Method and apparatus for inspecting compliance of a workpiece

Abstract

A method for checking the compliance of a workpiece, comprising the steps of: - estimating (14) a risk of non-compliance (TNC k ) of a characteristic (X) based on a probability law (L k ) associated with said characteristic (X); and - verifying (16) whether the estimated risk of non-compliance (TNC k ) meets a decision criterion, and if so, declaring that the workpiece is compliant for said characteristic (X); if not, measuring (20) a value of said characteristic, determining whether the workpiece is compliant based on the measured value, and updating the probability law (L k ) associated with said characteristic (X) based on said measured value.​​​​​

Description

Invention Field

[0001] This disclosure relates to a method for inspecting the compliance of workpieces, and more particularly to a method that allows for improved efficiency of quality control in industrial environments. Technical Background

[0002] During the production process, workpieces are defined by specifications related to their dimensions, materials, internal stresses, etc. These specifications typically include tolerances, which are the maximum permissible deviations from the nominal values ​​of the specifications. As part of quality control, workpieces are verified to conform to their specifications, meaning that the differences between the workpiece's characteristics and its nominal values ​​do not exceed the tolerances allowed.

[0003] Quality control can indicate excessive additional time in the production process. This time is especially important for parts produced in large quantities.

[0004] To reduce this time and increase productivity, key statistical methods have been envisioned for examining only a reduced sample of workpieces and inferring compliance from this to larger batches. These methods are satisfactory for workpieces with only moderately critical reliability.

[0005] On the other hand, these methods are not suitable for components where reliability is critical, because it is difficult, if not impossible, to prove that the risk of non-compliance going undetected is low. Therefore, a new method for checking workpiece compliance is needed that is time-efficient and reduces the risk of non-compliance going undetected.

[0006] Such a method has been proposed in FR 3 063 153 A1. Invention Summary

[0007] This disclosure is intended to at least partially meet this need.

[0008] This disclosure relates to a method for checking the compliance of a workpiece having at least one characteristic, the method comprising the following steps:

[0009] Estimating the non-compliance risk of characteristics based on probability laws associated with those characteristics; and

[0010] Verify whether the estimated non-compliance risk meets the decision criteria, and if it does, declare the workpiece compliant with that characteristic; if not, measure the value of that characteristic, determine whether the workpiece is compliant based on the measured value, and update the probability law associated with that characteristic based on the measured value.

[0011] By using a probabilistic law, such as the one described above, updated with each actual measurement performed on the workpiece, this inspection method is able to adapt to any drift in production. Furthermore, this adaptation does not require any creativity from human experts.

[0012] In some embodiments, the determination criteria include a first criterion that verifies the estimated noncompliance risk is below a fixed threshold.

[0013] In some embodiments, the determination criteria include a second criterion that verifies, as a function of the estimated non-compliance risk and the estimated non-compliance risk of workpieces to which the determination criteria have previously been applied, that the value is below a threshold as a function of the number of said workpieces.

[0014] In some embodiments, the value is a function of the estimated non-compliance risk and the estimated non-compliance risk of artifacts that have previously met the decision criteria.

[0015] In some embodiments, the probability law is updated using the sequential Monte Carlo method.

[0016] In some embodiments, the workpiece has p characteristics, where p is an integer at least equal to 2, and for each of the p characteristics, steps are performed to estimate the non-compliance risk and verify the judgment criteria.

[0017] In some embodiments, when the determination criterion is not met for at least one of the p characteristics:

[0018] Define a list of at least one characteristic that does not satisfy the determination criteria;

[0019] The measurement operation list is defined based on the list of at least one characteristic that does not meet the decision criteria, and a dependency tree that associates each of the p characteristics with at least one measurement operation that can be performed on the workpiece.

[0020] A list of additional characteristics of the workpiece is also defined, which includes at least some of the characteristics that can be measured by performing measurement operations in the measurement operation list based on the measurement operation list.

[0021] Measure the value of each of the at least one characteristic that does not meet the judgment criteria, and the value of each characteristic in the additional characteristic list; and

[0022] Based on the values ​​measured in this way, each of the probability laws associated with each of the at least one characteristic that does not meet the decision criterion is updated, and each of the probability laws associated with the additional characteristic is updated.

[0023] In this way, the method makes the most of the information that can be extracted from the measurement operations, which in all cases need to be performed after the results of the verification decision criteria steps.

[0024] The present invention also relates to a method for monitoring workpiece production, comprising the following steps:

[0025] - Provide multiple workpieces of the same type; and

[0026] - Use the previously described methods to check the compliance of each of these artifacts.

[0027] In some embodiments, the compliance of the artifacts is checked in the order in which they are provided.

[0028] In other embodiments, their compliance is checked in an order different from the order in which each of the workpieces is provided.

[0029] In some embodiments, compliance of the at least one characteristic is checked in an order different from the order in which the items are provided, independent of the estimated non-compliance risk of each of the at least one characteristic, and the value of at least one characteristic of the items is measured in an order opposite to the order in which the items are provided.

[0030] In other embodiments, compliance is checked in an order different from the order in which each of the workpieces is provided, all measured values ​​are stored in memory, and the probability law associated with each of the at least one characteristic of the workpiece is updated based on the measured values ​​stored in memory before estimating the non-compliance risk of the workpiece.

[0031] In this way, the estimation of non-compliance risk is based on more measurement data points and is therefore more reliable. Consequently, this monitoring method is more robust to drift in production.

[0032] In one particular embodiment, the different steps of the inspection or monitoring method are determined by computer program instructions.

[0033] Therefore, the present invention also relates to a program on a data medium that can be implemented in an inspection device or more generally in a computer, the program including instructions suitable for implementing steps such as the inspection or monitoring methods described above.

[0034] The program can use any programming language and can be in the form of source code, object code, or intermediate code between source code and object code, such as a partially compiled form or any other suitable form.

[0035] The present invention also relates to a data medium that can be read by a computer or microprocessor and includes program instructions such as those described above.

[0036] Data media can be any entity or device capable of storing programs. For example, the media can include storage devices such as ROM (e.g., CD-ROM) or ROM of microelectronic circuits, or magnetic recording devices (e.g., floppy disks or hard disks).

[0037] Furthermore, the data medium can be a transmissible medium capable of being transmitted via cable or optical fiber, wirelessly, or through other means, such as electrical or optical signals. The program of this invention can be downloaded, in particular, from an internet-type network.

[0038] This disclosure also relates to an apparatus for checking the compliance of a workpiece having at least one characteristic, comprising a measuring machine configured to measure the value of the at least one characteristic, a command module, a transmission device configured to transmit the measured value from the measuring machine to the command module and to transmit instructions from the command module to the measuring machine, the command module being configured to estimate the non-compliance risk of the characteristic based on a probability law associated with the characteristic, verify whether the estimated non-compliance risk meets a determination criterion, and if so, declare the workpiece compliant with the characteristic; if not, the command module is configured to instruct the measuring machine to measure the value of the characteristic to determine whether the workpiece is compliant based on the measured value, and to update the probability law associated with the characteristic based on the measured value.

[0039] The device can be used to implement the previously described inspection or monitoring methods. Additionally, the device may include all or some of the features previously detailed in relation to the method.

[0040] Brief description of the attached figures

[0041] A better understanding of the invention and its advantages will be gained from reading the following detailed description of embodiments of the invention given as non-limiting examples. This description refers to the accompanying drawings, in which:

[0042] [ Figure 1 ] Figure 1 This is a block diagram illustrating the inspection method and monitoring method according to the first embodiment;

[0043] [ Figure 2 ] Figure 2 This is a block diagram illustrating the inspection method and monitoring method according to the second embodiment;

[0044] [ Figure 3 ] Figure 3 It is an explanation Figure 2 A block diagram of a sub-part of the block diagram;

[0045] [ Figures 4A-4B ] Figures 4A to 4B This is illustrated in a practical example. Figure 3 A diagram illustrating each step of the flowchart in the document;

[0046] [ Figure 4C-4D ] Figures 4A to 4D This is illustrated in a practical example. Figure 3 A diagram illustrating each step of the flowchart in the document;

[0047] [ Figure 5A ] Figure 5A This is a block diagram illustrating the inspection method and monitoring method according to the third embodiment;

[0048] [ Figure 5B ] Figure 5B and 5C This is explained in the example. Figure 5A A diagram illustrating how the method works;

[0049] [ Figure 5C ] Figure 5B and 5C This is explained in the example. Figure 5A A diagram illustrating how the method works;

[0050] [ Figure 6A ] Figure 6A This is a block diagram illustrating the inspection method and monitoring method according to the fourth embodiment;

[0051] [ Figure 6B ] Figure 6B and 6C This is explained in the example. Figure 6A A diagram illustrating how the method works;

[0052] [ Figure 6C ] Figure 6B and 6C This is explained in the example. Figure 6A A diagram illustrating how the method works;

[0053] [ Figure 7 ] Figure 7 The inspection equipment was explained in a illustrative way.

[0054] Detailed description of the invention

[0055] Reference Figure 1 Describes a method for checking workpiece compliance according to a first embodiment. Figure 1 The steps of the method are illustrated schematically.

[0056] The method for checking the compliance of workpiece 10 includes step 12 of providing workpiece k.

[0057] In the first embodiment, workpiece k has a characteristic X for which compliance is to be verified. Throughout this document, the subscripts containing the letter k relate to workpiece k for which compliance is verified during inspection method 10. In one example, characteristic X is a dimensional value measured on workpiece k.

[0058] The compliance of feature X is determined by the nominal value N and the upper tolerance IT. + >N and lower tolerance IT - <N is defined. Therefore, if the value of characteristic X lies in the interval [IT] - IT +If N ∈ [IT], then it is considered compliant. - IT + ].

[0059] Provide workpiece k, followed by an estimate of the non-compliance risk of workpiece k by TNC. k Estimate step 14. This non-compliance risk TNC k It is estimated based on the probability law associated with characteristic X, as detailed below.

[0060] Inspection method 10 then includes a verification step 16, where the estimated non-compliance risk TNC obtained in estimation step 14 is verified. k The judgment criteria are met.

[0061] If the result of verification step 16 is positive, that is, the estimated non-compliance risk TNC k If the decision criterion is met, the method proceeds to step 18, where it is declared that workpiece k is compliant with characteristic X. As detailed below, the decision criterion is a function of the acceptable level of non-compliance risk α.

[0062] Conversely, if the result of verification step 16 is negative, that is, the estimated non-compliance risk TNC is not considered negative. k If the judgment criteria are not met, the method proceeds to measurement step 20, where the actual value X of characteristic X is measured. k .

[0063] The fact that the decision criterion is met at step 16 indicates the estimated non-compliance risk TNC. k This involves an actual non-compliance risk that is sufficiently low for the acceptability of considering risk level α. The fact that the decision criterion is not met at step 16 indicates that the estimated non-compliance risk is TNC. k This involves actual non-compliance risks that are not low enough for the acceptability of considering risk level α. Examples of judgment criteria are detailed below.

[0064] Once the actual value of characteristic X is X k Once detected, the method moves to step 22, where X is verified. k Is it compliant? If X k Compliance, i.e., X k ∈[IT - IT + If [the condition is met], the method moves to step 18, where it declares that workpiece k is compliant with characteristic X. Conversely, if X [is not compliant], the method moves to step 18. k If non-compliant, the method moves to step 19, where it declares that workpiece k is non-compliant.

[0065] As previously mentioned, based on the probability law L associated with characteristic X.k To estimate the estimated non-compliance risk TNC k ,Right now Where y k Assumed to follow the probability law L k .

[0066] In addition, the probability law L k Based on the actual value X of characteristic X k To update. More specifically, such as Figure 1 As explained in the text, if measurement step 20 has already been performed, then inspection method 10 also performs update step 24, in which the probability law L is updated. k To obtain the updated probability law L k+1 .

[0067] In this embodiment, the probability law L k The method is updated using a sequential Monte Carlo approach, also known as a "particle filter" or "granular filter." This method is well-known in the literature and therefore will not be described in detail here. To briefly recap, using this method, the measurement distribution model M(y|θ) is determined. The distribution can be of any suitable type (e.g., Gaussian, Weibull, or bimodal distribution) where the parameter θ is not precisely known. Therefore, the parameter θ itself is determined by the empirical distribution P. k (θ) Description. When a new observation is available, i.e., when the actual value X is obtained. k At that time, the empirical distribution P k (θ) Based on the actual value X k It is examined and subsequently drifts with Brownian motion. This gives a new empirical distribution P. k+1 (θ). The computation required for this operation is approximated using Monte Carlo simulation. Finally, knowing P k+1 In the case of (θ), by multiplying P over all possible parameter values ​​θ k+1 Integrating (θ)×M(x|θ) yields the measurement X on workpiece k+1. k+1 The expected distribution L k+1 .

[0068] In all cases, after updating step 24, the updated probability law L is obtained. k+1 It is stored in memory and used to estimate the non-compliance risk of workpiece k+1.

[0069] By using a probability law that is updated at each actual measurement on the workpiece in the manner just described, inspection method 10 is able to adapt to any drift in production.

[0070] As alternatives to the sequential Monte Carlo method, the weighted moving average (wMA) or the autoregressive comprehensive moving average (ARIMA) method may be used. These methods are well-known in the literature and will therefore not be described in detail here. However, the sequential Monte Carlo method is preferred for two reasons. First, resorting to the Monte Carlo method allows for the approximation of the empirical distribution P. k+1 (θ), even if it cannot be found exactly (this is the case once the distribution with a parameter θ that is not yet precisely known does not follow Gaussian law). Secondly, because it is a sequential method, it is associated with the risk of non-compliance (TNC). k The Bayesian estimation is compatible. Initially, the risk of non-compliance was mitigated by incorporating the first actual measurements taken on the first batch of workpieces. k The estimates become increasingly accurate; then, Brownian motion drift with each update offsets this improvement in accuracy, leading to increased risk of non-compliance (TNC). k The estimated state is close to stable.

[0071] Details of an example of the decision criteria used in decision step 16 will now be given to determine whether workpiece k must be effectively inspected, i.e., whether it is necessary to proceed to measurement step 20.

[0072] The determination criteria include the verification of the estimated non-compliance risk TNC. k Below a fixed threshold F α (i.e., verify TNC) k <F α The first criterion is that the threshold depends on the acceptable level of non-compliance risk α, but not on the number of items inspected, i.e., not on k. Therefore, the fixed threshold F... α It can be a multiple of α, i.e., F α = m × α, where m is an appropriately chosen actual value. Therefore, the first criterion can be written as TNC. k <m×a α For example, a value of m=10 can be retained. Therefore, a fixed threshold F α This represents the maximum non-compliance risk considered acceptable on a single workpiece.

[0073] By providing a first criterion, once the probability law indicates that the non-compliance risk of workpiece k is too high, an effective inspection of workpiece k is performed. This helps to make inspection method 10 robust to drift in production.

[0074] The judgment criteria also include the verification of the estimated non-compliance risk TNC. k The value of VNC is a function of the estimated non-compliance risk of the workpiece to which the judgment criteria have already been applied. k Below the threshold Vk，α (i.e., verify VNC) k <V k，α The second criterion. Threshold V k，α It is a function of the number of the workpieces. In other words, the threshold V k，α It depends not only on the acceptable level of non-compliance risk α, but also on the number of items inspected, i.e., k. Therefore, the threshold V k，α It can be a multiple of α and k, i.e., V k，α = m′×k×α, where m′ is an appropriately chosen actual value. For example, a value that can be retained is V. k，α = k×α.

[0075] VNC value k It can be a weighted sum of these non-compliance risks, i.e. Where the value a j These are actual values. Therefore, the second criterion can be written as...

[0076] Preferably, the value VNC k This is only the estimated non-compliance risk TNC k The estimated risk of nonconformity for workpieces that have already met the judgment criteria (i.e., if characteristic X has been measured on workpiece j, then a) j A function that is equal to 0. In this way, the value VNC is... k This represents the cumulative non-compliance risk associated with the fact that a certain number of previous workpieces (k) were not effectively inspected. Additionally, in this case, the value VNC... k For example, the non-compliance risk of each workpiece that has already met the judgment criteria can be considered equally, i.e., if characteristic X has already been measured on workpiece j, then a j =0, otherwise a j =1. In this way, the value VNC k The accumulated non-compliance risk is related to the number of workpieces k that have not been effectively inspected before. However, for value a... j It is possible to apply different weights to a = 0, for example by providing a lower value as j decreases. j The value, even provides zero for all values ​​j less than a certain integer.

[0077] By providing a second criterion, workpiece k is effectively inspected once the cumulative non-compliance risk of workpiece k and other workpieces that have not yet been effectively inspected exceeds a certain threshold. This helps make inspection method 10 even more robust to drift in production.

[0078] The inspection method 10 described above can be included in the production method 10P for monitoring workpieces. This monitoring method includes providing multiple workpieces k, where k is between 1 and Q, and checking the compliance of each workpiece k using the aforementioned inspection method 10. The iteration of the inspection method 10 for each workpiece k is as follows: Figure 1 The explanation is carried out through incremental step 32.

[0079] Inspection method 10 and / or monitoring method 10P can be performed through methods such as Figure 7 The inspection device 59 described herein is used to implement this. The inspection device 50 here includes a measuring machine 52, a command module 54, and a transmission device 56. The measuring machine 52 is configured to measure the values ​​of the characteristics of the workpiece 50, and the transmission device 56 is configured to transmit the values ​​measured by the measuring machine 52 to the command module 54 and transmit the instructions of the command module 54 to the measuring machine 52.

[0080] Command module 54 is configured to estimate the non-compliance risk of a characteristic based on a probability law associated with the characteristic, verify whether the estimated non-compliance risk meets the judgment criteria, and if so, declare that workpiece 50 is compliant with the characteristic; if not, command module 54 is configured to instruct measuring machine 52 to measure the value of the characteristic to determine whether the workpiece is compliant based on the measured value, and update the probability law associated with the characteristic based on the measured value.

[0081] Here, command module 54 has the physical architecture of a computer, such as Figure 7 The diagram illustrates the concept. Specifically, it includes a processor 62, a read-only memory 63, a random access memory 64, a non-volatile memory 65, and a communication device 66. The measuring machine 52 enables the command module 59 to perform measurements executed by the measuring machine 52 in response to received instructions from the command module 59. The command module 59 and the measuring machine 52 are linked, for example, via a digital data bus or a serial interface (e.g., a USB interface (Universal Serial Bus)) or via wireless communication known per se.

[0082] The read-only memory 63 of the command module 59 forms a recording medium according to the invention, which can be read by the processor 62 and on which a computer program conforming to the invention is recorded, the computer program including implementation of the previously referenced... Figure 1 Instructions for each step of the inspection and / or monitoring method of the present invention.

[0083] The computer program defines, in an equivalent manner, the functional modules of the command module 59 that can implement the steps of the inspection method 10.

[0084] Now for reference Figures 2 to 4D A description is given of a method 110 for checking the compliance of a workpiece according to a second embodiment.

[0085] In this second embodiment, the workpiece k has p characteristics X. 1 ,…,X p , where p is an integer of at least 2. In the remainder of this document, subscripts with the letter k indicate that the workpiece k for which compliance is verified during inspection method 110, and superscripts with the number i indicate that the p characteristics X of the workpiece k for which compliance is verified during inspection method 110. 1 ,…,X p It relates to the i-th characteristic.

[0086] In one example, there are p features X 1 ,…,X p Each of these is a dimension value measured on workpiece k.

[0087] Figure 2 The steps of inspection method 110 are illustrated schematically.

[0088] exist Figure 2 In this process, steps similar to those in the first embodiment are accompanied by... Figure 1 The same reference numerals are used in the figures, but the reference numerals are increased by 100.

[0089] Therefore, inspection method 110 includes the steps of providing workpiece k (step 112), estimation (step 114), verification (step 116), inspection (step 122), and declaring workpiece k for characteristic X. i Regarding compliance steps 118, declare the workpiece k for characteristic X. i Regarding non-compliant steps 119 and incremental steps 132.

[0090] These steps are similar to those in the first embodiment, except that they are for feature X. 1 ,…,X p Each of them, rather than a single feature X, is executed.

[0091] In fact, feature X 1 ,…,X p It is possible to perform the same measurement operation on workpiece k. For example, when p characteristics X 1 ,…,X p When the dimensions are measured on workpiece k, these dimensions can depend on the measurements of the same geometric elements on workpiece k, especially when p is high.

[0092] Taking this situation into account, preprocessing steps are performed before measuring workpiece k. These steps are... Figure 3 The block diagram is described in detail.

[0093] When at least one of the p characteristics does not meet the judgment criterion (step 116), Figure 2 When ), method 110 (marked R1, Figure 2 and 3 Proceed to step 117A, where the characteristic X that does not meet the decision criterion is defined. 1 ,…,X p A list.

[0094] Method 110 then proceeds to step 117B, where a list of measurement operations to be performed on workpiece k is defined. This list of measurement operations is based on the list from step 117A, and includes p characteristics X. 1 ,…,X p Each of them is associated with at least one measurement operation M that can be performed on workpiece k. 1 ,…,M n The associated dependency tree D. The dependency tree D specifies the characteristic X to be examined. 1 ,…,X p The value is determined beforehand and therefore can be stored in memory. Note that n is not necessarily equal to p.

[0095] Method 110 then proceeds to step 117C, where a list of additional characteristics of workpiece k is defined. This list of additional characteristics includes at least some, and preferably all, of the characteristics that can be measured when performing the measurement operations defined in step 117B, regardless of the fact that these additional characteristics may already meet the decision criteria.

[0096] Method 110 then proceeds to measurement step 120A, where the value of each characteristic in the list defined in steps 117A and 117C is measured.

[0097] Once measurement step 120A is completed, method 120( Figure 2 and Figure 3 The markers R2 and R3 in the table will move each characteristic measured at step 120A to check 122 and update step 124.

[0098] In this way, method 110 benefits from the data provided by the measurement of additional characteristics, even if the decision criteria have indicated that its measurement is not necessary. Therefore, method 110 derives maximum benefit from the measurement results, and the decision criteria indicate the performance of the measurement results. This is particularly advantageous because measurement operations (such as the measurement of geometric elements on workpiece k) are time-consuming. Furthermore, method 110 incorporates more information provided by the measurement operations, which helps to make it more robust.

[0099] To facilitate understanding of the above content, the following is provided: Figures 4A to 4D Explanation of actual examples in the explanation.

[0100] exist Figure 4A In the example, as a very simplified one, workpiece 2 consists of a disk with eight drill holes 2-0, 2-1, 2-2, 2-3, 2-4, 2-5, 2-6, and 2-7. Figure 4A It also explains an example of determining the geometry of workpiece 2 according to standards ISO 8015 and ISO 1101.

[0101] Figure 4B An example of dependency tree D is explained. In this example, it focuses only on verifying the geometry of boreholes 2-0 and 2-1, provided that this example can be easily generalized to verifying the geometry of boreholes 2-0 through 2-7, and then further generalized to the geometry of more complex parts. In this example, since the goal is to check the geometry of boreholes 2-0 and 2-1, four properties X are defined in Table 1 below. 1 X 2 X 3 and X 4 .

[0102] [Table 1]

[0103] Characteristic marking Meaning of characteristics <![CDATA[X 1 ]]> 2-0 Drilling Diameter <![CDATA[X 2 ]]> 2-0 Drilling position <![CDATA[X 3 ]]> 2-1 Drilling Diameter <![CDATA[X 4 ]]> 2-1 Drilling location

[0104] For reference Figure 4A and 4B These characteristics X will be understood. 1 X 2 X 3 and X 4 It depends on reference plane A, reference cylinder B, and reference plane C. Therefore, characteristic X 1 X 2 X 3 and X 4 The following four measurement operations M can be used as defined in Table 2. 1 M 2 M 3 and M 4 Come and check.

[0105] [Table 2]

[0106]

[0107]

[0108] As mentioned above, in the specified characteristic X 1 X 2 X 3 and X 4 The dependency tree D was previously determined.

[0109] Figure 4C and 4D Is with Figure 4BA similar diagram illustrates an example of the application of the dependency tree D.

[0110] In these figures, it is assumed that different decision criteria indicate that only characteristic X needs to be measured. 2 The value (location of borehole 2-0) (in step 116 for characteristic X) 2 Yes or no), without needing to measure characteristic X. 1 X 3 and X 4 The value (in step 116 for characteristic X) 1 X 3 and X 4 (Yes). Therefore, the list of characteristics defined in step 117A contains only characteristic X. 2 This feature is in Figure 4C and 4D There is shadow, other features are Figure 4C It is surrounded by a dotted line.

[0111] Using the dependency tree D, feature X can be easily understood. 2 With measurement operation M 1 M 3 and M 4 Related. Therefore, the list of measurement operations defined in step 117B includes measurement operation M. 1 M 3 and M 4 These in Figure 4C and 4D There is a shadow in the image; measurement operation M 2 exist Figure 4C and 4D It is surrounded by a dotted line.

[0112] This will now be easy to understand—because the list of measurement operations defined in step 117B includes measurement operation M containing the measurement of borehole 2-0. 1 —This measurement operation M 1 It also allows obtaining information about the borehole diameter 2-0 (i.e., about characteristic X). 1 The information includes property X. As a result, in this example, the list of additional properties defined at step 117C includes property X. 1 And with the already targeted feature X 1 The facts verified the determination criteria were irrelevant. Figure 4D The image illustrates this point using shadows and dashed lines.

[0113] Subsequently, in measurement step 120A, measurement operation M is performed. 1 M 3 and M 4 Measurement characteristic X 1 and X 2The value of. The following steps of method 110 are then performed as described above.

[0114] In the first and second embodiments, the compliance of the workpieces is checked in the order in which they are provided, which is equivalent to incrementing k by 1 at incrementing steps 32 and 132. However, in practice, it is possible to check the workpieces in a different order than the order in which they are provided, especially considering the physical organization of the checkpoints designed to perform compliance checks. Therefore, a description of an embodiment that takes this possibility into consideration is now given. Figures 5A to 5C The third embodiment of the compliance inspection method has been explained.

[0115] exist Figure 5A In this context, steps similar to those in the first embodiment are performed by... Figure 1 The same reference numerals are used to indicate this, except that the reference numerals are increased by 200.

[0116] Therefore, inspection method 210 includes the steps of providing workpiece k (step 212), estimation (step 214), verification (step 216), measurement (step 220), inspection (step 222), updating (step 224), workpiece k being declared compliant with characteristic X (step 218), and workpiece k being declared non-compliant with characteristic X (step 219).

[0117] As previously stated, in method 210, the compliance of the workpieces is checked in a different order than the order in which they were provided. Therefore, in increment step 232A, k is not incremented by +1, but by a non-zero integer b, which can be positive or negative.

[0118] Figure 5B A practical example is provided where the workpieces are inspected in an order different from the order in which they were provided. Figure 5B In the diagram, the compliance check sequence 900 is illustrated by check chains C1 to C7, and the ranking of the workpiece verified in each check is indicated by the numbers on each of checks C1 to C7. Therefore, at check C1, workpiece number 1 is checked; at check C2, workpiece number 2 is checked; at check C3, workpiece number 5 is checked; and so on. Figure 5A In incremental step 232A, for example, when inspection C3 is reached, k increases by b = +3 (because it moves from workpiece 2 to workpiece 5); when inspection C5 is reached, k increases by b = -3 (because it moves from workpiece 6 to workpiece 3).

[0119] When inspection method 210 and monitoring method 210P are performed, compliance inspection sequence 900 is stored in database DB. Database DB may also store the results of measurements performed at steps 213M and 220.

[0120] After step 212 of providing workpiece k, method 210 moves to step 213A, where compliance check sequence 900 is read from database DB.

[0121] Following step 213A, method 210 moves to step 213B, where the workpiece k being inspected is checked in reverse order of the order in which the workpieces were provided. "Checking in reverse order" means that the workpiece k being inspected is ranked lower than the workpiece preceding it. This is equivalent to step 213B, which determines whether b is greater than 1.

[0122] If condition b>1 is not true (No at step 213B), then the rank of the workpiece k being inspected is lower than that of the workpieces preceding it in inspection sequence 900. Still using... Figure 5B Taking inspection sequence 900 as an example, at inspection C5, b = -3, and since workpiece number 6 was previously inspected at inspection C4, workpiece number 3 is inspected in reverse order. In this case, method 210 moves to step 213M, where the values ​​of all characteristics of workpiece k are measured, regardless of the estimated non-compliance risk of these characteristics. In other steps not illustrated in the figure, workpiece k is determined and declared compliant based on the values ​​measured at step 213M, in a manner similar to steps 222, 218, and 219. Additionally, in step 224M, the probability law L is updated only by applying a drift that follows only Brownian motion. k In order to obtain the updated probability law L k+1 .

[0123] If condition b > 1 is true (yes at step 213B), then the rank of the workpiece k being inspected is higher than that of the workpieces preceding it in inspection sequence 900. Still using... Figure 5B Taking inspection sequence 90 as an example, at inspection C3, b = +3 because workpiece number 2 was previously inspected at inspection C2. In this case, method 210 moves to step 213C, where, based on the database DB, a list of all workpieces ranked j is defined such that j < k and workpiece ranked j has not yet undergone an inspection operation.

[0124] After step 213C, method 210 moves to step 214 to estimate the non-compliance risk (TNC) of the inspected workpiece k. k .

[0125] After step 214, method 210 moves to step 216, which is the verification step of the decision criteria.

[0126] The determination criteria include the first criterion described above in conjunction with the first embodiment. The first criterion will not be described in detail again.

[0127] The determination criteria also include a second criterion, which is similar to the criteria described above in conjunction with the first embodiment, but modified to take into account the fact that the workpiece being inspected is ranked higher than the workpiece being inspected in inspection sequence 900.

[0128] More specifically, the second criterion verifies VNC′ k <V k′，α ,in V k，α = m′×k′×α, where k′ is the value that has been checked using monitoring method 210P and in VNC′ k The calculation determines the number of workpieces in the process, and the summation over workpiece j is performed only on workpieces that have already been inspected in monitoring method 210P. k′ is obtained from the reference database DB along with a list of workpieces that have already been inspected in monitoring method 210P.

[0129] After step 216, method 210 proceeds to the same steps as in the first embodiment, and therefore will not be described in detail hereafter.

[0130] Figure 5C Is with Figure 5B Similar to the schematic diagram in the diagram, and based on the diagram in the diagram. Figure 5B The same inspection sequence 900 is explained according to the ranking order of the inspected workpieces, except that the inspection of C1 to C7 is not in the order of inspection of C1 to C7.

[0131] like Figure 5C As explained, at the moment of inspection C3 of workpiece 5, information about workpieces 3 and 4 is not yet available. In step 216, at the moment of inspection C3 of workpiece 5, assuming that the judgment criteria have been verified for workpieces 3 and 4, the judgment criteria are determined, such as... Figure 5C As shown by mark 910′ in the figure.

[0132] Furthermore, since workpieces 3 and 4 are inspected in reverse order, all characteristic values ​​of these workpieces are measured regardless of the estimated non-compliance risk associated with those characteristics. Figure 5C As shown by mark 910 in the figure.

[0133] It should be noted that in the third embodiment, the information provided by the measurements performed on workpieces 3 and 4 at inspections C5 and C6 was not taken into consideration in inspection 7 and subsequent inspections.

[0134] Figures 6A to 6C A fourth embodiment of the compliance inspection method has been explained. Unlike the third embodiment, this fourth embodiment allows for the incorporation of information provided by measurements performed on workpieces inspected in reverse order.

[0135] exist Figure 6AIn this context, steps similar to those in the third embodiment are performed by... Figure 5A The same reference numerals are used to indicate this, except that the reference numerals are increased by 100.

[0136] Therefore, inspection method 310 includes the steps of providing workpiece k (step 312), estimation (step 314), verification (step 316), measurement (step 320), inspection (step 322), updating (step 324), declaring workpiece k compliant with characteristic X (step 318), and declaring workpiece k non-compliant with characteristic X (step 319).

[0137] Inspection method 310 also includes steps 313A, 313B, 313C, 313M, and 324M, which are similar to steps 213A, 213B, 213C, 213M, and 224M of inspection method 210. Therefore, these steps will not be described in detail again. However, after step 313M, the results of the measurement performed in step 313M are recorded in the measurement database DBMV, which records all previously measured values ​​of characteristic X. The measurement database DBMV is typically contained within a database DB.

[0138] However, the difference between inspection method 310 and inspection method 210 is that if condition "b > 1" is not true ("No" at step 313B), then method 310 moves to step 313M.

[0139] Furthermore, if condition b > 1 is true (yes at step 313B), then method 310 moves to step 313C, and then to step 313D.

[0140] In step 313D, the method reads the measurement database DBMV and updates the probability law L associated with characteristic X based on the measured values ​​input to the measurement database DBMV. k .

[0141] After step 313D, method 310 moves to step 314 to estimate the non-compliance risk (TNC) of the inspected workpiece k. k After step 314, method 310 proceeds to step 316 to verify the decision criterion. Step 316 is the same as step 216 and therefore will not be described in detail again.

[0142] After step 316, method 310 moves to the same steps as in the third embodiment, and therefore will not be described in detail again, but when measuring the value of characteristic X at measurement step 320, method 310 also includes step 340 of adding the measured value to the measurement database DBMV.

[0143] Figure 6B Reproduced with Figure 5B The same inspection sequence 900, and Figure 6C Is with Figure 6B A similar schematic diagram to the inspection sequence 900 in the diagram and based on Figure 6B The inspection sequence 900 is explained according to the ranking order of the inspected workpieces, except that inspections C1 to C7 are not explained in the order of inspections C1 to C7.

[0144] like Figure 6C As explained in the third embodiment, at inspection C3 of workpiece 5, information regarding workpieces 3 and 4 is not yet available. In step 316, at inspection C3 of workpiece 5, assuming the decision criteria have been verified for workpieces 3 and 4, the decision criteria are determined, such as... Figure 6C As shown by mark 910' in the text.

[0145] Furthermore, since the values ​​that can be measured for inspections C3 and C4 of workpieces 5 and 6, and for inspections C5 and C6 of workpieces 3 and 4, are stored in the measurement database DBMV, these values ​​are read into the database DBMV at inspection C7 of workpiece 7, as shown below. Figure 6C As indicated by mark 910 in the diagram. This makes it possible to update the probability law associated with characteristic X before estimating the non-compliance risk of workpiece number 7. Since this estimation is based on more measurement data points, it is more reliable. As can be seen from the preceding text, method 310 is more robust to drift in production.

[0146] although Figure 5A and 6A Although not explained in the text, inspection methods 210 and 310, as well as the corresponding monitoring methods 210P and 310P, can obviously be extended to the following situation, where not only is one characteristic X of workpiece k checked, but also p characteristics X of workpiece k are checked. 1 ,…,X p As in the second embodiment. In this case, inspection methods 210 and 310 include steps similar to steps 117A, 117B, 117C and 120A, with other steps adapted as in the second embodiment.

[0147] Furthermore, the inspection device 59 can obviously be configured to implement the methods of the second, third, and fourth embodiments by appropriately adapting the command module 54 and / or the computer program executed by the module.

[0148] Although the invention has been described with reference to specific embodiments, modifications may be made to these embodiments without departing from the general scope of the invention as defined in the claims. Specifically, features of various described / mentioned embodiments may be combined in additional embodiments. Therefore, the description and drawings are to be interpreted as illustrative rather than restrictive.

Claims

1. A method for checking the compliance of a workpiece (k) having at least one property (X), said method comprising the steps of: - estimating a risk of non-compliance (R k ) of a property (X) based on a probability law (L ) associated with said property (X); and - verifying whether the estimated non-compliance risk (R ) fulfils a decision criterion, and if so, declaring the artefact compliant for the property (X); if not, measuring the value of the property, determining whether the artefact is compliant based on the measured value, and updating the probability law (L k ) associated with the property (X) based on the measured value.

2. The method of claim 1, wherein, The determination criteria include the verification of the estimated non-compliance risk therefrom. ) below a fixed threshold ( The first principle.

3. The method of claim 1 or 2, wherein, said decision criterion comprises a second criterion of verifying that a value (VNC k ) as a function of the estimated risk of non-compliance and of the estimated risk of non-compliance of the workpieces for which said decision criterion has previously been applied is lower than a threshold (V k,α ) as a function of the number of said workpieces.

4. The method of claim 3, wherein, The value (VNC k ) is simply a function of the estimated non-compliance risk and the estimated non-compliance risk of the artifacts that have previously satisfied the decision criteria.

5. The method of claim 1, wherein, The probability law (L k ) is updated using a sequential Monte Carlo method.

6. The method of claim 1, wherein, The workpiece has p characteristics ( ), wherein p is an integer at least equal to 2 and the steps of estimation of the risk of non-compliance and verification of the decision criteria are performed for each of the p characteristics.

7. The method of claim 6, wherein, - defining a list of the properties of the workpiece (k) that can be measured by performing a list of measurement operations on the basis of the list of measurement operations; - defining a list of the properties of the workpiece (k) that can be measured by performing a list of measurement operations on the basis of the list of measurement operations; - the list of measurement operations is defined on the basis of the list of the at least one characteristic not satisfying the decision criterion and of a dependency relation tree (D) associating each of the p characteristics ( ) with at least one measurement operation (M) that can be performed on the workpiece. ) - measuring the value of each of the at least one property that does not satisfy the decision criterion and of each of the properties of the additional list of properties; and - based on the values thus measured, each of the probability laws associated with each of the at least one property that does not satisfy the decision criterion is updated and each of the probability laws associated with the additional properties is updated.

8. A method for monitoring the production of workpieces comprising the steps of: - providing a plurality of workpieces (k) of the same type and each having at least one property (X); and - checking the compliance of each of the workpieces with the checking method according to any one of claims 1 to 7. The compliance of each of the workpieces (k) is checked in a different order from the order in which they are provided and the values of the at least one property of each of the workpieces being checked are measured in the opposite order from the order in which they are provided, independently of the estimated risk of non-compliance of each of the at least one property.

9. The method of claim 8, wherein, The compliance of each of the workpieces is checked in a different order from the order in which they are provided, all the measured values are placed in a memory and, before estimating the risk of non-compliance of a workpiece, the probability laws associated with each of the at least one property of the workpiece are updated on the basis of the measured values placed in the memory.

10. The method of claim 8, wherein, 12. A program comprising instructions for performing the steps of the checking method according to any one of claims 1 to 7 when the program is executed by a computer or microprocessor.

11. An apparatus for inspecting the compliance of a workpiece having at least one characteristic (X), comprising a measuring machine configured to measure a value of said at least one characteristic, a command module, and a transmission device configured to transmit measured values ​​from said measuring machine to said command module and to transmit instructions from said command module to said measuring machine, said command module being configured to apply a probability law (L) associated with the characteristic (X). k Estimate the non-compliance risk of characteristic (X) based on ) The command module verifies whether the estimated non-compliance risk meets the judgment criteria, and if so, declares the workpiece compliant with the characteristic; if not, the command module is configured to instruct the measuring machine to measure the value of the characteristic to determine whether the workpiece is compliant based on the measured value, and to update the probability law (L) associated with characteristic (X) based on the measured value. k ).

13. A program comprising instructions for performing the steps of the monitoring method according to any one of claims 8 to 10 when the program is executed by a computer or microprocessor. ​

Images