I ADAPTIVE MACHINING METHOD FOR SMELTED BLADES
The present invention relates to the manufacture of turbine engine components and, in
particular, the finishing of components produced by smelting or forging.
5
More generally, the invention relates to the finishing of blank components, of final
dimensions, which still require final machining, or finishing. In particular it applies to smelted
or forged turbojet engine blades.
10 Turbine blades of turbine engines are generally made using the so-called "lost wax"
technique, which directly produces the required shape for most of the component, without the
need to produce a blank which then has to be machined to create the final shape. However,
this technique still necessitates final machining, which comprises operations on the leading
edge to remove transferred grains, operations on the trailing edge to remove a trailing edge
15 strip and operations on the platform for its connection to the surface of the vane.
The smelting technique is not completely effective for producing very thin faces (or partitions),
such as, for example at the trailing edges of blades and their ventilation air evacuation vents.
In order to overcome this problem, the thin ends of the blades are smelted with a thickened
20 portion, which is then removed by machining at the end of the manufacturing process. The
problem with this method is that it is necessary to calibrate the reference for positioning the
blade in the space used for this machining with that of the vane as it comes from the foundry.
In the prior art the reference used is taken on the root of the blade and the vane is machined
from the reference points taken on this root. This method gives rise to numerous dimensional
25 irregularities and is time-consuming to implement. Because of the possible discrepancy
between the reference taken on the root and that of the vane, there is often a lack of precision
at the connection between the surface of the vane and the machined surface, with the
appearance of projections on the surface of the component.
30 Moreover, a digitally controlled machine tool would ideally be used for these finishing
- . .
operations to avoid dimensional irregularities.
-
The object of the present invention is to overcome these problems by proposing a machining
method for finishing smelted or forged components, whilst limiting dimensional irregularities
35 and guaranteeing a good connection of the smelted or forged components with the machined
I
I parts of the component surface. This type of machining is known as adaptive machining
I because the path of the machine tool adapts to the geometry of the blank component.
The object of the invention is therefore a method for finishing the shape of a turbine engine
5 blade by machining, in which at least one area is produced thicker than the required profile,
the said thickened portion forming a first surface with the surrounding profile, the machining
of the method being aimed at removing this thickened portion, using a suitable tool, to obtain
a third surface, the theoretical profile of the component being, moreover, defined by a second
surface, characterised in that it comprises the following steps to define the position of the bit
10 of the said tool during machining of the said area:
- defining, on the second surface, a grid formed by intersecting lines forming nodes and
delimiting squares, the said grid covering the area to be machined and extending so
i.
that the outer limit of the machining area is completely enclosed within the outer limits
of the grid,
15 - defining the position of each point of contact P between the tool/component for
machining along the theoretical profile, by weighting with reference to the position of
the nodes Ni of the square in which it is located,
- measuring, for each node Ni located outside the outer limit, the delta(Ni) difference
between the first surface at said node and the theoretical position of the node on the
20 second surface, the said difference being measured normal to the second surface,
- calculating, step by step, the differences delta(Nj) for each node Nj located within the
outer limit and selecting the square to which said node belongs and for which the
delta (Ni) or delta (Nj) differences of the other nodes are already known, and
interpolating from the said delta (Ni) or delta (Nj) differences, weighted according to
25 the length of the segments connecting the said nodes,
- defining the delta(P) difference to be applied to each point P to obtain a point of
contact P' between the component and the tool, relative to the reference input into the
machine tool, by the weighted sum of the deltas(Ni) of the nodes of the square to
which the said point P belongs, using the said weighting performed with reference to
30 the nodes Ni.
This method ensures that no projections are created between the machined surface and the
surrounding surface and that these two surfaces join on the same tangential plane.
Preferably weighting is done by assigning weighting coefficients CPi to each point P,
35 corresponding to the weight to be assigned to the nodes of said square so that it is the
barycentre of said nodes assigned said coefficients. Advantageously the lines forming the
grid are straight lines.
The polygons forming the squares are preferably four-sided polygons.
5 In an alternative embodiment the polygons forming the squares are three-sided polygons.
In a particular embodiment at least one of the nodes located outside the outer limit is not
borne by the first surface, the delta(Ni) difference for this point therefore being calculated by
interpolation between the delta(Ni) differences of the adjacent nodes borne by said first
10 surface, weighted according to the length of the segments connecting said nodes.
The following detailed explanatory description of an embodiment of the invention, given by
way of an illustrative and non-limiting example and with reference to the attached schematic
drawings, will provide a better understanding of the invention and better illustrate other
15 applications, details, characteristics and advantages of it.
In these drawings:
- Figure 1 is a front view of a high-pressure turbine engine blade, as a blank casting:
- Figure 2 is a front view of the blade from Figure I, after finishing using the method
20 according to the invention;
- Figure 3 is a perspective schematic view of an area to be machined using the method
according to the invention;
- Figure 4 is a schematic view of the area shown in Figure 3 and its surrounding area;
- Figure 5 is a perspective view of the intrados of a turbine blade machined using the
25 method according to the invention;
- Figure 6 is a schematic view of a section taken from the area shown in Figure 3.
Figure 1 shows a high pressure turbine rotor blade 1 at the end of "lost wax" smelting
operations, seen from the intrados side. It comprises a root 2 and a vane 3 separated from
30 each other by a platform 4. To ensure the strength of thinner parts, thickened areas have
been left during smelting at the trailing edge outlet vents and at the connection between the
vane 3 and the platform 4. A retaining strip 5, which is to be removed, has also been arranged
parallel to the trailing edge to facilitate smelting operations.
Figure 2 shows the same turbine blade after final machining, which consisted of removing the
retaining strip 5 and thinning out the trailing edge 6. On the blade intrados one can see the
machining area 8 which, by thinning out the intrados, reveals the ventilation outlets 7 of the
trailing edge. For reasons of aerodynamic quality it is desirable to have tangential continuity
between the machined surface and the rough smelted surface of the vane intrados at the
margin of this machining area. Lastly, this machining operation finalised the connection
between the vane 3 and the platform 4, in particular in the vicinity of the trailing edge, and the
tip of the blade was planed down.
10 In Figure 3 one can see two outer surfaces and a machining area corresponding to two steps
in the manufacture of the same smelted or forged component, part of which has a thickened
area that is to be removed. A first surface 10 corresponds to the component following
smelting or forging and a second surface 11 corresponds to the theoretical definition as
designed by the digital model of the component to be produced. It is desirable for the surface
15 10 to be as close as possible to the surface 1 1, but without creating tangential discontinuities
at the margin of the machining area 8. The outer limit 12 of the thickened area to be
machined 8 and a grid with straight or curved segments 13, represented by its nodes
referenced N1 to N30, are defined on the second surface 11.
20 A milling tool 20 is shown resting against the second surface 11 at a point P located in one of
the squares of the grid 13. This tool, which is intended to machine the-area 8 so as to thin the
vane intrados, inside the limit 12, until the first surface 10 approaches the second surface 11,
is inclined relative to the surface to be machined to obtain optimum cutting conditions.
25 Figure 4 shows the same second surface 1 I and the machining area 8 with the distribution of
the nodes of the grid 13 drawn on it. This is drawn so that it completely encloses the
machining area 8, even if certain points on its four sides do not physically lie on the second
surface 11. Thus all the points of the grid 13 positioned over the area 8 are located between
two other points of the grid which are outside the machining area.
30
Figure 5 shows the application of a grid onto the intrados of a high pressure turbine blade 1
positioned on a digitally-controlled machine tool, the tool 20 of which is shown in contact with
its trailing edge. Only one row of squares is represented, the vertices, or nodes, of which
enclose the machining area 8. In the case shown one can see that only the nodes referenced
35 N8 to N11 correspond to points located on the surface of the blade intrados, whilst points N1
to N6 of the outer edge and points N7 and N12 of the lateral edges of the grid 13, are located
in a completely physically unsupported area associated with the vane 3.
Figure 6 shows a point P of the surface 11 located in a square of the grid 13, which is
5 identified by its nodes N1, N2, N7 and N8. Its position in this square is defined by the weights
to be assigned to each of the vertices of the square so that the point P is the barycentre of the
four vertices, each of these being assigned the corresponding weight. Thus one obtains a set
of four values (CI, C2, C7 and C8) which unequivocally determine the position of the point P
on the second surface 11, knowing the square in which it is located.
10
We are now going to describe the machining method according to the invention that, starting
from the first surface 10, enables us to obtain a third surface, that is as close as possible to
the second surface 11 and which observes tangential continuity with the part of the blade not
touched by machining.
15
The profile of the blade to be produced, that is to say that of the second surface 11, is defined
by a theoretical digital model, contained for example in a CATlA program file, whereas the
actual component, leaving smelting or forging, has a profile that is thickened at various
points, in accordance with the first surface 10, to take account of potential fragilities or local
20 working difficulties. The theoretical profile associated with the second surface 11 is
represented in the theoretical CATlA digital file by the position of its points and the orientation
, of the vector normal to the second surface 11 at these points.
The first step in the process according to the invention is to define a grid 13 on the second
25 surface 11, surrounding the entire machining area 8, that is to say the peripheral nodes of
said grid all being outside the outer limit 12 of the machining area 8. This grid is a mesh of
squares expressed by nodes N1 to N30, the coordinates of which are integrated in the CATlA
file, in a Cartesian system associated with the machine tool. Similarly, the CATlA file includes
data corresponding to the position of the bit of the tool 20 of the machine that is to perform the
30 finishing operation and the orientation of its axis.
The second step consists of modelling the path to be followed by the tool 20 during machining
of the first surface 10 so that the point of contact between it and the component describes the
second surface 11 in the case where the points outside the machining area 8 are ideally
35 placed, that is to say on the first surface 10. This path is defined firstly by the succession of
positions of a point P which circulates inside the machining area 8 and thereby passes
through each of the squares of the grid 13 and secondly by the required orientation of the tool
axis at each of these points P. For this purpose each point P is defined relative to the four
nodes of the square in which it is located by means of four coefficients, known as weighting
5 coefficients Cpi (i being the reference number of the node in question). Each weighting
coefficient corresponds to the weight to be assigned to the corresponding node in order for
point P to be the barycentre of these four nodes. In other words, the closer the point P is to a
node, the higher the coefficient assigned to this node and, conversely, the nodes farthest
away are assigned a low coefficient. In order for these weighting coefficients to be uniquely
10 defined, they are reduced proportional to each other so that their sum is equal to 1. For
example, if the point P is in the centre of the square, the four coefficients are all equal to 0.25;
if it is close to one of the nodes, as shown in Figure 6, the coefficient Cpi would equal 0.5,
whilst the other three would be 0.35 for Cp2, 0.1 0 for Cp7 and 0.05 for Cp8.
15 The file comprising the weightings of the points P to be scanned by the machine tool and the
corresponding orientations of the tool axis is then converted into a format that can be
understood by the digitally-controlled machine and loaded into its software.
The next step consists of calibrating the digital model input into the CATlA file with the actual
20 surface of the blade, that is to say with the first surface 10. This stage serves to ensure that
the third surface will coincide perfectly with the first surface 10 at the margin of the machining
area 8 and that there is therefore no projection between the part of the vane outside the
machining area 8 and this area. To achieve this the invention makes provision for
measurement by probing the position of the nbdes located at the margin of the machining
25 area 8 on the component itself, in the normal way at each point on the vane surface. For each
of these points this probing produces a differential value, known as a delta, between the
position of the relevant point in the theoretical CATlA file and the real point measured by the
probe. Application of a correction equal to this-delta to the controlled position of the bit of the
tool 20, along a direction normal to the surface, serves to guarantee that, for all the nodes
30 outside the machining area 8, the tool bit will be flush with the surface of the vane without
machining it or standing off from it.
To give an example relating to Figure 3, the deltas are measured for nodes N1 to N6, nodes
N7, N13, N19 and N25 which are located on the right-hand edge of area 8, nodes N6, N12,
N18, N24 and N30 which are located on the left-hand edge and nodes N25 to N30 which are
located on the bottom edge.
The invention then recommends defining the deltas, known as delta(Ni), for all the nodes Ni,
5 based on the deltas measured on the nodes outside the machining area 8. The deltas of the
nodes located inside the machining area are thus defined by interpolating between the deltas
of adjacent nodes for which a delta has already been measured or calculated.
For example, with reference to Figure 3, the delta of node N8 is defined by interpolating
10 between the deltas of nodes N1, N2 and N7, taking account of the relative length of the
segments connecting this node N8 to these other three nodes. The delta of node N9 is then
calculated from the measured values of the deltas of nodes N2 and N3 and the calculated
value of the delta of node N8. By repeating this operation for all the squares where three of
the four nodes have a delta that is already known, one can gradually determine a delta for all
15 the nodes.
In the particular case where one wanted to machine the theoretical profile of the vane, it
would then be sufficient to give a value of zero to the deltas of the points inside the machining
area 8.
20
The next step according to the invention consists of defining a positioning delta for all the
points of contact P' between the component and the tool 20 during finishing. For this, the
calculation of this delta takes account of the weighting coefficients previously calculated for
the point P and the delta(Ni) differences of the nodes of the square in which the point P is
25 located. The delta at point P, that is to say the correction to be applied to the point P of the
second surface 1 1, is defined as being equal to the sum of the values obtained by multiplying
each delta of a node by the weighting coefficient associated with it.
In the example of the point P located in the square formed by the four nodes N1, N2, N7 and
30 N8, the value of delta(P) is equal to Cpl*delta (NI) + Cpz*delta (N2) + Cp7*delta (N7) +
Cp8*delta (N8).
This delta(P), difference which extends along the normal to the second surface 11 at point P,
is then projected over the reference axes of the machine tool to determine the three
components of the correction to be applied to the Cartesian coordinates given to the program
controlling positioning of the tool bit during the finishing operation.
Thus machining is done by a digitally-controlled machine tool for which the bit position is at all
5 times corrected by the value from the deltas of the nodes of the relevant square and
weighting coefficients of the point where the said tool bit is located.
Due to readjustment of the position of the tool bit on the nodes close to the machining area 8
and interpolation from these nodes, it is ensured that the third surface thus obtained will be
10 tangential to the un-machined surface of the intrados and that there will be no projection
between the machining area 8 and the surface of the vane 1 around this area.
The process according to the invention has been described above in a case, as illustrated by
Figure 4, where all the nodes outside the machining area 8 are physically borne by the first
15 surface 10, that is to say the nodes for which it is possible to measure the delta using a probe.
In referring to Figure 5, it can be seen that the grid 13 may comprise nodes that are outside
the machining area 8 but are not borne by the intrados surface of the blade 1. The invention
recommends that these nodes be assigned a delta obtained by linear interpolation between
20 the deltas of the adjacent nodes that are borne by the first surface 10, once again taking
account of the length of the segments connecting the relevant node to the nodes from which
the interpolation is done.
In the example of the grid shown in Figure 5, ;Ay nodes N8 io N 1 1 can be measured by a
25 probe. The delta value assigned to nodes 7 and 12 is that of nodes 8 and 11 respectively, and
that of nodes 1 to 6 is taken to be equal to that of nodes 7 to 12 respectively.
Although the invention has been described using a grid formed by four-sided squares, it can
also be implemented with a grid formed by triangles or any other closed polygon.
30 Furthermore, it has been described by application to a turbine engine blade, but the process
may be applied to any other component and remain within the scope of the invention.

CLAIMS
1. Method for finishing the shape of a turbine engine blade by machining, in which at least
one area (8) is produced thicker than the required profile, the said thickened portion forming a
first surface (10) with the surrounding profile, the machining of the method being aimed at
removing this thickened portion, using a suitable tool (20), to obtain a third surface, the
theoretical profile of the component being moreover defined by a second surface (1 I),
characterised in that it comprises the following steps to define the position of the bit of the
said tool during machining of the said area:
- defining, on the second surface (I I), a grid (13) formed by intersecting lines forming
nodes and delimiting squares, the said grid covering the area to be machined (8) and
extending so that the outer limit (12) of the machining area (8) is completely enclosed
within the outer limits of the grid,
- defining the position of each point of contact (P) between the component and the tool
(20) for machining along the theoretical profile, by weighting with reference to the
position of the nodes (Ni) of the square in which it is located,
- measuring, for each node Ni located outside the outer limit (12), the delta(Ni) between
the first surface (10) at said node and the theoretical position of the node on the
second surface (II), the said difference being measured normal to the second
surface (1 I),
- calculating, step by step, the differences delta(Nj) for each node Nj located within the
outer limit (12), by selecting the square to which said node belongs and for which the
delta(Ni) or delta(Nj) differences of the other nodes are already known, and
interpolating from the said delta(Ni) or delta(Nj) differences, weighted according to the
length of the segments connecting the said nodes,
- defining the delta(P) to be applied to each point P to obtain a point of contact P'
between the component and the tool (20), relative to the reference input into the
machine tool, by the weighted sum of the deltas(Ni) of the nodes of the square to
which the said point P belongs, using the said weighting performed with reference to
the nodes N(i).
2. Method according to claim I , in which weighting is performed by assigning weighting
coefficients (CPi) to each point P, corresponding to the weight to be assigned to the nodes of
said square so that it is the barycentre of said nodes assigned said coefficients.
3. Method according to either claim I or claim 2, in which the lines forming the grid are
straight lines.
4. Method according to any of claims 1 to 3, in which the polygons forming the squares are
5 four-sided polygons.
5. Method according to any of claims I .to 3, in which the polygons forming the squares are
three-sided polygons.
10 6:~ethod according to any of claims 1 to 5, in which at least one of the nodes located outside
the outer limit (12) is not borne by the second surface (II), the delta(Ni) difference for this
point therefore being calculated by interpolation between the delta(Ni) differences of the
adjacent nodes borne by said second surface, weighted according to the length of the
segments connecting said nodes.
Dated this 17/12/2013 NEHA SRIVASTAVA
OF REMFRY & SAGAR
ATTORNEY FOR THE APPLICANTS