text
stringlengths
0
43
detected O
in-situ S-CONPRI
via O
pyrometry S-CHAR
during O
laser B-MANP
powder I-MANP
bed I-MANP
fusion I-MANP
additive I-MANP
manufacturing E-MANP
and O
correlated S-CONPRI
with O
voids S-CONPRI
observed O
using O
post-build O
micro-computed B-CHAR
tomography E-CHAR
. O
Large O
two-color O
pyrometry S-CHAR
data S-CONPRI
sets O
were O
used O
to O
estimate O
instantaneous O
temperatures S-PARA
, O
melt B-MATE
pool E-MATE
orientations O
and O
aspect B-FEAT
ratios E-FEAT
. O
Machine B-ENAT
learning I-ENAT
algorithms E-ENAT
were O
then O
applied O
to O
processed S-CONPRI
pyrometry O
data S-CONPRI
to O
detect O
outlier O
images S-CONPRI
and O
conditions O
. O
It O
is O
shown O
that O
melt B-MATE
pool E-MATE
outliers O
are O
good O
predictors O
of O
voids S-CONPRI
observed O
post-build O
. O
With O
this O
approach O
, O
real O
time O
process B-CONPRI
monitoring E-CONPRI
can O
be S-MATE
incorporated O
into O
systems O
to O
detect O
defect S-CONPRI
and O
void S-CONPRI
formation O
. O