical testing
Zhengkai Xu, C. J. Hyde, A. Thompson, R. K. Leach, I. Maskery, C. Tuck, A. T. Clare
Faculty of Engineering, The University of Nottingham, University Park, NG7 2RD, United Kingdom
zhengkai.xu@nottingham.ac.uk
Abstract
The mechanical performance of additively manufactured (AM) components remains an issue, limiting the implementation of AM
technologies. In this work, a new method is presented, to examine the evolution of defects in an Inconel 718 two-bar test specimen,
manufactured by laser powder bed fusion AM, during thermo-mechanical testing. The test was interrupted at specific
extensions of the specimen, and X-ray computed tomography measurements performed. This methodology has allowed, for the first
time, the evolution of the defects in an AM specimen to be studied during a thermo-mechanical test. The number and size of the
defects were found to increase with time as a result of the thermo-mechanical test conditions, and the location and evolution of
these defects have been tracked. Defect tracking potentially allows for accurate prediction of failure positions, at the earliest possible
stage of a thermo-mechanical test. Ultimately, when the ability to locate defects in this manner is coupled with manipulation of build
parameters, laser powder bed fusion practitioners will be able to further optimise the manufacturing procedure in order to produce
components of a higher structural integrity.
Creep, Two-bar specimen, X-ray computed tomography, Laser powder bed fusion, Mechanical testing, Nickel superalloys
1. Introduction
Laser powder-bed fusion (LPBF) additive manufacturing (AM)
is a near-net-shape technique which provides a high design
flexibility to components such as aero-engine turbine blades.
However, few studies have considered thermo-mechanical
properties such as creep, of specimens fabricated by such
techniques. Rickenbacher et al. [1] investigated the creep
performance of an LPBF fabricated In738LC specimen; the
specimen anisotropic, and also
slightly inferior to conventionally manufactured specimens.
Pröbstle et al. [2] studied the creep properties of LPBF
manufactured Inconel 718 specimens and found these
specimens possess better creep performance than cast and
wrought specimens, though fracture results were not included
in their study. Conventionally, creep testing cannot provide
direct evidence regarding defect evolution during a test, thus
staged creep testing has been implemented. Through the use of
X-ray computed tomography (XCT) at each stage, it is possible to
study defect evolution in a non-destructive manner and
continue mechanical testing. Babout et al. [3] designed an in-situ
tomography system and applied it in room temperature tensile
testing, obtaining seven scans at different stages of a tensile
test. The propagation of defects in the specimen was
successfully measured. Hangai et al. [4] applied a similar concept
and compared the conditions of pores in an aluminium specimen
pre and post fatigue testing of the specimen. This method gave
direct evidence of defect development and a better
understanding of the specimen behaviour during testing
through determination of the fracture origin. The method can
also be extended to other mechanical test types.
2. Methodology
In this study, a two-bar small specimen (TBS) was used to study
creep (as shown in figure 1, in which the bold arrow indicates
the building direction, L for left bar and R for right bar), and
prove the practicality . The TBS
is a relatively small-sized specimen, designed by Hyde et al. [5],
which can be used to obtain both creep strain rate and fracture
life. Specimens of this geometry were built using a Renishaw
AM250 LPBF system with commercial Inconel 718 powder
provided by Renishaw PLC and milled to the specific dimensions
shown. A specimen was tested with a tensile stress of
747.45 MPa at 650 and compared to a reference study by
Sugahara et al. [6]. The test rig is shown in figure 2, where the
TBS is mounted in the loading element and then encased in the
heating element.
Figure 1. Dimensions of the TBS.
113
Figure 2. (a) Loading element and (b) heating element that make up the
test rig.
A staged mechanical test was designed based on extension data
obtained from preliminary experiments. The creep test was
interrupted when the extension reached a certain distance, and
the specimen was removed from the testing kit for XCT
measurement. A Nikon MCT 225 was used to perform XCT
measurements (settings: source voltage 225 kV, source current
×,
yielding a reconstructed voxel size of 10 ± 0.2 ). The test
strategy is shown in figure 3.
Figure 3. Staged creep testing strategy.
XCT image processing was performed using MATLAB [7] and
ImageJ [8] to examine pores, based on thresholding via the
ISO50 surface determination method [9]. The two-bar section
(as shown in figure 4) was considered and measured.
Figure 4. XCT images and the convention for the two-bar section.
3. Results and discussion
3.1. Porosity distribution in the TBS before mechanical testing
The right bar has higher porosity than the left (as shown in
figure 5) and the porosity distribution has no regular pattern,
which indicates that the pores in the AM manufactured parts are
randomly distributed. The study made by et al. [10]
showed that the pore distribution was largely dependent on
. The section which has the highest
porosity, i.e. peak 1, can be identified in the curve. The defects
found at this position were considered at the first stage to be the
likely cause of eventual fracture.
Figure 5. Porosity distribution in the TBS before testing.
3.2. Cross-sectional area changes
Figure 6 shows the decrease of the cross-sectional area in the
two-bar section over the course of the creep test (necking). In
this case, the most obvious reduction in the cross-sectional areas
are located near the two ends. There is a 30% reduction of the
cross-sectional area in the third stage when compared to the
first stage and these positions represent likely potential fracture
points. The negative end shown in figure 6 has an approximately
8% higher area reduction than the positive end. The negative
end is more likely to
Figure 6. Cross-sectional area changes along the build direction.
3.3. Porosity distribution changes during mechanical testing
114
Figure 7 shows the porosity distribution change over time
along the two bars of the specimen, with the greatest porosity
increases being near the two ends. T
occurred in the negative end, as shown in figure 7. Peak 1 (same
as the peak 1 in figure 5) has the highest porosity in the first
stage (as-build condition). This peak disappeared in the fourth
stage, indicating that the defects at this point resulted in the
eventual fracture. This result agrees with the prediction made in
section 3.1.
Figure 7. Porosity distribution change during testing.
3.4. Performance of the fracture section during the test
The reconstructed 3D model (as shown in figure 8) of the
fracture section clearly indicates defects evolution. Pore 1 has
an obvious ring shape in the first stage, which grew into a
spherical pore during the test. Some newly developed pores are
also observed in the third stage, which grew much faster than
pore 1. These defects together cause the eventual fracture of
the specimen. The small pore (pore 2 in figure 8) experienced
almost no change during the testing.
Figure 8. Reconstructed 3D model of the fracture section.
4. Conclusion
Defects in the as-build LPBF manufactured specimen are
normally distributed randomly, and sections with the highest
porosity are more likely lead to fracture. In the tested specimen,
the most porous sections were found to be located at each end
of the gauge sections, and the area of the respective cross-
sections dropped significantly in these sections during testing.
The evolution of pores in the fracture section involved the
growth of existing pores in addition to the generation of the new
pores. These mechanisms both contribute to specimen fracture.
The combination of the staged testing and XCT can be applied
to estimate the potential fracture points and gain information of
the subsequent specimen response, such as the pore evolution,
in testing.
Acknowledgement
The authors would like to acknowledge the Engineering and
Physical Sciences Research Council (EPSRC Grants
EP/L017121/1, EP/M008983/1 and EP/L01534X/1) for funding
this work. In addition, the authors would like to thank Alexander
Jackson-Crisp for his invaluable technical contributions in
machining specimens, Shane Maskill for facilitating the creep
testing and to Renishaw Plc, process engineering team for
providing Inconel 718 samples.
References
[1] Rickenbacher L, Etter T, Hövel S and Wegener K 2013 Rapid
Prototyping J. 19 282-29013
[2] Pröbstle M, Neumeier S, Hopfenmüller J, Freund L P, Niendorf T,
Schwarze D and Göken M 2016 Mat. Sci. Eng. A 674 299-307
[3] Babout L, Maire E, Buffière J Y and Fougères R 2001 Acta
Materialia 49 2055-2063
[4] Hangai Y, Kuwazuru O, Yano T, Utsunomiya T, Murata Y, Kitahara S,
Bidhar S and Yoshikawa N 2010 Mat. Trans. 51 1574-1580.
[5] Hyde T H, Ali B S M and Sun W 2013 J. Eng. Mat. Tech. 135 041006-
041006
[6] Sugahara T, Martinolli K, Reis D AP, Neto C de Moura, Couto A A,
Neto F P and Barboza M J R 2012 Defect and Diffusion Forum 326-
328 509-514
[7] Marques O 2011 Practical Image and Video Processing Using
MATLAB (Hoboken, New Jersey, Wiley) 368-370
[8] Schneider C A, Rasband W S and Eliceiri K W 2012 Nat Meth 9 671-
675
[9] Otsu N 1979 IEEE Transactions on Systems, Man, and Cybernetics 9
62-66
[10] G, Chlebus E, Szymczyk P and Kurzac J 2014 Archives of
Civil and Mechanical Engineering 14 no. 4 608-614
115
