FR3037168A1 - METHOD FOR SIMULATION OF AUB DISTRIBUTION ON A TURBOMACHINE DISK - Google Patents

Abstract

A method of blade distribution simulation (P) on a turbomachine disk (D) is provided, the method comprising the steps of: providing a plurality of blade configurations, each blade configuration comprising minus one balancing parameter measurement (R, T, A) measured on a respective blade (P); • defining a plurality of bladed wheel splits, each splitting associating the blade configurations provided with positions on the blade; disk (D), • searching and selecting (106) a bladed wheel distribution satisfying a criterion defined according to a predetermined cost function dependent on the balancing parameter measurements, the method being characterized by the fact that the search and the selection are conducted by successive iterations by means of a Tabou algorithm.

Description

FIELD OF THE INVENTION The invention relates to a method of simulation of blade distribution around a disk so as to form a turbomachine bladed wheel. STATE OF THE ART Blades intended to be mounted around a turbomachine disk are subject to dispersions during their manufacture. These dispersions can be measured cold on each blade manufactured, that is to say before mounting on the disk and rotating the bladed wheel obtained at the end of this assembly. These dispersions can also be measured hot on the various blades, that is to say during the rotation of the bladed wheel. Such dispersions may render the bladed wheel unbalanced. The consequences of such imbalance are multiple: vibration of the entire engine, and / or generation of a significant acoustic noise during the rotation of the bladed wheel. To distribute vanes in a balanced manner around a turbomachine disk, it has been proposed to implement an algorithm 20 configured to search, in a set of possible blade distributions, a distribution optimizing a predetermined cost function. For n blades to be mounted around a disk, such a search is performed in a set of (n-1)! / 2 blade distributions, considering the identical wheel by rotation and symmetry. Such a search therefore requires a large number of calculations if the number n is high. By way of example, for n = 18, which is a conventional number of blades in a turbomachine bladed wheel, the set of selectable distributions comprises more than 2.1014 distributions. SUMMARY OF THE INVENTION An object of the invention is to distribute blades around a turbomachine disk in a balanced manner, with a limited calculation load. In order to achieve this object, the invention proposes a blade distribution simulation method on a turbomachine disk, the method comprising the steps of: providing a plurality of blade configurations, each blade configuration comprising at least one balancing parameter measurement measured on a respective blade, 10 - defining a plurality of bladed wheel splits, each splitting associating the blade configurations supplied with positions on the disc, - searching and selection of a bladed wheel distribution satisfying a criterion defined according to a predetermined cost function depending on the balancing parameter measurements, the search and the selection being conducted by successive iterations by means of a Tabou type algorithm. The Tabou algorithm makes it possible to explore the space defined by the set of possible distributions in a fast manner. Indeed, during each iteration of the search and selection, some distributions that have already been tested previously are not processed again. As will be seen in the following, a Tabou-type algorithm used during your search makes it possible to reduce the number of blade distributions tested during the search. The invention may also be supplemented by the following features, taken alone or in any of their technically possible combinations. A current iteration of the search and selection may include steps of: generating at least one neighboring distribution from a reference distribution, each neighboring distribution satisfying a predetermined resemblance criterion with the distribution of reference, - selecting, from neighboring distributions, a neighboring distribution optimizing the cost function as a new reference distribution, in which at least one bladed wheel distribution having participated in at least one election during a previous iteration n is not treated as a neighboring distribution during the current iteration. At least one neighboring distribution may result from one or two exchanges of blade positions between blade configurations of the reference distribution. One of the positions exchanged in the context of one of the two exchanges and one of the positions exchanged in the context of the other of the two exchanges may be diametrically opposed. The neighboring distribution may result from one or two circular blade configuration permutations applied to the reference distribution, each circular permutation being performed on a subset of blade configurations associated with consecutive positions around the disk in the reference distribution. Each of the two circular permutations applied may be representative of a blade configuration displacement from an initial position to a destination position, the two initial positions being diametrically opposed in the reference distribution, and the two end positions being diametrically opposed. in the neighboring distribution obtained. The similarity criterion used for the generation of the neighboring distribution is selected by an equiprobable draw between several predetermined resemblance criteria. The method may comprise, for the reference distribution and for each adjacent distribution, the calculation of at least one unbalance from balancing parameter measurements of the same type contained in the blade configurations of the distribution. , and from the positions associated with the blade configurations in the distribution, the cost function depends on the unbalance. The cost function may depend on several unbalances, each unbalance relating to a respective balancing parameter type of measurement, and in which at least one of the unbalances is thresholded by applying the cost function. . The balancing parameter measurements are at least one of the following types: blade angle, axial blade moment, radial blade moment, and tangential blade moment. It is also proposed a method of manufacturing a turbomachine bladed wheel from blades and a disc, the method comprising the steps of implementing a simulation method, and assembling blades on the disk according to the bladed wheel distribution selected by the simulation method. There is further provided a computer program product comprising program code instructions for performing the steps of the above simulation method, when this program is executed by a computing unit. DESCRIPTION OF THE FIGURES Other features, objects and advantages of the invention will be apparent from the following description, which is purely illustrative and not restrictive, and should be deleted with reference to the accompanying drawings, in which: FIG. bladed wheel for turbomachine. FIG. 2 partially represents a blade for a turbomachine. FIGS. 2a, 2b and 2c detail certain measurable parameters on a turbomachine blade. Figure 3 schematically shows a blade distribution device according to one embodiment of the invention. Figure 4 is a flow diagram of steps of a blade distribution method according to one embodiment of the invention. Fig. 5 is a flowchart detailing a step shown in Fig. 4. Fig. 6 schematically shows a reference bladed wheel distribution and adjacent distributions of this reference distribution, generated during a blast distribution process. dawn according to one embodiment of the invention. Fig. 7 shows a cost function according to one embodiment of the invention. In all the figures, similar elements bear identical references. DETAILED DESCRIPTION OF THE INVENTION Referring to Figure 1, a turbomachine bladed wheel 20 comprises a disk D and a plurality of blades P distributed on the disk about its axis. FIG. 2 illustrates a blade P intended to be assembled on the disk D. The blade P has in particular the following intrinsic characteristics: an axial momentum A, a tangential moment T, a radial moment R. The blade P also has a wedging angle. With reference to FIG. 2a, the wedging angle α can be described as the angle between the blade of the blade P and an axis of rotation X of the disk, also called "motor axis" in what follows. With reference to FIGS. 2b and 2c, blade P has a center of gravity. A static moment is the product of the mass of the dawn P and of one of the coordinates (DXcdg, DYcdg, DZcdg) of the center of gravity G of the dawn P taken in a direct orthonormal coordinate system (0, Ux , Uy, Uz) with Ux being the motor axis, and 0 being the center of the reference on the motor axis. For the radial moment R, the coordinate multiplied by the mass of the blade is that of the center of gravity of the blade taken on the axis U. For the axial moment A, the coordinate multiplied with the mass of the blade is the center of gravity of the dawn taken on the axis U. For the tangential moment T, the coordinate multiplied at the mass of the blade is that of the center of gravity of the blade taken on the axis Us, .

With reference to FIG. 3, a device 1 for distributing a plurality of blades P on the disk D to form a turbomachine wheel, comprises a storage memory 2, a calculation unit 4, a buffer memory 6 and a human interface 8. The storage memory 2 is configured to store a data base containing data representative of the blades (for example the moments and cold setting, the mass of the blade, the angle of registration of the blade). dawn, the height of the blade, in a nonlimiting manner) and positions of these blades on a disk D. The calculation unit 4 is configured to implement a Tabou type search algorithm, which will be detailed here. afterwards, from the data stored in the memory 2. The calculation unit 4 has read and write access to your storage memory 2. The buffer memory 6 is also accessible for reading and writing by the storage unit. calculation 4. This buffer memory 6 is configured 25 in. To store temporary data calculated by the calculation unit 4. Alternatively, the memories 2 and 6 form a single memory. The human-machine interface 8 comprises for example a display screen and input means such as a keyboard manipulable by an operator. The human-machine interface 8 offers the possibility to this operator of modifying the execution parameters of the algorithm implemented by the calculation unit 4. The following will be described with reference to FIG. In this invention, it is assumed that n individual P vanes have previously been fabricated, as well as the disk D, comprising n free slots for vane, distributed about its axis. In a preliminary step 100, one or more balancing parameter measurements are measured on each of the n blades. In the following, it is considered that a balancing parameter is a physical quantity characterizing an individual blade. It may be envisaged as a measure of balancing parameters at least one of the following types of measurements: calibration angle measurement, axial moment measurement, tangential moment measurement, radial moment measurement, by means of measures known to those skilled in the art (for example using scales).

The measurement values are stored by the memory 2 in a database, during a storage step 102. The database is for example organized in the form of a table with several entries, each entry relating to one of the four considered. Each entry comprises one or more measurements made on the same blade, the assembly formed by these measurements being in the following called "blade configuration". In addition to these measurements, an entry may include a single blade identifier and / or a serial number. SN Angle (°) Radial (inch Axial (inch Tangential (inch g) Posi_Instrum Fixed? G) g) KLJLKAER -0.22 12345.-2000 507 1 Yes FHFGHFGHDFH -0.31 12500 -2500 420 5 No FHGFH -0.250 13000 -1500 560 12 No HGHGGH -0.02 12000 -1800 670 6 No GHFG -0.101 12700 -1702 450 2 Yes FHGF -0.200 11800 -1504 590 10 No FHGF -0.500 13500 -1203 360 17 No ... .... .. ... ... ... ...

In the table above, the rows represent respective entries in the database. The columns of the table contain, from left to right: a unique SN serial number identifying a blade, a calibration angle value (relative to a reference angle) measured on the blade, a value of radial moment measured on the blade, an axial moment value measured on the blade, a tangential moment value measured on the blade, and information indicating whether the blade must be positioned at an imposed or fixed position, or that this dawn can instead be freely placed at any of the n 10 predetermined positions. Also initiated during step 100 in the database, n position information, each information defining the position around the axis of the disk D of a respective blade location. In the following, this position information will simply be called "dawn positions", or more simply positions. A blade position can be expressed as an angle, or as an index i varying from 1 to n in a reference linked to the bladed wheel. In the following, it will be considered that your positions are indices from which corresponding angles in the reference of the bladed wheel can be determined. At this stage of the process, the n stored blade positions and the n blade configurations are not yet associated. In a search and select step 106, the computing unit 4 produces a plurality of blade distributions on the bladed wheel, and searches for and selects one of those bladed wheel arrays that optimizes a predetermined cost function. This cost function is representative of a balancing criterion and depends on the balancing parameters considered. A bladed wheel distribution is defined as a n-tuple of associations, each association associating one of the n blade configurations and one of the n blade positions. (n-1)! 12 different blade distributions 3037168 9 can thus be theoretically generated during step 106. As will be seen in the following, a Tabou algorithm used during your research makes sure that decrease the number of blade distributions tested during the search.

In an assembly step 108, the individual blades are assembled on the disk D, in its n locations, according to the distribution selected by the device 1 during the previous step 106 search and selection.

10 Search and selection by means of a Tabou algorithm We will now detail in connection with FIG. 5 the step of searching and selecting 106 the blade distribution to be used for the assembly. As a preliminary, two lists are allocated in the buffer memory 6: a list of candidates, and a list called "taboo". The search is implemented by means of a Tabou algorithm, including successive iterations. A current iteration includes the following substeps. The computing unit determines a reference bladed wheel distribution. In the particular case of the first iteration, the reference distribution can for example be generated randomly and stored in the buffer memory 6 and / or in the database of your memory 2. As a variant, the first reference distribution is defined by The operator during a parameterization step 104, via the man / machine interface 8. In any iteration except the first one, the reference distribution was generated during a previous iteration. The calculation unit 4 generates, from the reference position, one or more distributions that are close to the reference distribution, according to a predetermined similarity criterion. This predetermined resemblance criterion corresponds to a bijective mathematical function of permutation (composition of transpositions) between two distributions. In other words, a distribution that respects the predetermined resemblance criterion is a similar distribution.

The neighboring distributions are stored in the list of candidates by the buffer memory 6, after emptying this list of candidates. The calculation unit 4 calculates the cost of each neighboring configuration, by applying the cost function to the balancing parameter measurements contained in the neighboring configuration. More specifically, the computing unit 4 calculates, for each adjacent distribution, at least one unbalance relating to one or more types of balancing parameter measurement; the cost function is applied to the unbalance of each adjacent distribution.

An imbalance corresponds to the eccentricity of the center of gravity of the entire wheel. This center of gravity is calculated by vector summation of all the moments of each of the vanes P. The calculation unit 4 then selects from the list of candidates a distribution that optimizes the cost function the most. (that is, the distribution that offers the best balance among the neighboring distributions). Depending on the type of function considered, the selected distribution will be the distribution presenting a minimum or maximum cost. The neighboring distribution selected is then considered as a new reference distribution. The neighboring distributions generated are furthermore stored in the "taboo" list by the buffer memory 6. These steps are then repeated in a next iteration, on the basis of this new reference distribution, and so on in subsequent iterations. , each new iteration considering a new reference distribution.

The algorithm ends when a predetermined output criterion is checked, for example, a value of the cost function is reached by a reference distribution and / or a number of iterations is reached. In both cases, the optimal distribution chosen is the last reference distribution considered by the Tabou algorithm implemented. A peculiarity of the Tabou-type algorithm used is that, during any iteration except the first, it is ensured that your blob distributions having participated in at least one election during a previous iteration are not not treated as neighboring distributions during a current iteration. This exclusion principle can be implemented in a current iteration by comparison, after the step of generating neighboring distributions, between the content of the "taboo" list (updated during the immediately preceding iteration). and the contents of the list of candidates. Any distribution found in both lists is then deleted from the list of candidates. The exclusion can be limited to the neighborhoods generated during the immediately preceding iteration, in order to limit the memory consumed by the taboo list. As a variant, the exclusion can relate to the neighboring distributions processed during the last K iterations preceding the current iteration. This variant makes it possible to explore the space of the possible bladed wheel distributions more quickly.

This exclusion operation may furthermore lead to increasing the value of the function (in a minimization problem): this is the case when all the neighboring distributions have a higher cost. than your reference distribution, which then forms a local minimum. The risk, however, is that at the next stage, we fall back into the local minimum that we have just escaped. This is why the heuristic must have memory: the mechanism consists of forbidding (hence the name taboo) to return to the last positions explored. The size of the Tabou list may be an executing parameter of the algorithm, which may be changed by an operator via the man / machine interface 8. Generation of Neighboring Distributions In one embodiment, four criteria of similarities are used to determine / generate neighboring distributions.

A neighboring distribution defined according to a first similarity criterion is a bladed wheel distribution obtained after exchange of the positions associated with two blade configurations in the reference distribution. FIG. 6 shows an exemplary reference distribution REF and a neighboring distribution V1 generated according to this first similarity criterion. Positions 2 and 3 of the REF distribution are exchanged to obtain your V1 distribution. A similar distribution defined according to a second similarity criterion is a bladed wheel distribution obtained after two exchanges of the positions associated with two blade configurations in the reference distribution. Two pairs of positions are exchanged, for example distinct. On the other hand, one of the traded positions on one of the two trades, and one of the traded positions in the other trades, may be diametrically opposed. By way of example, positions 2 and 3 of the distribution REF are exchanged, likewise the positions 12 and 11, to obtain the distribution V2, for a number of blades equal to 18. The maximum number of exchanges allowed for the second criterion can be increased to a value greater than 2.

The positions defining an exchange may be adjacent (in the application of the first criterion or the second criterion). A similar distribution defined according to a third similarity criterion is a bladed wheel distribution resulting from a circular permutation of blade configurations applied to a subset of blade configurations associated with consecutive positions around the disk in the reference distribution. Such permutation virtually represents the movement of a blade of a bladed wheel defined according to the reference distribution from a first position to a second position, and an offset of the blades interposed between the first position and the second position. By way of example, the distribution V3 illustrated in FIG. 6 is obtained by a circular permutation applied to the set [2, ... 13] of the reference configuration REF. This circular permutation 15 represents the displacement of a blade virtually positioned at the position 13, at the position 2. A neighboring distribution defined according to a fourth similarity criterion is a bladed wheel distribution resulting from two circular permutations according to the third criterion. .

Each of the two circular permutations applied being representative of a blade configuration displacement from an initial position to a destination position, it can be provided that the two initial positions are diametrically opposed in the reference distribution, and the two positions. diametrically opposite ends in the neighboring distribution obtained. By way of example, the distribution V4 illustrated in FIG. 6 is obtained according to the fourth criterion. The maximum number of circular permutations allowed to meet the fourth criterion can be increased to a value greater than 2.

In the step of generating the neighbors, one of the four criteria described above is selectively used after a random draw between the four criteria. The four different resemblance criteria set out above define neighborhoods different from the reference distribution. The fact of using different similarity criteria (and therefore of defining different neighborhoods for a reference distribution) is advantageous in the case where the cost function depends on several types of balancing parameter.

Indeed, a given neighborhood may result in the election of a relatively balanced distribution from the point of view of a first type of balancing parameter, but relatively unbalanced from a point of view of another type of parameter. balancing. Accordingly, the fact of using these different criteria makes it possible to explore the space of the possible distributions according to different possible strategies. The algorithm can be implemented with predetermined constraints that the blade distribution to be chosen must respect. The fixed or blocked nature of this or that dawn, already discussed above, constitutes an example of constraint. Furthermore, it can be chosen that at least one of the following constraint stress parameters must be less than a predetermined threshold in the distribution: Residual "Radial" unbalance: resulting from the radial moments on the wheel; - Residual "Radial + Tangential" imbalance: resulting from the radial and tangential moments on the wheel; - "Radial imbalance generated by the Axial torque in the plane of the residual bearing 1": resultant of the radial components generated by the axial torque in the plane of the bearing 1 on your wheel; - Radial + Radial + Radial Balun generated by the Axial torque in a plane of the residual bearing 1: resultant radial, tangential and radial moments related to the axial torque on the entire wheel; 5 - Delta moment Radial max vanes face to face: the maximum of the differences of moment between 2 vanes face to face raised on the wheel; - Delta moment Axial max vanes face to face: the maximum of the differences of moment between 2 vanes face to face raised on the wheel; - Delta moment Tangential max vanes face to face: the maximum of the 10 differences of moment between 2 vanes face to face raised on the wheel; - BalQurd "Calibration" residual: by assimilating the wheel blade spacings to radial vectors, we calculate the radial resultant wedges on your wheel (calibration measured at a certain cutting height); 15 - Delta Calibration max dawn to dawn: te maximum of the differences in timing raised between 2 successive blades on the whole wheel (calibration measured at a certain height of cut). Finally, four families of constraints can be examined: - The unbalance: on each position of your wheel, a characteristic 20 dawn (angle, moment, mass, ...) is represented by a vector. Unbalance is the length of displacement generated by the addition of all vectors associated with one or more blade characteristics. - The blade characteristic deltas face to face: on some 25 wheel configurations, in front of each blade is a blade. In this document, a pair of vanes, will designate 2 vanes which are mounted opposite each other. To constrain pairs is to look at the difference in the characteristics of paired vanes. 3037168 16 - Adjacent blade characteristic deltas: the maximum difference in characteristics of a blade with its nearest neighbors is a quantity of interest. - The blocked positions: the optimization algorithm will not test 5 wheel with blocked blades that would not be in their place. For example, "moving" a blade, in the sense of the third criterion, causes many blades to be moved from their initial position in the reference configuration, breaking the initial pairing of the blades in the reference configuration. This has the consequence of being incompatible with the constraint of having vanes (that is to say non-displaceable vanes) or, to a lesser extent, constraints on diametrically opposed pairs of vanes on the wheel. If a certain number of vanes lose their initial position (generally displaced by one cell), their position with respect to their neighbor is little modified, so this neighborhood favors an improvement on the adjacent delta type criteria. ". The table below summarizes the advantages and disadvantages of the different neighborhoods according to the balancing criterion sought.

20 Neighbors fixed algebra Delta criterion adjacent Delta criterion pair unbalance displacement - - + - + 1 permutation ++ - _ + 2 blades displacement - - + + + 1 permutation pair ++ - + + 2 pairs Selection of a neighboring distribution « The imbalance of a given adjacent configuration is calculated from the n equilibrium parameter measurement values of the same type (e.g., n stagger angles) contained in each of the n blade configurations, and from the n positions associated with the n vane configurations of the neighboring distribution. For example, the imbalance of a neighboring configuration relating to the wedge angles can be calculated as the sum of n elementary vectors; each elementary vector of index i ranging from 1 to n is defined by means of the calibration angle measurement and is recorded in the index blade configuration i, and the index position j. The orientation of the calculated sum vector, with respect to the virtual axis of the disk D is representative of a degree of equilibrium of the neighboring configuration, from the point of view of the wedging angles. A collinear sum vector on this axis is representative of a perfectly balanced wheel. Unbalances related to a moment (tangential, axial, or radial) can be calculated in the same way. When the blade configurations contain more than one type of balancing parameter, several unbalances can be computed for each of the neighboring distributions, each unbalance relating to a type of balancing parameter. For example, for each of the neighboring distributions, an unbalance relative to the angles of wedging, an imbalance relative to the axial moments, an imbalance relative to the tangential moments, etc., can be calculated. Other more complex unbalances, representative of a combination of balancing parameter measurements, can also be calculated. Cost function In one embodiment, the cost function depends on the following predetermined values: - ba ba: an axial unbalance reference - C_br: a radial unbalance reference - C_bt: a tangential unbalance reference 30 - C_brt: a Radial + Tangential unbalance reference - C_brta: An unbalance reference Radial + Tangential + Axial 3037168 18 - C_bang: An angular unbalance reference - C_adjacent_ang: An adjacent angle deviation reference - C_paire_a: An axial end pair criterion reference - C_pair_t: a reference of criterion pair moments tangential 5 - C_pair_r: a reference of criterion pair moments radial - C_paire_ang: a reference of criterion pair angle of calibration These values are determined by an expert according to the thresholds of vibration and acoustics that he wants to reach. The cost function also depends on the following parameters: 10 - BA (Wheel): calculated axial unbalance for the bladed wheel distribution Wheel - BR (Wheel): radial unbalance calculated for the bladed wheel distribution Wheel - BT (Wheel): Tangential equilibrium calculated for the bladed wheel distribution Wheel - BRT (Wheel): radial + tangential unbalance calculated for the bladed wheel distribution Wheel - BRTA (Wheel): radial + tangential + axial equilibrium calculated for the bladed wheel distribution Wheel 20 - BAng (Wheel): calculated angular unbalance for the bladed wheel distribution Wheel - Dadja_Ang (Wheel): maximum deviation of adjacent wedging angles found in the wheel distribution. The cost function is constructed using the following terms, called "distances": - E_ba (Wheel) = max (1, BA (Wheel) / C_ba) E_br (Wheel) = max (1, BR (Wheel) / C_br) - E_bt (Wheel) = max (1, BT (Wheel) / C_bt) - E_brt (Wheel) = max (1, BRT (Wheel) / C_brt) 30 - E_brta (Wheel) = max (1, BRTA ( Wheel) / C_brta) - E_bang (Wheel) = max (1, Bang (Wheel) / C_bang) 3037168 19 - E_delta_ang (Wheel) = max (1, Dadja_Ang (Wheel) / C_adjacent_ang) These distances have the form illustrated in FIG. It will be understood that each reference of the abscissa axis defines a balancing criterion. Whatever the distance considered, when your wheel respects the associated criterion, the distance stagnates at 1. The interest of this lower bound, or thresholding, lies in the idea that as long as a criterion is reached the algorithm focuses on the other criteria. The cost function also depends on another parameter Nb_pair (Wheel), called "delta pair pair tangential moment out of 10 criterion radial or tangential", and is defined as follows: Nb_pair (Wheel) n -2 n / 2 (<C_paire_a} X 1 {1 Angi-Ang i + n <C_pair_t}) <C_pair_ang} X 111mx_m T t-1-- n M_Ri-M - R. + C_paire_r} 2 - - 2 X where: - 1 [1 is the indicator function, - M_Ri the radial moment of the blade of index i - M_Ti the tangential moment of the blade of index i 15 - M_Ai the axial moment of the blade of index i - Angi the angle dawn calibration index i - n is the number of blades. The cost function is also constructed using a distance from the pairing criteria (C_paire_xxx, where xxx represents the quantity of interest), of the form: pair (Wheel) The function of cost on the wheel can then be defined as follows: 3037168 20 Distance (Wheel) = _ba (Wheel) Ek L _br (Wheel) x E _bt (Wheel) x _brt (Wheel) x _brta (Wheel) x ibang (Wheel) x e_delta_ang (Wheel)) .11E_pair (Wheel) Or alternatively: Distance (Wheel) = Wnax (E _ba (Wheel) Mli _br (Wheel); IE _bt (Wheel); 11E_brt (Wheel); L _brta (Wheel) x _bang (Wheel)) _delta_ang (Wheel) 5 <E _pair (Wheel) One indicator per family of criteria to be respected, makes it possible not to drown a poorly represented family.

In a particular embodiment, only the references using the radial moment are used. In another embodiment, only references using the moments are used. More generally, depending on the values used for the references C_b * and C_p *, the focus can be on one or more criteria in particular. The dispatching method 10 may be encoded into a computer program product comprising program code instructions for the execution of its steps, when this program is executed by the calculation unit 4.

Claims (12)

REVENDICATIONS1. A method of blade distribution simulation (P) on a turbomachine disk (D), the method comprising the steps of: - providing a plurality of blade configurations, each blade configuration comprising at least one measurement of balancing parameter (R, T, A) measured on a respective blade (P), - defining a plurality of bladed wheel splits, each splitting associating the blade configurations provided with positions on the disk (D) - searching and selecting (106) a bladed wheel distribution satisfying a criterion defined according to a predetermined cost function dependent on the balancing parameter measurements, the method being characterized by the fact that the search and your selection are conducted by successive iterations using a Tabou type algorithm. The method of claim 1, wherein a current iteration of the search and selection (106) comprises the steps of: - generating at least one neighboring distribution (V1, V2, V3, V4) from a reference distribution (REF), each neighboring distribution satisfying a predetermined resemblance criterion with the reference distribution (REF), - selecting, from neighboring distributions, a neighboring distribution optimizing the cost function as a new reference distribution, wherein at least one bladed wheel splitter having participated in at least one election during a previous iteration is not treated as a neighboring distribution during the current iteration. 3037168 22 3. Method according to claim 2, wherein at least a neighboring distribution (V1, V2) results from one or two exchange of blade positions between configurations of vanes of the reference distribution (REF). 5 4. The method of claim 3, wherein one of the positions exchanged in the context of one of the two exchanges and one of the positions exchanged in the context of the other of the two exchanges are diametrically opposed. 10 5. Method according to one of claims 2 to 4, wherein the neighboring distribution (V3, V4) results from one or two circular configurations of blade configuration applied to the reference distribution, each circular permutation being performed on a subset of 15 blade configurations associated with consecutive positions around the disk in the reference distribution. The method of claim 5, wherein each of the two circular permutations applied is representative of a blade configuration displacement from an initial position to a destination position, the two initial positions being diametrically opposed in the reference distribution. and the two end positions being diametrically opposed in the neighboring distribution obtained. 25 7. Method according to one of claims 2 to 6, wherein the similarity criterion used for the generation of the neighboring distribution is selected by an equiprobable draw between several predetermined resemblance criteria. 8. Method according to one of claims 2 to 7, comprising, for the reference distribution and for each adjacent distribution, the calculation of at least one unbalance from balancing parameter measurements of the same type. contained in the blade configurations of the distribution, and from the positions associated with the blade configurations in the distribution, the cost function depends on the unbalance. 5 9. Method according to one of claims 2 to 8, wherein the cost function depends on several unbalances, each unbalance relating to a type of balancing parameter measurement respective, and wherein at least one of the unbalance is the l object of a thresholding by application of the cost function. 10. Method according to one of claims 2 to 9, wherein the balancing parameter measurements are at least one of the following types: blade pitch angle, axial blade moment, radial moment of blade and tangential moment of dawn. 11. A method of manufacturing a turbomachine bladed wheel from blades and a disc, the method comprising the steps of - implementing a simulation method according to one of the preceding claims, and - assembly of the blades on the disk according to the bladed wheel distribution selected by the simulation method. A computer program product comprising program code instructions for performing the steps of the simulation method according to one of claims 1 to 10, when this program is executed by a computing unit (4). 30