Method for identifying a quality defect, implemented using a manufacturing machine, and associated implementing manufacturing machine

Abstract

The invention relates to a manufacturing machine, a computer program and storage for implementing a method for identifying a quality defect, which method is implemented using a manufacturing machine, and comprises the following steps: S1: acquiring measurement data for selected operating parameters of the manufacturing machine via the sensors during a manufacturing sequence, which operating parameters comprise spatial coordinates (CE) for a plurality of points, as well as at least one other operating parameter (Px) of the manufacturing machine, forming an acquired data set (LDA); S2: classifying the acquired data set (LDA); S3: computing at least one differential data set (DDI) for each point of the classified data set (LDC); S4: generating at least one functional graphical model (MGF); and S5: identifying the point of the associated functional graphical model (MGF) as having a quality defect.

Description

METHOD FOR IDENTIFYING A QUALITY DEFECT IMPLEMENTED BY MEANS OF A MANUFACTURING MACHINE AND ASSOCIATED MANUFACTURING MACHINE TECHNICAL FIELD OF THE INVENTION

[0001] The invention relates to the general technical field of machines and processes for machining and material deposition for the manufacture of parts, and more particularly to the field of surface modeling of machined or deposited parts and its application in the detection of surface condition defects.

[0002] TECHNOLOGICAL BACKGROUND OF THE INVENTION

[0003] Manufacturing machines are defined as any machine capable of removing or depositing material. Programmable machining centers typically offer multiple degrees of freedom to allow for numerous types of movements, thus enabling the machining of more complex geometries without the need to reposition the workpiece during a production run.

[0004] Programmable material deposition machines can also feature a plurality of degrees of freedom to allow many types of movement and thus enable material deposition according to more complex geometries without the need to reposition the part during a manufacturing process.

[0005] Such machines are used by entering a series of trajectory instructions in program form into a machine control unit, thus serving to move one or more tools on the surface of a workpiece, or to move one or more deposition tools on the surface of a mold in order to form the surfaces defining the part to be produced.

[0006] Trajectory instructions typically include commands regarding position, travel speed, cutting / deposition speed, and tool type, which must be observed during manufacturing. The program, or manufacturing sequence, is established to perform a series of operations using the machine to generate a specific geometry from a substantially uniform blank or a mold installed in the machine.

[0007] In practice, defects sometimes appear during the machining of certain blanks within the same production run. These defects can originate from inconsistencies in the geometry of the blanks, but can also be caused by other phenomena, such as a combination of machining parameters within the range leading to degraded performance that can damage the surface of the part.

[0008] It also happens that defects appear during the deposition of material onto the mold for the same production run. The origin of these defects can, for example, stem from bonding issues, but can also be caused by other phenomena resulting from a combination of manufacturing parameters within the run, leading to degraded performance that can damage the surface of the part, such as delamination, the presence of a foreign body, or cracking of the plies.

[0009] However, detecting these defects does not always allow for the identification of their causes, and the degradation of these parts represents a direct loss for manufacturers. Similarly, modifying machining / deposition parameters to incorporate large safety margins to prevent part degradation leads to underutilization of the machine's production capacity, resulting in wasted time and energy in part production.

[0010] In addition, the development of machining / deposition programs is currently poorly instrumented and results from iterative processes inducing significant and therefore costly development time, and is often associated with large safety margins at the level of implementation parameters to ensure that the program will never cause an error once implemented.

[0011] Therefore, there is a need to improve the methods of responding to a quality defect on a manufactured part, and to improve the processes of developing a manufacturing program.

[0012] The present invention remedies these drawbacks.

[0013] The invention relates to a method for identifying a quality defect implemented by means of a manufacturing machine capable of determining operating parameters of said manufacturing machine during a manufacturing range of a part by means of a tool having a point.

[0014] According to a general definition of the invention, the method according to the invention further comprises the following steps: - acquiring selected operating parameter data of the manufacturing machine during the manufacturing range, said operating parameters comprising on the one hand coordinates in space of the tool tip during its movement forming a plurality of points, and on the other hand at least one other operating parameter of the manufacturing machine associated with each of these points, forming a set of acquired data; - classifying the set of acquired data by association with at least one representative marker of the manufacturing range to form at least one set of classified data;- Matching each point of the classified data set with at least one point of a reference data set, having a marker representing the manufacturing range identical to that of the classified data set, based on the coordinates in space of each point of the classified measurement set and the coordinates in space of at least one point of the reference data set according to a chosen association rule, - Calculating at least one differential data set for each point of the classified data set with the point of the reference data set to which it has been associated; - Generating at least one point-by-point functional graphical model of the manufactured part from each calculated differential data set, the functional graphical model being configured to model the geometry of the part in space, each point of the graphical model of the part thus generated being associated with the differential data set;and - if the differential data set of a point in the functional graphical model has at least one value greater than a chosen quality threshold for at least one operating parameter, identify the point in the functional graphical model as having a quality defect.

[0015] The Applicant observed that the process according to the invention simultaneously allows for precise modeling of the manufactured part while also allowing for the detection and precise modeling of any manufacturing defect, both at the level of the manufactured part itself, which would not conform to the expected quality requirements, and at the level of performance in the control program used for the manufacture of the part, which could be optimized.

[0016] Furthermore, the method according to the invention makes it possible to locate instantly and precisely in space the operating parameters of the manufacturing machine with respect to reference data which do not necessarily have the same exact coordinates, and therefore makes it possible to adjust the number of points acquired to reference data of different resolution.

[0017] For example, the operating parameters of the manufacturing machine belong to the group formed by vibrations at the tool, power of a tool spindle, feed rate of the tool tip, current on each motor of the machine, torque at the tool spindle, spindle speed, noise generated, temperature at the tool tip, pressure applied by the tool, force applied by the tool, force relative to an effector, measured thickness of the part, deviation from contour, material deposition rate, tension applied to the deposited material, or similar parameter.

[0018] In addition, the representative marker of the manufacturing range includes machining context information belonging to the group formed by machined part / placed part, manufacturing orders, operations performed, at least one tool used, machine used, sensors used.

[0019] According to an embodiment of the invention, the calculation step further includes an additional substep of calculating a quality score for each point in the differential data set, the quality score being representative of a normalized deviation between the classified data and the reference data of associated points for each operating parameter of the manufacturing machine.

[0020] Advantageously, this step makes it possible to obtain a standardized scoring system for all operating parameters, and thus detect whether said operating parameters are compliant with the target quality score, compliant but potentially optimizable, or insufficient with respect to the target quality score, making it possible to identify ways to optimize the manufacturing process in addition to identifying a quality defect.

[0021] By way of non-limiting example, the association rule chosen for the matching step of the process according to the invention is carried out via the application of a method of extracting points belonging to the group formed by nearest neighbor, nearest neighbors, radius around the point, followed by an interpolation algorithm according to a mean, a median or a normal distribution.

[0022] In practice, the process also includes a step of recording the functional graphic model thus generated in at least one database.

[0023] Such a step allows for an acceleration of the process of identifying the causes of the identified quality defect while ensuring better repeatability of the iterations of the manufacturing processes of a part as well as traceability of the manufacturing process.

[0024] In practice, the data acquisition step for measurement of selected operating parameters of the manufacturing machine is implemented at a frequency between 100 Hz and 500 Hz, and preferably at a frequency between 150 Hz and 250 Hz.

[0025] Advantageously, acquisition at such frequencies provides sufficient resolution to ensure a proper match between points in the acquired data set and reference points during the differential data calculation step. Such acquisition frequencies also ensure that the allocated computing resources are reasonable and that the computation time does not exceed a value chosen for industrial-scale implementation.

[0026] The invention further relates to a manufacturing machine capable of acquiring selected operating parameter data during a manufacturing range for the implementation of the process according to the invention.

[0027] According to a second general definition of the invention, the manufacturing machine comprises: - data processing means, configured to acquire data of selected operating parameters of the manufacturing machine during the manufacturing range, said operating parameters comprising on the one hand coordinates in space of the tip of the tool during its movement forming a plurality of points, and on the other hand at least one other operating parameter of the manufacturing machine associated with each of these points, forming a set of acquired data; classify the set of acquired data by association with at least one marker representative of the manufacturing range to form at least one set of classified data;to map each point of the classified data set with at least one point of a reference data set, having a marker representing the manufacturing range identical to that of the classified data set, based on the coordinates in space of each point of the classified measurement set and the coordinates in space of at least one point of the reference data set according to a chosen association rule; to calculate at least one differential data set for each point of the classified data set with the point of the reference data set to which it has been associated; to generate at least one point-by-point functional graphical model of the manufactured part from each calculated differential data set, the functional graphical model being configured to model the geometry of the part in space, each point of the graphical model of the part thus generated being associated with the differential data set;if the differential data set of a point in the functional graphical model has at least one value greater than a chosen quality threshold for at least one operating parameter, identify the point in the functional graphical model as having a quality defect; and - means for storing data, configured to store at least one database integrating data belonging to the group formed by: a plurality of classified data sets, a plurality of differential data sets, causes of quality defect associated with the differential data indicating a quality defect, a plurality of reference data sets, the acquired data set(s), the classified data sets, the differential data sets, and each functional graphical model thus generated.

[0028] The invention and its various applications will be better understood by reading the following description and examining the accompanying figures. BRIEF DESCRIPTION OF THE FIGURES

[0029] Other advantages and features of the invention will become apparent from an examination of the description and drawings in which: schematically represents the steps of the process according to the invention; schematically represents the step of acquiring the process according to the invention; schematically represents the step of classifying the process according to the invention; schematically represents the step of matching and calculating the process according to the invention; schematically represents a particular embodiment of the step of calculating the process according to the invention; schematically represents the step of generating a functional graphical model of the process according to the invention; schematically represents the step of identifying a quality defect of the process according to the invention; and schematically represents an additional step of querying a database and recording the process.The figures are presented for illustrative purposes only and are in no way limiting to the invention. DETAILED DESCRIPTION

[0030] Unless otherwise specified, the same element appearing on different figures has a unique reference.

[0031] With reference to Figures 1 to 8, the invention relates to a method for identifying a quality defect implemented using a manufacturing machine capable of determining operating parameters of said manufacturing machine during a manufacturing process of a part using a tool having a tip

[0032] Manufacturing machines are understood to mean machining machines and material deposition machines, configured to allow the production of a part according to a manufacturing range.

[0033] A quality defect is defined as any divergence between reference parameters and the operating parameters of the manufacturing machine, and any divergence beyond a chosen threshold, of a chosen shape, on the manufactured part compared to the expected part.

[0034] The method according to the invention first includes an acquisition step S1 of selected operating parameter data of the manufacturing machine during a manufacturing range, said operating parameters comprising on the one hand coordinates in CE space of the tool tip during its movement forming a plurality of points, and on the other hand at least one other operating parameter Px of the manufacturing machine associated with each of these points, forming a set of acquired data LDA.

[0035] The term "point" or "machined / deposited point" refers to each action of acquiring operating parameter data having unique coordinates in CE space, for which at least one operating parameter Px of the manufacturing machine is acquired.

[0036] The tool tip is defined as the end or surface of the tool configured to come into contact with a workpiece during machining and the end or surface configured to apply the material to be deposited during material deposition manufacturing.

[0037] By way of non-limiting example, the manufacturing machine operating parameters Px other than the coordinates in CE space belong to the group formed by vibration at the tool, tool spindle power, tool tip feed rate, current on each machine motor, torque at the tool spindle, spindle speed, noise generated, tool tip temperature, pressure applied by the tool, force applied by the tool, force relative to an effector, measured thickness of the part, deviation from contour, material deposition rate, tension applied to the deposited material, or similar parameter.

[0038] A tool is defined as any material capable of removing material in order to transform the raw material (block of material) into a finished part with the desired shape, dimensions and surface quality, or capable of depositing material in order to obtain a finished part with the desired shape, dimensions and surface quality.

[0039] According to a particular embodiment of the invention, the acquisition step S1 is implemented via sensors integrated into the manufacturing machine.

[0040] According to a second embodiment of the invention, the acquisition step S1 comprises the acquisition of data obtained via sensors integrated into the manufacturing machine in combination with data acquired via at least one remote sensor, which records at least one operating parameter of the manufacturing machine Px other than the coordinates in CE space. The recording of the manufacturing machine's operating parameter Px by the remote sensor is performed on a common time base with the sensors of the manufacturing machine, said acquired data being synchronized and correlated with the data from the sensors of the manufacturing machine.

[0041] A sensor is defined as any means of measuring a parameter, calculating from a parameter measurement, or acquiring data available or produced by the manufacturing machine.

[0042] In practice, the S1 acquisition step of the selected operating parameter data CE, Px of the manufacturing machine is implemented at a frequency between 100 Hz and 500 Hz.

[0043] This specific frequency range enables the acquisition of real-time measurement data. Furthermore, such a frequency range allows for the acquisition of a batch of LDA data with a resolution high enough to eliminate the need for time-domain measurements requiring additional calculation and conversion steps, thus minimizing the physical computing resources allocated to this task.

[0044] Advantageously, this frequency is adapted to the speed of movement of the tip, and the targeted accuracy of acquisition.

[0045] According to a preferred mode, the S1 acquisition step of the selected operating parameter data CE, Px of the manufacturing machine is implemented at a frequency between 150 Hz and 250 Hz.

[0046] The acquisition step S1 further includes sub-steps among which:-S12 Acquire data of the operating parameter(s) of the manufacturing machine Px; and-S13 record the acquired data on storage means in the form of an LDA acquired data batch.

[0047] According to a particular embodiment, the acquisition substep S12 is preceded by a substep consisting of calibrating S11 the sensors involved in said acquisition substep S12.

[0048] The LDA acquired data set for a given manufacturing range therefore includes for each acquisition point: - the coordinates in CE space of the tool at the acquisition point; and - the acquisition data of at least one measured operating parameter Px of the manufacturing machine associated with the coordinates in CE space of the tool at the acquisition point.

[0049] As a non-limiting example, the LDA data set for a given manufacturing range therefore includes for each acquisition point: - the coordinates in CE space of the tool of the acquisition point; - the data of a plurality of operating parameters Px of the manufacturing machine, each associated with the coordinates in CE space of the tool of the acquisition point.

[0050] The applicant observed that acquiring Px operating parameter data for the manufacturing machine at such frequencies, in combination with coordinates in CE space, made it possible to obtain a batch of LDA data with sufficient resolution for accurate modeling of the machined part. This particular approach also avoids generating an excessive volume of data that could significantly slow down any processing equipment configured to handle this / these batches of LDA data, thus enabling the implementation of the process on an industrial scale when numerous parts need to be manufactured.

[0051] The method according to the invention further includes a step of classifying the acquired LDA data by association with at least one MARQ marker representative of the manufacturing range to form at least one batch of classified LDC data.

[0052] In practice, the MARQ marker representing the manufacturing range includes machining context information M1, M2, M3, Mx belonging to the group formed by machined part / deposited part, manufacturing orders, operations performed, at least one tool used, machine used, sensors used, or any other information allowing the implementation context of the manufacturing range to be defined.

[0053] As an example, the LDC classified data set for a given manufacturing range therefore includes for each acquisition point: - the coordinates in CE space of the tool at the acquisition point; - the data of the operating parameter(s) Px of the manufacturing machine associated with the coordinates in CE space of the tool at the acquisition point; and - at least one MARQ marker including the machining context information M1, M2, M3, Mx.

[0054] Advantageously, this S2 classification step makes it possible to associate the LDA acquired data set with a specific implementation context of the manufacturing range, and thus make comparable any data set having a MARQ marker representative of the identical manufacturing range.

[0055] The process according to the invention then includes a step of matching S2B each point of the LDC classified data lot with at least one point of a DREF reference data lot, having a MARQ marker representative of the manufacturing range identical to that of the LDC classified data lot.

[0056] In practice, the S2B matching step includes a first sub-step of identification S2B1 of at least one set of DREF reference data from a BDD database, said set of DREF reference data having an identical MARQ marker.

[0057] The S2B matching step includes a second association substep S2B2 implemented on the basis of the coordinates in CE space of each point in the LDC classified measurement set and the coordinates in CE space of at least one point in the DREF reference data set by applying at least one chosen association rule.

[0058] As an example, the association rule includes a method for extracting points belonging to the group formed by nearest neighbor, nearest neighbors, radius around the point, followed by an interpolation algorithm based on a mean, a median, or a normal distribution.

[0059] According to a first embodiment, the reference data set DREF includes at least: - the coordinates in CE space of each theoretical point of the tool tip path; - the theoretical data of the operating parameter(s) Px of the manufacturing machine in absolute values ​​associated with said coordinates in CE space of the tool tip; and - the MARQ marker including machining context information M1, M2, M3, Mx identical to the machining context information M1, M2, M3, Mx of the classified data set LDC.

[0060] The DREF reference data set therefore includes theoretical nominal values ​​of the manufacturing machine's operating Px parameters associated with a set of coordinates in CE space for each theoretical point of the part to be manufactured.

[0061] According to a second embodiment, the DREF reference data set includes at least: - the CE space coordinates of the tool of each acquisition point of a master part; - the measurement data of the operating parameter(s) Px of the manufacturing machine measured associated with the CE space coordinates of the tool of the measurement point of the master part; and - the MARQ marker including machining context information M1, M2, M3, Mx identical to the machining context information M1, M2, M3, Mx of the classified data set LDC.

[0062] According to a third embodiment, the DREF reference data set includes at least: - the CE space coordinates of the tool tip for each acquisition point of a previously manufactured part; - the acquisition data of the operating parameter(s) Px of the manufacturing machine associated with said CE space coordinates for each acquisition point of a previously manufactured part; and - the MARQ marker including machining context information M1, M2, M3, Mx identical to the machining context information M1, M2, M3, Mx of the classified data set LDC.

[0063] In practice, each measurement point in the LDC classified data set will have at least one corresponding point in the DREF reference data set, thus forming a set of associated points.

[0064] Advantageously, this matching makes it possible to calculate a differential data set DDI even if the classified data set LDC and the reference data set DREF do not have identical resolution.

[0065] The method according to the invention then includes a calculation step S3 of at least one differential data set DDI for each point of the classified data set LDC with the at least one point of the reference data set DREF to which the point of the classified data set LDC has been associated.

[0066] The S3 calculation step includes calculating the absolute value of the difference between the value of each manufacturing machine operating parameter Px at the acquisition point of the classified data set LDC and the value for that same parameter Px at the associated point of the reference data set DREF, this operation being repeated for each set of associated points.

[0067] The Applicant observed that the calculation step made it possible to verify that the compared data are comparable, of the same context, and corresponding so that the result of said calculation is representative of a deviation or not with a chosen quality reference and thus accurate.

[0068] According to one embodiment of the invention, the calculation step S3 further comprises an additional substep for calculating a quality score S33 for each parameter Px of each point in the differential data set DDI, the quality score representing a normalized deviation between the result of the calculation of the differential data DDI for each parameter Px and a threshold set for each parameter Px. The differential data DDI therefore further comprises a quality score associated with each parameter Px for each point.

[0069] Advantageously, calculating a quality score allows quality scores for all operating parameters Px of the manufacturing machine to be normalized according to a single scale, thus obtaining a more uniform batch of differential DDI data.

[0070] The method according to the invention then includes a generation step S4 of at least one point-to-point functional graphical model MGF of the machined part from each batch of calculated differential data DDI.

[0071] Point-to-point refers to a functional MGF graphical model whose graphical representation is a point cloud or a point cloud ordered in the sequential order of manufacture.

[0072] The MGF point-to-point functional graphical model is configured to model the geometry of the manufactured part in space, with each point thus generated being associated with at least one set of DDI differential data.

[0073] According to a particular embodiment of the invention, the generation step S4 of the functional graphical model MGF includes a filter substep S41 of the differential data DDI, configured to compare the value of the calculation result for each parameter Px to a noise threshold SB, and to exclude the points whose calculation result of each parameter Px of the differential data DDI is greater than or equal to said noise threshold SB.

[0074] Advantageously, if the operating parameters Px of the manufacturing machine for the batch of differential data (DDI) associated with at least one point of the functional graphical model (MGF) are above the noise threshold SB, this differential data (DDI) is considered an anomaly / noise, and the point(s) are not included in the generation of the MGF. This allows for the exclusion of outliers that could interfere with the identification of a quality defect.

[0075] As a non-limiting example, the differential data (DDI) can be represented using a color code defined according to the magnitude of the deviation, or as a vector extending from the point in question, allowing for a four-dimensional representation. This step therefore enables the creation of a graphical representation of the manufactured part, highlighting the areas where a parametric deviation for at least one parameter Px has been observed compared to the reference data (DREF).

[0076] Advantageously, a user can therefore immediately observe, via the variations in the differential DDI data, the condition of the part and note the presence or absence of a quality defect, and can thus establish a potential correlation between the presence and type of defect observed, and the deviation of the observed parameter Px.

[0077] The method according to the invention further includes a step S5 of identifying a quality defect consisting of applying a quality threshold SQ chosen for each of the parameters Px, and comparing said quality threshold SQ to the differential data set DDI calculated for each of the points of the functional graphical model MGF.

[0078] If the differential data set DDI of a point in the MGF functional graphical model has at least one value greater than the chosen SQ quality threshold for at least one parameter Px, identify the associated point in the MGF functional graphical model as having a quality defect and identify the associated point in the MGF functional graphical model for that parameter Px as non-compliant.

[0079] If the differential data set DDI of a point in the MGF functional graphical model has a value less than or equal to the SQ quality threshold chosen for a parameter Px, identify the point in the MGF functional graphical model associated for that parameter Px as compliant.

[0080] According to a first embodiment of the invention, each quality threshold SQ is compared to the calculation results of the differential data set DDI for each corresponding parameter Px of the differential data set DDI.

[0081] According to an alternative embodiment of the invention, the quality threshold SQ applied to the quality score is chosen and compared to the quality score for each point of the differential data set DDI.

[0082] Advantageously, the process according to the invention makes it possible to guarantee the stability and robustness of manufacturing processes such as machining and material deposition while generating a digital twin of the manufactured part.

[0083] The MGF functional graphic model forms a digital twin of the manufactured part, allowing the manufacturing process parameters to be located instantly and precisely in space, while also ensuring traceability of manufacturing operations to certify that the manufactured part follows the manufacturing instructions.

[0084] The process according to the invention also makes it possible to reduce the industrialization time of parts by decreasing the number of iterations required until the manufacturing process is validated for industrialization.

[0085] Advantageously, the process according to the invention also makes it possible to reduce the machining cycle time by allowing the optimization of trajectories and operating conditions of the manufacturing machine.

[0086] Finally, the process according to the invention makes it possible to reduce the time required to inspect the part by focusing the monitoring and control plan of the manufactured part on what is strictly necessary, namely the areas of the manufactured part for which a quality defect has been identified.

[0087] The identification step S5 is followed by a registration step S6 of the functional graphical model MGF in the BDD database.

[0088] In practice, the BDD database integrates a plurality of functional graphical models MGF, a plurality of classified data sets LDC, a plurality of differential data sets DDI, a plurality of quality scores associated with the differential data sets DDI, at least one list of causes of quality defect associated with the differential data DDI indicating a quality defect, and the reference data sets DREF.

[0089] In practice, the S6 recording step of the MGF functional graphical model includes an additional S7 query step of the BDD database based on the generated and recorded MGF functional graphical model.

[0090] In practice, for each quality defect identified on the MGF functional graphic model, compare the LDD differential data set of said MGF functional graphic model recorded with the DDI differential data sets stored in the BDD database.

[0091] If a match is found for a parameter Px, check if at least one potential cause is associated with it in said database BDD.

[0092] If at least one potential cause is associated with the matching data, mark S71 the quality defect identified on the MGF functional graphic model as having an identified potential cause and save the MGF functional graphic model.

[0093] If no potential cause is associated with the matching data, mark S72 the quality defect identified on the MGF functional graphic model as having an unidentified potential cause and save the MGF functional graphic model.

[0094] The invention further relates to a computer program comprising instructions which, when executed on computer-type processing means, cause the computer to implement the process as described.

[0095] The invention further relates to a computer-readable data storage medium on which the computer program according to the invention is recorded.

[0096] The invention further relates to a manufacturing machine capable of acquiring selected operating parameter data during a manufacturing range for the implementation of the process according to the invention.

[0097] In practice, the manufacturing machine according to the invention includes at least one tool having a point.

[0098] The tool tip is understood to be the end or surface of the tool configured to come into contact with a workpiece during machining and the end or surface configured to apply the material to be deposited.

[0099] The manufacturing machine according to the invention further includes data processing means, configured to: - acquire S1 selected operating parameter data of the manufacturing machine during the manufacturing range, said operating parameters comprising on the one hand coordinates in CE space of the tool tip during its movement forming a plurality of points, and on the other hand at least one other operating parameter Px of the manufacturing machine associated with each of these points, forming a set of acquired data LDA;

[0100] -classify S2 the batch of acquired LDA data by association with at least one MARQ marker representative of the manufacturing range to form at least one batch of classified LDC data;

[0101] - to match S2B each point of the LDC classified data set with at least one point of a DREF reference data set, having a MARQ marker representative of the manufacturing range identical to that of the LDC classified data set, based on the coordinates in space of each point of the LDC classified measurements set and the coordinates in space of at least one point of the DREF reference data set according to a chosen association rule,

[0102] -calculate S3 at least one differential data set DDI for each point of the classified data set LDC with the point of the reference data set DREF to which it was associated;

[0103] -generate S4 at least one point-to-point functional graphical model MGF of the manufactured part from each batch of calculated differential data, the functional graphical model being configured to model the geometry of the part in space, each point of the graphical model of the part thus generated being associated with the batch of differential data DDI;

[0104] -if the differential data set DDI of a point of the functional graphical model MGF has at least one value greater than a quality threshold SQ chosen for at least one operating parameter Px, identify S5 the point of the functional graphical model MGF as having a quality defect.

[0105] Finally, the manufacturing machine also includes data storage means, configured to store at least one database BDD incorporating data belonging to the group formed by: a plurality of LDC classified data sets, a plurality of DDI differential data sets, of quality defect causes associated with the DDI differential data indicating a quality defect, a plurality of DREF reference data sets, the LDA acquired data set(s), the LDC classified data sets, the DDI differential data sets, and each MGF functional graphic model thus generated.

[0106] According to a particular embodiment of the invention, the manufacturing machine includes integrated sensors configured to measure operating data during the acquisition step S1.

[0107] According to a second embodiment of the invention, the manufacturing machine includes integrated sensors configured to measure operating data during the acquisition step S1 in combination with at least one remote sensor, which records at least one operating parameter of the manufacturing machine Px other than the coordinates in CE space. The recording of the manufacturing machine's operating parameter Px by the remote sensor is performed on a common time base with the sensors of the manufacturing machine, said acquired data being synchronized and correlated with the data from the sensors of the manufacturing machine.

[0108] A sensor is defined as any means of measuring a parameter, calculating from a parameter measurement, or acquiring data available or produced by the manufacturing machine.

[0109] Advantageously, the manufacturing machine according to the invention makes it possible to precisely follow the manufacturing process while reducing the industrialization time of a part to be manufactured and reducing the time of quality control operations.

[0110] It also allows verification of compliance with manufacturing instructions and reduces the development cost of a series of parts to be manufactured by ensuring manufacturing homogeneity and cost reduction by reducing the number of manufacturing iterations of the first parts before mass production.

Claims

Method for identifying a quality defect implemented by means of a manufacturing machine capable of determining operating parameters of said manufacturing machine during a manufacturing range of a part using a tool having a point, comprising the following steps: S1: acquire selected operating parameter data of the manufacturing machine during the manufacturing range, said operating parameters comprising on the one hand coordinates in space (CE) of the tool tip during its movement forming a plurality of points, and on the other hand at least one other operating parameter (Px) of the manufacturing machine associated with each of these points, forming a set of acquired data (LDA); S2: classify the set of acquired data (LDA) by association with at least one marker (MARQ) representative of the manufacturing range to form at least one set of classified data (LDC);S2B: Matching each point of the classified data set (LDC) with at least one point of a reference data set (DREF), having a marker (MARQ) representative of the manufacturing range identical to that of the classified data set (LDC), based on the coordinates in space of each point of the classified measurement set (LDC) and the coordinates in space of at least one point of the reference data set (DREF) according to a chosen association rule including the application of a point extraction method belonging to the group formed by nearest neighbor, nearest neighbors, radius around the point, followed by an interpolation algorithm according to a mean, a median or a normal distribution, S3: Calculate at least one differential data set (DDI) for each point of the classified data set (LDC) with the point of the reference data set (DREF) to which it has been associated;S4: generate at least one point-by-point functional graphical model (FGM) of the manufactured part from each batch of calculated differential data, the functional graphical model being configured to model the geometry of the part in space, each point of the generated part graphical model being associated with the batch of differential data (DDI); S5: if the batch of differential data (DDI) of a point of the functional graphical model (FGM) has at least one value greater than a chosen quality threshold (QT) for at least one operating parameter (Px), identify the point of the functional graphical model (FGM) as having a quality defect. A method according to claim 1, characterized in that the operating parameters (Px) of the manufacturing machine belong to the group formed by vibrations at the tool level, power of a tool spindle, feed rate of the tool tip, current on each motor of the machine, torque at the tool spindle, spindle speed, noise generated, temperature at the tool tip, pressure applied by the tool, force applied by the tool, force relative to an effector, measured thickness of the part, deviation from the contour, material deposition rate, tension applied to the deposited material, or similar parameter. Method according to claim 1 or 2, characterized in that the marker (MARQ) representative of the manufacturing range includes machining context information (M1, M2, M3, Mx) belonging to the group formed by machined part / deposited part, manufacturing orders, operations performed, at least one tool used, machine used, sensors used. A method according to any one of claims 1 to 3, characterized in that the calculation step S3 further comprises an additional calculation substep (S33) of a quality score for each point of the differential data set (DDI), the quality score being representative of a normalized deviation between the classified data (LDC) and the reference data (DREF) of associated points for each operating parameter (Px) of the manufacturing machine. A method according to any one of claims 1 to 4, characterized in that it further comprises a step S6 of recording the functional graphic model (MGF) in at least one database (BDD). A method according to any one of claims 1 to 5, characterized in that the acquisition step S1 of the selected operating parameter data (Px) of the manufacturing machine is implemented at a frequency between 100 Hz and 500 Hz. A method according to any one of claims 1 to 5, characterized in that the acquisition step S1 of the selected operating parameter data (Px) of the manufacturing machine is implemented at a frequency between 150 Hz and 250 Hz. Manufacturing machine capable of acquiring selected operating parameter data during a manufacturing range for the implementation of the process according to claims 1 to 7, characterized in that it comprises: - data processing means, configured to acquire (S1) selected operating parameter data of the manufacturing machine during the manufacturing range, said operating parameters comprising on the one hand coordinates in space (CE) of the tool tip during its movement forming a plurality of points, and on the other hand at least one other operating parameter (Px) of the manufacturing machine associated with each of these points, forming a batch of acquired data (LDA); classify (S2) the batch of acquired data (LDA) by association with at least one marker (MARQ) representative of the manufacturing range to form at least one batch of classified data (LDC);to match (S2B) each point of the classified data set (LDC) with at least one point of a reference data set (DREF), having a marker (MARQ) representative of the manufacturing range identical to that of the classified data set (LDC), on the basis of the coordinates in space of each point of the classified measurement set (LDC) and the coordinates in space of at least one point of the reference data set (DREF) according to a chosen association rule, calculate (S3) at least one differential data set (DDI) for each point of the classified data set (LDC) with the point of the reference data set (DREF) to which it has been associated;generate (S4) at least one point-by-point functional graphical model (FGM) of the manufactured part from each batch of calculated differential data, the functional graphical model being configured to model the geometry of the part in space, each point of the graphical model of the part thus generated being associated with the batch of differential data (DDI); if the batch of differential data (DDI) of a point of the functional graphical model (FGM) has at least one value greater than a quality threshold (QT) chosen for at least one operating parameter (Px), identify (S5) the point of the functional graphical model (FGM) as having a quality defect;and data storage means, configured to store at least one database (DB) incorporating data belonging to the group formed by: a plurality of classified data sets (LDC), a plurality of differential data sets (DDI), of quality defect causes associated with differential data (DDI) indicating a quality defect, a plurality of reference data sets (DREF), the acquired data set(s) (LDA), the classified data sets (LDC), the differential data sets (DDI), and each functional graphical model (MGF) thus generated.; Computer program comprising instructions which, when executed on a computer, cause the computer to implement the method according to any one of claims 1 to 7. Computer-readable data storage medium on which the computer program according to claim 9 is stored.

Images