Prakt. Met. Sonderband 50 (2016) 21
cently. Even the Depending on the source, there appears to be at least a moderate mismatch in the coefficient of
) printing shall be thermal expansion (CTE) between the main burner body, consisting mainly of the low-alloy heat
A has been around resistant steel 16Mo3, 1.5415, and the Y pipe. The former has a CTE of between 11 x 10° K” at
n gaining impetus room temperature and 13.5 x 10 K! at 400°C, the latter of approximately 14 x 10° K!, while
es for prototyping other sources claim there was no mismatch. Because of the CTE mismatch, a thermal stress will
if small series of develop upon each and every start-up and shut-down of the burner rig. Since the Y pipe will expand
of actual design more than the main burner body, a bulging of the pipe wall, resulting in a hoop stress, is expected to
build up. This hoop stress is hypothesized to have finally cracked the Y pipe by tensile overload,
after having grown a fatigue crack into the wall thickness of the tube material. In a burner rig, such
me thing, such as alternating stresses are more pronounced than in regular engine service.
1), laser sintering
ng (LM), to name While the microstructural imperfections described above are considered to be a weak contributing
or other materials factor only, the CTE mismatch is suspected to be a more significant issue contributing to the failure.
resentation of the What is more, residual stresses that normally develop during the SLM processing also might have
ural materials has contributed to the failure by superposition with applied cyclic stresses. Additive manufacturing
mited geometrical processes are known for their rather high after-process residual stress state that may or may not
materials, giving have been relieved by post-SLM stress relief annealing. Measuring residual stresses in this case was
e turbomachinery pointless since the cracking had already relieved these stresses. This should have been done post-
al efficiency. For SLM processing but no need for this was seen at the time.
‘metal powder are
y laser beam”, the The subject Y pipe failed by thermal-mechanical fatigue (TMF). Two long longitudinal cracks ran
along both cavities of the Y pipe. They both initiated on the inside and propagated into the wall
thickness. The one in the larger cavity penetrated the wall thickness and caused the indications upon
us alloys, such as penetrant testing. CTE mismatch, frequent cycling of the burner rig, temperature difference between
al properties close fuel gas on the inside and compressor exhaust on the outside, thermal shock upon purging for fuel
have to take into switchovers, and possibly residual stresses all contributed to the failure. The rough as-built internal
particular, SLM pipe surface and the notch effect that caused may well have been a contributing factor, too.
dients and rapid
ns requiring high If it is desired that the Y pipe be made by SLM, it is recommended to select a material with no or
) strength may not very low CTE mismatch or to insert in the burner assembly components designed to remove
idual stresses that restraint as much as possible, such as bellows. In any case, 17-4 PH is overdesigned for this
e superposition of particular application. However, in the opinion of the authors, the subject Y pipe is not a good
g case study will candidate component for additive manufacturing, since conventional production routes, such as
rolling and welding, still have greater damage tolerance than components made by non-
conventional manufacturing. It is known from the literature that the build strategy employed for
v microscopically additive manufacturing might be critical to the low cycle fatigue resistance of components made by
tic rough as-built SLM. It is quite possible that improvements to the microstructure could be made for the subject Y
pipe by developing a different build strategy, in an attempt to remove contributing microstructural
factors as far as is feasible.
a long and gaping
parent that the Y
c features that are
s. Microstructural
ections, such as
to the failure. An
ee Fig. 8: Y-pipe as built