The present invention relates to a method for determining the position of a core in an injection mold, in particular a wax injection mold. This method is for the manufacture of parts for turbomachines, such as turbine blades.
Conventionally, the technique of lost-wax casting is first to make a wax pattern, or any other easily removable material subsequently to the part to be produced; This model comprises an inner part forming a ceramic core contained cavities that it is desired to appear within the blading. The wax pattern is then dipped repeatedly in slurries consist of a suspension of ceramic particles for making, by said operations stucco and drying, a shell mold.

We then proceeded to the dewaxing of the shell mold, which is an operation by which the wax or the material constituting the original model is removed from the shell. After this removal, there is obtained a ceramic mold whose cavity reproduces all forms of the blade and which still contains the ceramic core for generating the internal cavities thereof. The mold is then subjected to high temperature heat treatment or "cooking" which gives it the necessary mechanical properties.

The shell mold is then ready for the production of the metal part by casting. After monitoring the internal and external integrity of the shell mold, the next step consists in pouring a molten metal, which comes to occupy the voids between the inner wall of the shell mold and the core, and then to solidify. In the area of ​​lost wax casting, it is presently are several solidification techniques, so several casting techniques, depending on the nature of the alloy and the expected properties of the workpiece resulting from the casting. It may be directionally solidified columnar structure (DS), directionally solidified single crystal structure (SX) or equiaxed solidification (EX).

After casting of the alloy, the shell is broken by a shake-out operation. In a further step, chemically eliminates the ceramic core which remained locked in the blade obtained. The metal blade obtained then undergoes finishing operations that allow to obtain the finished part.

Exemplary embodiments of turbine blades by casting technique to the lost wax are given in FR2875425 and FR2874186 patent applications of the applicant.

To form the wax pattern of the blade tooling is used, wax or injection mold, wherein the core is placed and then injects the liquid wax through a channel provided for this purpose.

Should be placed this core extremely accurately in the injection mold because any shift in positioning will result in non-compliance of the wall thickness of the blade. The metal of the blade being subjected to very high temperatures, these defects would result in lifetimes of dawn greatly reduced. It is therefore appropriate to ensure with high precision the place of the core in the mold. For this purpose, the molds currently used comprise support means of statically core, said support means may include rods whose ends form supporting points for supporting the core in the mold.

If this type of core positioning is effective, it can still pose a number of challenges. In fact, the cores are made in a mold whose imprint corresponds to that of the core. However, a manufacturing defect, however small, the core of the manufacturing mold, a sufficiently precise estimation of the coefficients of shrinkage and repeatability of the baking step following the molding operation, can lead to differences in thicknesses wall on the final part, which while not leading to non-compliance, however, are not desirable. An obvious solution would obviously be to build a new core injection mold that would respect the manufacturing tolerances for the kernel. However, this solution is not desirable since the step of forming a core manufacturing mold is very expensive, both from a financial perspective that the point of view of the time needed for its implementation.

The invention particularly aims to provide a simple, efficient and economical problems of the prior art described above.

To this end, it proposes a method for determining the position of the cores in an injection mold, comprising the steps of: a) Collect k rated cores R 1 ... R i ... R k in a population of nuclei all made from the same theoretical three dimensional model core,

b) establishing a three-dimensional model of each of the cores,

c) reset in space each three-dimensional models with respect to a plurality of support points T ... T q ... T the core into the mold so as to obtain a three-dimensional spatial model V1 recalibrated for each core ,

d) selecting the R rated core rep whose spatial three-dimensional model recalé V1 has the least deviation from the theoretical spatial three-dimensional model,

e) performing a recalibration of the three-dimensional model of the core R rep with the theoretical spatial three-dimensional model taking into account at least one functional face of the theoretical model of the core so as to obtain a three-dimensional spatial model recalé V2 core R rep ,

f) repositioning the support points T q so as to support the ring R rep in the position corresponding to the spatial three-dimensional model recalé V2 core R rep .

According to the invention, a core geometry defect is compensated by repositioning of a representative core relative to the faces

functional the theoretical model. All cores are then positioned in an injection mold in the same manner that is positioned Representative core in a mold. The method is thus particularly advantageous when the (or) fault nuclei of geometry corresponds to a drift of a dimension with respect to a nominal value. The collection of cores k is made randomly.

The term "functional surface" means a core core face to form, prior to assembly of the part, a face of the final geometry of the part. Such functional face is an outer face of the core which allows the shaping of the internal or external faces of the metal part and which has an impact on the aerodynamics and heat the workpiece in operation. In the case of a turbine blade, a functional surface may refer to an outer face of the core forming an internal face of a wall of the core, such as a wall intrados or extrados for example. The internal cavity of the blade may be a cooling cavity of the blade.

The term "three-dimensional model" in reference to a ring can be interpreted as a set of digital data to a digital three-dimensional reconstruction of the core, for example by a geometric mesh.

The term "space" refers to three-dimensional model positioned in space.

The term "recalibrated" refers to a spatial three-dimensional model that has been positioned or repositioned in space.

According to another feature, each three-dimensional model may be obtained from a three-dimensional survey of the outer surface of the core, for example obtained from a non-contact measurement that can be achieved by optical triangulation. In such a configuration, a central projector illuminates a room with a fringe pattern that are observed by two CCD cameras. We deduce a polygon mesh of the outer surface of each of the cores.

In a particular embodiment of the invention, step d) may comprise the steps of:

- select n points denoted P 1 ... P i ... P n on at least one of the functional surfaces of the theoretical model of the nucleus,

- Select the nucleus R rep with n points of the spatial three-dimensional model recalé V1 has the least deviation with the same n points of the space theoretical model.

In this alternative embodiment, determining the representative core is thus performed by measuring the differences on a functional face after resetting of the bearing points. It is indeed interesting to measurements of deviations from at least one functional surface since it is a face having a direct impact on a corresponding face of the final part.

Also, step d) may comprise the steps of, for each kernel R t :

i. Z¾ determine the difference between each point Pj of the theoretical model and the model V1,

ii. calculating the average M j (E¾) = ^ _Σ = 1 E¾

iii. calculate Δ¾ E¾ = - M, - (E¾)

iv. calculating for each R ring, S = Δ Σ ^ ^ 2

v. consider the ring R t which is assigned the lowest S value as the representative core R rep of the core population.

Preferably, the method comprises a checking step, inserted between steps e) and f) of verifying that the spatial pattern recalé V2 core R rep is better positioned than the spatial model recalé V1 R core rep .

If the registration V2 proves worse than the registration V1, then it should start the registration V2 on a number of faces

functional than the number of functional surfaces used previously.

The verification step comprises the following steps:

i. determining the deviation E? ep j between each point P j of the theoretical three-dimensional spatial model and the spatial three-dimensional model recalé V2 core R ref with respect to functional faces,

ii. calculate S? ep = Σ = 1 £ 2 >

iii. compare S ep S ep so check that S ep is less

The gap E¾ and / or the deviation E? ep j can be determined according to the normal to the theoretical spatial three-dimensional model at the point P j .

The repositioning of points of support of step f) can be performed as follows for each of the points T q support:

- projecting a point T q , along the normal to the theoretical three-dimensional model space passing through the contact point of the fulcrum T q with the theoretical three-dimensional model space, the spatial three-dimensional model recalé V2, so as to obtain a point T q - modify the bearing points into the mold so as to be brought at the points T q .

In a practical embodiment of the invention, k is greater than or equal to five and / or / is greater than or equal to six and / or n is greater than or equal to three. In practice, n is a function of the curvature and tolerance of functional face considered. The greater the curvature, the smaller n is large. Thus, the minimum number n is three, which corresponds to the minimum number of points for positioning a isostatically plane in space.

In the context of the manufacture of a turbomachine component, the injection mold is a wax injection mold. The core may be a core turbine blade for example.

The invention will be better understood and other details, advantages and features of the invention appear on reading the following description given by way of nonlimiting example, with reference to Figure 1 showing the main steps of the method of 'invention.

In a first step a) of the process are selected k nuclei rated R R ... T ... R k in a population of nuclei, all made from the same theoretical model three-dimensional core. The term "population" here refers to a set of nuclei whose number can be determinate or indeterminate.

In a second step b) of the method, a three-dimensional survey of the outer surface of each of the cores is obtained from a non-contact measurement which may be an optical measurement, for example by optical triangulation as discussed above. Of course, other measurement methods may be used without departing from the object of the invention. For example, another method may be to use a sensor device or measuring machine (CMM say) more accurate but much slower. The three-dimensional survey makes it possible to establish a three-dimensional model of each of the cores, that is to say a numerical model comprising a set of surface points of coordinates of a core, allowing relative positioning of points.

In a third step c), the method comprises a step of positioning in space of each of the three-dimensional models with respect to / fulcrums T ... T q ... T l of the core in the mold so as to obtain a three-dimensional model space V 1 for each kernel. This positioning thus consists of a spatial adjustment relative to the support points.

In practice, this recalibration can be done by performing a minimization of the difference between the theoretical model and three-dimensional space the spatial three-dimensional model recalé V1 of each of the cores, at points T q . The minimization can be performed according to the least squares method.

The method further includes, in a fourth step, to select the R rated core rep whose spatial three-dimensional model recalé V1 has the least deviation from the average differences calculated between the actual model and the theoretical three-dimensional spatial model. This step is performed on n points P j rated P 1 ... P j ... P n from at least one of the functional faces of the theoretical model of the theoretical ring. Preferably, the n points are distributed over a maximum number of functional surfaces. Preferably, the n points are distributed over the selected functional surfaces and selecting a number of points per side depending on the curvature and the tolerance applied to the face in question.

This kernel selection step representative of the k nuclei group is carried out by performing the following steps:

i. Z¾ determine the difference between each point P j of the theoretical model and the model V1, along the normal to the theoretical model passing through the point P j ,

ii. calculating the average M j (E¾) = ^ _Σ = 1 E¾

iii. Answering Δ¾ E¾ = - Μ (Ε¾)

iv. calculating for each R ring, S = Δ Σ ^ ^ 2

v. consider the core R t which is assigned the value If the lower as the representative core R rep of the core population.

In order to determine the new kernel desired position Representative R rep into the mold, the next step is to carry out, in a fifth step, a second reset of the three-dimensional model taking into account at least one functional face of the theoretical model of the core so to obtain a spatial three-dimensional model recalé kernel V2 R rep ■

Unlike V1 registration made for each of the cores, the core of the resetting Representative R rep is performed only over at least one of the functional surfaces and does not take into account the points of support T q . This is to allow a representative core repositioning R rep to minimize the shape differences between the part obtained from representative core and a theoretical part after the theoretical core, the stress points support T q being eliminated.

Before proceeding to the sixth step, that is to say, in step f), a prior step registration of the verification of the spatial three-dimensional model recalé kernel V2 R rep is performed. This verification step comprises the following steps:

i. determining the deviation E? ep between each point P j of the theoretical three-dimensional spatial model and the spatial three-dimensional model recalé V2 core R ref with respect to functional faces, this gap being measured along the normal passing through the point P j of the theoretical three-dimensional model space,

ii. calculer S; ep = D] = x R ^ 2 ,

iii. compare S ep S ep so check that S ep is less

When S? ep is greater than S £ ep several situations arise. If the registration of the three-dimensional model of the core R rep was performed on one functional surface, then it is necessary to establish that the spatial three-dimensional model recalé V1 R core rep is preferable since it demonstrates that the resetting of the cores on the functional faces do not make for a better positioning of the nucleus. If the registration of the three-dimensional model is carried out on a plurality of functional surfaces, that is to say functional faces F, F> 2, then performs the resetting step e) F-1 functional faces then determines if the new registration V2 of spatial three-dimensional model R rep is better than V1 core recalibration R rep by comparing S 2 ep to ^ ep .

In a complementary approach, it would be possible to perform a classification of the functional faces in at least two groups, a first functional group of main faces and a second group of second functional faces. The main functional faces are faces where the production tolerances are lower than for the second functional faces such that the registration made in step e) may preferably be carried out on the main functional faces. Thus, if the recalibration step e) is to be performed again, then it is best to remove the constraint of the registration against a secondary functional face. Finally, you should check that the discrepancies in secondary functional sides, between the theoretical spatial three-dimensional model and the new V2 dimensional model does not exceed the allowable manufacturing tolerances.

The sixth step f) comprises repositioning the support points T q so as to support the ring R rep in the position corresponding to the spatial three-dimensional model recalé V2 core R rep .

This repositioning is done by performing the following steps:

- projecting a point T q , along the normal to the theoretical three-dimensional model space passing through the contact point of the fulcrum T q with the theoretical three-dimensional model space, the spatial three-dimensional model recalé V2, so as to obtain a point T q ,

- modify the bearing points into the mold so as to be brought at the points T q .

In practice, to perform the above second sub-step, one proceeds to a determination of the distance between each pair of points T q and Tq, giving / distances. These distances correspond to the position corrections to be applied to the ends of core support rods.

CLAIMS
A method of determining the position of the cores in an injection mold, comprising the steps of:
a) Take k rated cores R 1 ... R ... R k in a population of all cores made from the same theoretical three dimensional model core,

b) establishing a three-dimensional model of each of the cores,

c) reset in space each three-dimensional models with respect to / fulcrums T ... T q ... T the core into the mold so as to obtain a three-dimensional spatial model V1 recalibrated for each core,

d) selecting the R rated core rep whose spatial three-dimensional model recalé V1 has the least deviation from the theoretical spatial three-dimensional model,

e) performing a recalibration of the three-dimensional model of the core R rep with the theoretical spatial three-dimensional model taking into account at least one functional face of the theoretical model of the core so as to obtain a three-dimensional spatial model recalé V2 core R rep ,

f) repositioning the support points T q so as to support the ring R rep in the spatial position corresponding to the spatial three-dimensional model recalé V2 core R rep .

The method of claim 1, wherein each three-dimensional model is obtained from a three-dimensional survey of the outer surface of the core, for example obtained from a measurement without contact.

A method according to claim 1 or 2, wherein step d) comprises the steps of:

- select n points denoted P ... P j ... P n on at least one of the functional surfaces of the theoretical model of the nucleus,

- Select the nucleus R rep with n points of the spatial three-dimensional model recalé V1 has the least deviation with the same n points of the space theoretical model.

The method of claim 3, wherein step d) comprises the following steps, for each kernel R t :

i. Z¾ determine the difference between each point P j of the theoretical model and the model V1,

ii. calculating the average M j (E¾) = ^ _Σ = 1 E¾

iii. calculate Δ¾ E¾ = - M, - (E¾)

iv. calculating for each R ring, S = Δ Σ ^ ^ 2

v. consider the ring R t which is assigned the lowest S value as the representative core R rep of the core population.

Method according to one of the preceding claims, comprising a verification step, inserted between steps e) and f) of verifying that the readjusted V2 spatial model of the core R rep is better positioned than the spatial model recalé V1 R core rep .

The method of claim 5 when dependent on claim 4, wherein the verification step comprises the following steps: i. determining the deviation E? ep j between each point P j of the theoretical three-dimensional spatial model and the spatial three-dimensional model recalé V2 core R ref with respect to functional faces,

ii. calculer S; ep = D] = x R 2 ,

iii. compare S ep S ^ ep so check that S ep is less

7. A method according to claim 4 or claim 6, wherein the E¾ gap and / or the deviation E? ep j are determined according to the normal to the theoretical spatial three-dimensional model at the point P j .

8. A method according to one of claims 1 to 7, wherein step f) comprises the following steps for each of the points T q support:

- projecting a point T q , along the normal to the theoretical three-dimensional model space passing through the contact point of the fulcrum T q with the theoretical three-dimensional model space, the spatial three-dimensional model recalé V2, so as to obtain a point T q ,

- modify the bearing points into the mold so as to be brought at the points T q .

9. A method according to one of claims 1 to 8, wherein k is greater or equal to five and / or / is greater than or equal to six.

10. A method according to one of claims 1 to 9 when dependent on claim 3, wherein n is greater than or equal to three.

January 1. Method according to one of the preceding claims, wherein the injection mold is a wax injection mold.