FIELD OF THE INVENTION
5 The present description relates to a method of
fabricating a part by additive fabrication, in particular
by melting or sintering particles of powder by means of a
high energy beam.
The method is particularly suitable for fabricating
10 parts that present asymmetry or large mass disparities,
in particular for the field of aviation.
STATE OF THE PRIOR ART
It is already known, in particular in the field of
15 aviation, to use additive fabrication methods in order to
make certain parts of shapes that are fine or complex.
A conventional example of additive fabrication is
fabrication by melting or sintering powder particles by
means of a high energy beam. Among such high energy
20 beams, mention may be made in particular of laser beams
and of electron beams.
The term "selective laser melting" (SLM) is used to
designate a method having the main characteristics as set
out below with reference to Figure 1, which shows a
25 conventional device for fabricating a part by selective
melting or sintering of beds of powder by means of a
laser beam.
A first layer 10a of powder of a material is
deposited, e.g. with the help of a roller 20 (or any
30 other deposition means) on a fabrication plate 21 (which
may be a single plate or which may be surmounted by a
solid support, or by a portion of another part, or by a
support grid used to facilitate constructing certain
parts).
' Translation of the title as established ex officio.
The powder is transferred from a feed bin 22 during
go and return movement of the roller 20 and it is then
scraped and possibly also compacted a little during one
or more return movements of the roller 20. The powder is
5 made up of particles 11. Excess powder is recovered in a
recycling bin 23 situated adjacent to the construction
bin 24 in which the fabrication plate 21 moves
vertically.
Use is also made of a generator 30 for generating a
10 laser beam 31 and of a control system 32 for directing
the beam 31 onto any region of the fabrication plate 21
so as to scan any region of a previously deposited layer
of powder. The laser beam 31 is shaped and its diameter
in the focal plane is varied respectively by means of a
15 beam expander 33 and by means of a focusing system 34,
which together constitute the optical system.
Thereafter, a region of the first layer 10a of
powder is raised to a temperature higher than the melting
temperature of the powder by being scanned with a laser
20 beam31.
The SLM method may use any high energy beam instead
of the laser beam 31, and in particular it may use an
electron beam, providing the beam has sufficient energy
to melt the particles of powder and a portion of the
25 material on which the particles rest.
By way of example, the beam may be caused to scan by
a galvanometer head that forms part of a control system
32. For example, the control system includes at least
one steerable mirror 35 on which the laser beam 31 is
30 reflected before reaching a layer of powder where each
point of the surface is always situated at the same
height relative to the focusing lens contained in the
focusing system 34, with the angular position of the
mirror being controlled by a galvanometer head so that
35 the laser beam scans at least a region of the first layer
of powder, and thus follows a pre-established profile for
a part. For this purpose, the galvanometer head is
controlled using information contained in the database of
the computer tool used for the computer assisted design
and fabrication of the part that is to be fabricated.
Thus, the particles of powder 11 in this region of
5 the first layer 10a are melted and they form a first
single-piece element 12a that is secured to the
fabrication plate 21. At this stage, it is also possible
to use the laser beam to scan a plurality of independent
regions of the first layer so as to form a plurality of
10 mutually disjoint first elements 12a once the material
has been melted and has solidified.
The fabrication plate 21 is lowered through a height
corresponding to the thickness of the first layer of
porider 10a (through 20 micrometers (pm) to 100 pm, and
15 generally through 30 pm to 50 urn).
Thereafter, a second layer lob of powder is
deposited on the first layer 10a and on the first singlepiece
or consolidated element 12a, and the second layer
lob as situated in part or completely over the first
20 single-piece or consolidated element 12a as shown in
Figure 1 is heated by being exposed to the laser beam 31
in such a manner that the powder particles in this region
of the second layer 10b are melted together with at least
a portion of the element 12a so as to form a second
25 single-piece or consolidated element 12b, with these two
elements 12a and 12b together forming a single-piece
block in the example shown in Figure 1.
Such an additive fabrication technique, or other
techniques such as fabrication by porider projection, thus
30 provides excellent control over the shape of the part
that is to be fabricated and enables parts to be made
that are very fine.
Nevertheless, those techniques require carefully
thought-out construction strategies that are specific to
35 the part that is to be fabricated in order to comply with
its dimensional tolerances and in order to ensure it has
good mechanical strength. It is thus necessary to use
computer assisted design (CAD) software in order to
define the best position and the best orientation for the
part for the purpose of fabricating it layer by layer.
Although fabrication strategies are still developing and
5 being refined, they are not at present satisfactory for
obtaining parts that are asymmetrical or that present
large mass disparities.
When making such parts by additive fabrication, it
is frequently observed that there are irreversible
10 metallurgical defects, e.g. the appearance of cracks,
and/or dimensional defects, with certain portions of the
part not complying with the specified tolerances. Mass
disparities, as a result in particular from asymmetries
of the part, lead to residual stresses accumulating in
15 certain zones of the part, which residual stresses then
lead to deformations: these poorly balanced residual
stresses thus give rise to geometrical dispersions that
are responsible for the defects observed in the resulting
parts. Unfortunately, such defects are often
20 unacceptable and lead to the resulting part being
scrapped, thereby giving rise to considerable losses and
thus high overall fabrication costs.
There therefore exists a real need for a method of
fabricating a part by additive fabrication that is
25 adapted to such parts that are asymmetrical or that
present large mass disparities.
SUMMARY OF THE INVENTION
The present description relates to a method of
30 fabricating a part by additive fabrication, the method
comprising the following steps: supplying a digital model
of a part to be fabricated; orienting the model relative
to a construction direction for constructing the part;
modifying the model by adding a sacrificial balancing
35 fraction configured so as to balance the residual
stresses that appear in the part while it is being
fabricated; making a rough part layer by layer using an
additive fabrication technique on the basis of the model
as modified in this way, said layers being stacked in the
construction direction; and using a material-removal
method to eliminate the sacrificial portion from the
5 rough part as results from the sacrificial balancing
fraction of the model, thereby obtaining said part that
is to be fabricated.
By means of this method, it is possible during the
stage of computer assisted design to detect a potential
10 risk of residual stresses accumulating during fabrication
as a result in particular of asymmetries within the part,
or at least of large mass disparities, and then to
correct the model of the part artificially so as to give
it an overall shape that is more regular and better
15 proportioned so as to enable residual stresses within the
part to be balanced during fabrication.
Thus, during layer-by-layer fabrication, residual
stresses become distributed within the part in more
uniform manner: this avoids these residual stresses
20 becoming concentrated in certain regions of the part to
above a certain threshold likely to lead to critical
deformations of the part. For example, adding such a
sacrificial fraction can make it possible to reduce
certain edge effects or to shift a region of stress
25 concentration towards a portion of the part that is less
sensitive to deformation, e.g. a portion that is thicker
or that possesses a shape that is particularly simple, or
towards a portion of the part in which mechanical or
dimensional tolerances are greater.
30 Under such circumstances, the rough part that is
obtained presents fewer defects, both dimensionally and
mechanically: it then suffices to use a conventional
material-removal method to remove the sacrificial
balancing portion from the rough part that results from
35 the sacrificial balancing fraction of the model, thereby
obtaining the desired part.
By means of this method, it is thus possible to use
additive fabrication for obtaining a part that is
asymmetrical or that presents great disparity, 161hile
benefiting from all of the advantages of additive
5 fabrication, and while nevertheless presenting few or no
defects.
In certain implementations, the part to be
fabricated possesses an asymmetrical portion, and the
sacrificial balancing fraction is configured in such a
10 manner that the sacrificial portion of the rough part
possesses mass lying in the range 70% to 130% of the mass
of the asymmetrical portion, preferably in the range 90%
to 110%.
The term "asymmetrical portion" is used to mean a
15 portion that, if it were to be removed from the part,
would leave a residual part possessing at least one more
element of symmetry than the original part. This
definition can be transposed directly to the model.
The term "element of symmetry" is used to mean
20 symmetry relative to a given plane, symmetry relative to
a given point, invariance relative to a given rotation,
or indeed any other invariance as a result of a given
geometrical relationship.
Throughout this description, the concept of symmetry
25 should be understood with a certain amount of tolerance:
thus, an element or a pair of elements is said to be
symmetrical providing at least 90% of the element or of
the pair is indeed symmetrical in the strict geometrical
sense; minor local differences, if any, are thus not
30 taken into consideration. This definition extends to the
planes of symmetry and more widely to all elements of
symmetry.
By means of this sacrificial portion that possesses
mass relatively close to the mass of the asymmetrical
35 portion, it is possible to correct, at least in part, the
mass disparity that results from the asymmetrical
portion: such an approach makes it possible to obtain
better balancing of residual stresses in the rough part,
both easily and in substantial manner, thereby obtaining
a remarkably favorable impact on the unwanted occurrence
of defects in the rough part.
5 In certain implementations, the sacrificial
balancing fraction is added over a height substantially
equivalent to the height of the asymmetrical portion.
Thus, the residual stresses are rebalanced for
substantially all of the layers that were originally
10 asymmetrical and that thus originally presented mass
disparity.
In certain implementations, the sacrificial
balancing fraction extends the model in its longest
direction.
15 In certain implementations, the sacrificial
balancing fraction is configured in such a manner that
the resulting sacrificial portion is constructed opposite
from the asymmetrical portion relative to the rough part.
In this way, the distribution of mass within the rough
20 part is rebalanced, thereby rebalancing the residual
stresses that are caused to appear during fabrication,
with these stresses being shifted in part towards the
center of the part.
In certain implementations, the sacrificial
25 balancing fraction is added so as to provide the model
with at least one additional element of symmetry,
preferably an additional plane of symmetry. The
distribution of residual stresses is thus better
distributed since it too benefits from an additional
30 element of symmetry. Artificially restoring symmetry in
this way in the meaning of the description thus makes it
possible in striking manner to reduce the occurrence and
the magnitude of defects in the rough part.
In certain implementations, the sacrificial
35 balancing fraction is located in such a manner that the
resulting sacrificial portion is situated in a region
that is symmetrical to the region of the asymmetrical
portion relative to a plane passing through the center of
gravity of the rough part.
In certain implementations, the sacrificial
balancing fraction is configured so that the sacrificial
5 portion of the rough part is symmetrical to the
asymmetrical portion relative to a plane, this plane
being a plane of symmetry of the rough part. Such a
plane of symmetry is particularly easy to put into place
with the help of the CAD software while working on the
10 model of the part.
In certain implementations, the step of modifying
the model includes a step of defining an equilibrium
plane parallel to the construction direction and
corresponding to a plane of symmetry that the model would
15 have if it did not have its asymmetrical portion
corresponding to the asymmetrical portion of the part.
This step makes it easy to identify a candidate plane of
symmetry for the rough part.
In certain implementations, the step of modifying
20 the model includes a balancing step during which a
balancing segment is added to the model in each layer
perpendicular to the construction direction, the
balancing segment restoring symmetry to the layer under
consideration of the model relative to the equilibrium
25 plane. In this way, it is possible to ensure that masses
on either side of the equilibrium plane are uniform layer
by layer along the construction direction: each layer
that is made by additive fabrication thus has a
symmetrical distribution of mass, thereby minimizing
30 deformation.
In certain implementations, said additive
fabrication technique is a method of fabrication by
selective melting or selective sintering of beds of
powder.
35 In certain implementations, said method of
fabrication by selective melting or selective sintering
of beds of powder makes use of a laser beam.
In other implementations, said method of fabrication
by selective melting or selective sintering of beds of
powder uses an electron beam.
In other modes of implementation, said additive
5 fabrication technique is a method of fabrication by
powder projection.
In certain implementations, during the orientation
step, the model is oriented in such a manner as to
minimize the number of fabrication supports and/or to
10 minimize their sizes. Such fabrication supports are
necessary in particular when a layer of the part projects
laterally from the support formed by the immediately
underlying layer of the part. In this way, it is
possible to limit the number of machining operations that
15 need to be performed on the rough part in order to obtain
the part: this also serves to save powder. In addition,
this serves to limit the impact of the roughness that
results from the method of using layers.
In certain implementations, during the orientation
20 step, the model is oriented in such a manner as to
minimize the height of the part in the construction
direction. This serves to minimize the number of layers
and thus the quantity of powder used and also the time
required for fabrication. In addition, any risk of
25 deformation is also reduced and the resulting surface
state is more uniform.
In certain implementations, the part to be
fabricated is a blading part having a leading edge, a
trailing edge, and an airfoil.
3 0 In certain implementations, the digital model of the
blading part is oriented in such a manner that its
leading edge or its trailing edge faces towards the
construction table. In this way, it is possible to
minimize recourse to fabrication supports.
35 In certain implementations, the equilibrium plane of
the blading part intersects the blading part
substantially halfway along its airfoil.
The above-mentioned characteristics and advantages,
and others, appear on reading the following detailed
description of implementations of the proposed method.
This detailed description refers to the accompanying
5 drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
The accompanying drawings are diagrammatic and seek
above all to illustrate the principles of the invention.
10 In the drawings, from one figure to another,
elements (or portions of elements) that are identical are
identified by the same reference signs.
Figure 1 is an overall view of a device for additive
fabrication by selectively melting beds of powder.
15 Figures 2A and 2B are a perspective view and a plan
vieti of an original model for an example part for
fabrication.
Figures 3A and 3B are a perspective view and a plan
view of the model of Figures 2A and 2B for which an
20 equilibrium plane has been defined.
Figure 4A is a perspective view of the model of
Figures 3A and 3B to which a sacrificial balancing
fraction has been added.
Figure 48 is a plan view of a layer of the Figure 4A
25 model.
Figure 5 is a perspective view of the rough part
obtained using the model of Figure 4A.
Figure 6 is a perspective view of the final part
after removing the sacrificial portion from the rough
30 part.
DETAILED DESCRIPTION OF IMPLEMENTATION(S)
In order to make the invention more concrete, an
example method is described in detail below with
35 reference to the accompanying drawings. It should be
recalled that the invention is not limited to this
example.
In the context of this example, the objective is to
fabricate a blade 90 as shown very diagrammatically in
Figure 6. The blade 90 comprises an airfoil 91 having a
leading edge 93 and a trailing edge 94, together with a
5 root 92 that is provided at one end of the airfoil 91.
The airfoil 91 is fine and elongate while the root 92 is
thick and compact: the root 92 thus constitutes an
asymmetrical portion of the blade 90. If the blade 90
did not have its root 92, then it would have a plane of
10 symmetry intersecting the airfoil 91 halfway along.
During a first step, the digital model 50 of the
blade 90 is received in computer assisted design (CAD)
software. As shown in Figures 2A and 2B, the digital
model 50 has a first portion 51 that is fine and
15 elongate, corresponding to the airfoil 91, and a second
portion 52 that is thick and compact, corresponding to
the root 92. Thus, the second portion 52 of the model 50
constitutes an asymmetrical portion of the model 50.
During a second step, the model 50 is oriented
20 relative to the digital image 61 of the fabrication plate
21 and to the construction direction 62 perpendicular to
said image of the plate 61. In order to minimize
recourse to fabrication supports, the model 50 is
oriented in such a manner that its edge 53 corresponding
25 to the leading edge 93 of the blade 90 faces towards the
image of the plate 61. Nevertheless, it is also possible
to orient the model 50 that its edge 54 corresponding to
the trailing edge 94 of the blade 90 is directed towards
the image of the plate 61.
3 0 During a third step shown in Figures 3A and 3B,
attention is given to the residual portion of the model
50 when it does not have its asymmetrical portion 52:
specifically, this is the first portion 51 that
corresponds to the airfoil 91. A plane of symmetry 70
35 for this residual portion 51 is then identified that is
parallel to the construction direction 62 and it is
defined as the equilibrium plane of the model 50.
Specifically, the equilibrium plane 70 intersects the
first portion 51 of the model 50 corresponding to the
airfoil 91 of the blade 90 halfway along the airfoil in
the airfoil height direction.
5 On this topic, it may be observed that the residual
portion 51 of the model 50 presents a second plane of
symmetry 71 that is longitudinal and likewise parallel to
the construction direction 62: nevertheless, this other
plane of symmetry 71 cannot be selected as an equilibrium
10 plane insofar as the original model 50 of the blade 90 is
already symmetrical about this plane 71.
During a fourth step shown in Figure 4B, attention
is given in succession to each fabrication layer along
the construction direction 62 starting from the image of
15 the plate 61, and in each layer, a balancing segment 72a
is added to the model 50 so as to restore symmetry to the
layer relative to the equilibrium plane 70. Thus, in
each layer perpendicular to the construction direction
62, the balancing segment 72a is symmetrical to the
20 corresponding segment 52a of the asymn~etrical portion 52.
Once this operation has been performed for all of
the layers of the model 50, a modified model 50' is
obtained as shown in Figure 4A that includes a
sacrificial balancing fraction 72 made up from the stack
25 of balancing segments 72a. The modified model 50' is
thus noti symmetrical relative to the equilibrium plane
70, the sacrificial balancing fraction 72 being
symmetrical to the asymmetrical portion 52 relative to
the equilibrium plane 70.
30 Under such circumstances, it is possible to launch
fabrication of the rough part 80 using layer-by-layer
additive fabrication on the basis of the modified model
50'. In this example, and as shown in Figure 1, the
method is a method of fabrication by selectively
35 sintering beds of powder. Nevertheless, it could in
analogous manner be a method of fabrication by powder
projection .
A first layer 10a of powder of the desired material,
specifically nickel-based powder, is thus deposited on
the fabrication plate 21.
A first region of said first layer 10a is scanned
5 with the laser beam 31 so as to heat the powder of said
region locally to a temperature higher than the sintering
temperature of the powder, such that the particles of
said powder as melted or sintered in this way and that
are located in said first region then form a first
10 single-piece element 12a.
A second layer lob of powder of said material is
deposited on said first layer of powder 10a.
A second region of said second layer 10b overlapping
said first single-piece element 12a at least in part is
15 scanned by the laser 31 so as to heat the powder in this
second region to a temperature higher than the sintering
temperature of the powder, so that the particles of the
powder as sintered or melted in this way form a second
single-piece element 12b connected to the first single-
20 piece element 12a and overlying it.
The two above steps are then repeated for each new
layer of powder that is to be deposited over a preceding
layer, and until the rough part 80 shown in Figure 5 has
been formed in full.
25 This rough part 80 comprises the desired airfoil 91
and root 92 together with a sacrificial balancing portion
82 that is symmetrical to the root 92 relative to the
plane of symmetry 81 of the rough part 80, which plane
corresponds to the equilibrium plane 70. Ideally, the
30 sacrificial balancing portion 82 thus possesses the same
shape, in reflection, and the same mass as the root 91
constituting the asymmetrical portion of the blade 90.
Because of the presence of this sacrificial
balancing portion 82 that is fabricated at the same time
35 as the remainder of the rough part 80, the configuration
of the residual stresses in the rough part 80 is
distributed symmetrically, and thus in balanced manner,
on either side of the plane of symmetry 81. The rough
part 80 thus does not have any major defects of the kind
that usually results from the deformations caused by
residual stresses.
5 Finally, once the rough part 80 has been obtained,
it suffices to remove the sacrificial balancing portion
82 by machining in order to obtain the desired blade 90.
In certain circumstances, additional machining steps may
be needed prior to obtaining the final part, in
10 particular when fabrication supports are necessary.
The implementations described in the present
description are given by way of non-limiting
illustration, and in the light of this description, a
person skilled in the art can easily modify these
15 implementations, or envisage others, while remaining
within the ambit of the invention.
Furthermore, the various characteristics of these
implementations can be used on their own or they can be
combined with one another. When they are combined, these
20 characteristics may be combined as described above or in
other ways, the invention not being limited to the
specific combinations described in the present
description. In particular, unless specified to the
contrary, a characteristic described with reference to
25 any one implementation may be applied in analogous manner
to any other implementation.

CLAIMS
1. A method of fabricating a part by additive
fabrication, the part (90) to be fabricated possessing an
asymmetrical portion (92), the method being characterized
5 in that it comprises the following steps:
supplying a digital model (50) of a part (90) to be
fabricated;
orienting the model (50) relative to a construction
direction (62) for constructing the part (90);
10 modifying the model (50) by adding a sacrificial
balancing fraction (72) configured so as to balance the
residual stresses that appear in the part (90) while it
is being fabricated;
making a rough part (80) layer by layer using an
15 additive fabrication technique on the basis of the model
(50') as modified in this way, said layers being stacked
in the construction direction (62); and
using a material-removal method to eliminate the
sacrificial portion (82) from the rough part (80) as
20 results from the sacrificial balancing fraction (72) of
the model (501), thereby obtaining said part (90) that is
to be fabricated;
wherein the sacrificial balancing fraction (72) is
configured in such a manner that the sacrificial portion
25 (82) of the rough part (80) possesses mass lying in the
range 70% to 130% of the mass of the asymmetrical portion
(92).
2. A method according to claim 1, characterized in that
30 the sacrificial balancing fraction (72) so configured so
that the sacrificial portion (82) of the rough part (80)
possesses mass 1,ying in the range 90% to 110% of the mass
of the asymmetrical portion (92).
35 3. A method according to claim 2, characterized in that
the sacrificial balancing fraction (72) is added over a
height substantially equivalent to the height of the
asymmetrical portion (92).
4. A method according to claim 2 or claim 3,
5 characterized in that the sacrificial balancing fraction
(72) is configured in such a manner that the resulting
sacrificial portion (82) is constructed opposite from the
asymmetrical portion (92) relative to the rough part
(80).
10
5. A method according to any one of claims 2 to 4,
characterized in that the sacrificial balancing fraction
(72) is added so as to provide the model (50') with at
least one additional element of symmetry, preferably an
15 additional plane of symmetry.
6. A method according to any one of claims 2 to 5,
characterized in that the sacrificial balancing fraction
(72a) is configured in such a manner that the sacrificial
20 portion (82) of the rough part (80) is symmetrical to the
asymmetrical portion (92) relative to a plane (81), said
plane being a plane of symmetry of the rough part (80).
7. A method according to any one of claims 2 to 6,
25 characterized in that the step of modifying the model
(50) includes a step of defining an equilibrium plane
(70) parallel to the construction direction (62) and
corresponding to a plane of symmetry that the model (50)
would have if it did not have its asymmetrical portion
30 (52) corresponding to the asymmetrical portion (92) of
the part (90) .
8. A method according to claim 7, characterized in that
the step of modifying the model (50) includes a balancing
35 step during which a balancing segment (72a) is added to
the model (50) in each layer perpendicular to the
construction direction (62), the balancing segment (72a)
restoring symmetry to the lByer under consideration of
the model (50) relative to the equilibrium plane (70).
9. A method according to any one of claims 1 to 8,
5 characterized in that said additive fabrication technique
is a method of fabrication by selecting melting or
selecting sintering of beds of powder, or indeed a method
of fabrication by powder projection.
10 10. A method according to any one of claims 1 to 9,
characterized in that- during the orientation step, the
model (50) is oriented so as to minimize the number of
fabrication supp-orts and/or so as to minimize their
sizes.
15
11. A method according to claim 7 or claim 8,
characterized in that the part to be fabricated is a
blading part (90) having a leading edge ( g 3 ) , a trailing
edge (94),. and an airfoil (9');
20 in that the digital model (50) of the balding part
(90) is oriented in such a manner that its leading edge
(53) or its trailing edge (54) faces'towards the
construction table (61); and
in that the equilibrium plane (70) of the blading
25 part (90) intersects the blading part substantiallyhalfway
along its airfoil (91).