JP2013015964A - Pass rate estimation device, pass rate estimation method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a pass rate estimation device capable of reducing an amount of a material and the like produced excessively, in quality control in which a plurality of inspections are conducted.SOLUTION: A pass rate estimation device inspects inspection objects for each lot by a first inspection method to estimate the possibility that rejected lots could pass a second inspection method different from the first inspection method if they were inspected by the second inspection method. The pass rate estimation device comprises: measurement value acquisition means that acquires measurement values of a current lot; past measurement value storage means DB5 and DB4 that store past measurement values and re-inspection pass rates of inspection items for past lots; deviation value calculation section 11 that calculates the deviations of the measurement values of the current lot with respect to standard values; past deviation storage means DB2 that stores past deviations of the past measurement values; a similarity calculation section 12 that calculates the magnitude of the difference between the current and past deviation for each inspection item; and re-inspection pass rate estimation section 14 that reads out the re-inspection pass rates of the past lots each having the difference within a prescribed value for each inspection item, to estimate a re-inspection pass rate of the current lot.

Description

  The present invention is capable of passing inspection when an inspection object is inspected for each lot by the first inspection method, and a lot that does not pass the inspection is inspected by a second inspection method different from the first inspection method. The present invention relates to a pass rate estimation apparatus that estimates sex.

  In the product manufacturing process, quality control of materials and parts is performed. In quality control, an inspection device or the like determines whether or not quality is satisfied by measuring performance to be satisfied by materials and parts and comparing it with some standard value. Since materials and parts that do not satisfy quality are generally discarded, it is important to reduce the number of defective products. However, since the defective rate itself is not so large, it is also required to improve the quality control.

  For example, a method for grasping the tendency of occurrence of defective products has been considered (for example, see Patent Document 1). Patent Document 1 discloses a quality information collection / analysis system that displays in a table which device is used when a defective part is assembled in a quality management analysis period, and what part shape is assembled. Has been. According to such a method, for example, when a failure occurs when it takes longer than usual between the A device and the B device, it is possible to predict that there are many failures between the facilities.

  By the way, there is a method (secondary evaluation method) for inspecting a lot that has not passed a plurality of times in order to reduce the defect rate and improve the quality control efficiency.

FIG. 13 is an example of a diagram schematically illustrating conventional quality control.
First inspection: A simple inspection using strict standard values is performed, and lots that do not pass the inspection are sent to the second inspection.
Second inspection: Inspection with higher measurement accuracy than the first inspection is performed on all materials and parts to inspect whether the standard value is satisfied.

  By performing the inspection multiple times in this way, the inspection time can be shortened for materials and parts that have passed the first inspection, and the materials and parts that have passed the second inspection can contribute to a reduction in the defect rate. In the figure, the inspector took out a part of the lot this time and obtained the actual value of dielectric constant “108”, but did not see the standard value. For this reason, the person in charge of inspection stores the lot this time and re-inspects it later (for example, one week to six months later). Depending on the result of the re-inspection, for example, a final inspection result is obtained such that 100% of the current lot has passed, 80% has passed, or all of the lots are discarded because the acceptance rate is low.

  However, the secondary evaluation method has a problem that it cannot be determined whether or not a lot that has not passed the first inspection can be used for a product during the period until the second inspection. In the production of products, the number of products produced by a line or the like is determined by the production plan, so it is necessary to prepare materials and parts necessary for the production of the products. If it is not determined whether or not the product can be used, it becomes necessary to create or order all the lots that have not passed the first inspection. For this reason, there is a problem that extra materials and parts that are created cause an increase in cost, and inventory management of extra materials and parts that are created is necessary.

  In view of the above problems, an object of the present invention is to provide a pass rate estimation device that can reduce the amount of extra materials and parts that are created in quality control in which inspection is performed a plurality of times.

  The present invention inspects an inspection object for each lot by the first inspection method, and passes a lot that did not pass the inspection by a second inspection method different from the first inspection method. A pass rate estimation device for estimating the possibility, a measurement value acquisition means for acquiring a measurement value of the current lot in one or more inspection items, and a past that did not pass the inspection by the first inspection method The past measurement value of the inspection item of the lot and the past measurement value storage means for storing the re-inspection pass rate that the past lot has passed by the second inspection method, and standardized by the past measurement value for each inspection item The deviation amount calculating means for calculating the deviation amount of the measured value of the current lot with respect to the standard value, the past deviation amount storing means for storing the past deviation amount of the past measured value, and the deviation for each inspection item Said past deviation to quantity From the difference calculation means for calculating the difference magnitude, and the past measurement value storage means, the retest pass rate of the past lot in which the difference for each inspection item falls within a predetermined value is read, and Re-inspection pass rate estimation means for estimating a re-inspection pass rate.

  In the quality control in which the inspection is performed a plurality of times, it is possible to provide a pass rate estimation device capable of reducing the amount of extra materials and parts created.

It is an example of the figure explaining the outline of the inspection method of this embodiment. It is an example of the figure explaining a secondary evaluation method. It is an example of the schematic block diagram of a test | inspection system. It is an example of the hardware block diagram of a pass rate estimation apparatus. It is an example of the functional block part of a pass rate estimation apparatus. It is an example of the flowchart figure which shows the whole flow of the procedure which a pass rate estimation apparatus estimates a pass rate. It is an example of the flowchart figure explaining the calculation procedure of a deviation value. It is an example of the figure explaining the specific example of calculation of a deviation value. It is a figure explaining an example of the calculation result of the deviation value of a lot this time. It is an example of the flowchart figure explaining the calculation procedure of similarity. It is an example of the figure explaining similarity determination. It is an example of the figure explaining estimation of the acceptance rate of this lot by a weighted average. It is an example of the figure which illustrates the conventional quality control typically.

Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.
FIG. 1 is an example of a diagram illustrating an outline of estimation of a pass rate by characteristic re-examination in the present embodiment. The person in charge of inspection inspects inspection objects such as materials and parts by operating a predetermined procedure using a predetermined apparatus. The inspection items are, for example, dielectric constant, magnetic property, cohesion degree, fluidity and the like. Here, since inspection objects are often produced in lots (a production form that is produced in a fixed quantity unit), a part of the inspection objects produced in lots is extracted and inspected based on ISO or JIS standards. Inspection is performed.

  A measurement value (hereinafter referred to as an actual value) such as a physical property obtained by inspecting an inspection object for each inspection item by the measuring device is compared with a standard value for each inspection item, and pass / fail is determined on a lot basis. Lots determined to be acceptable are used for manufacturing the next process (including product manufacturing processes), but lots determined to be unacceptable are sent for re-inspection. For this reason, lots that were rejected after the first inspection in the past (hereinafter referred to as the previous lot) are now subject to the secondary evaluation, rather than the lot that was inspected for the first time (hereinafter referred to as the current lot). It has been re-inspected according to the method, and the re-inspection pass rate has already been calculated.

The acceptance rate estimation device of the present embodiment predicts the probability that the rejected lot will pass the re-inspection, thereby accurately predicting the production amount of the inspection object of the product produced systematically, and Reduce administrative costs.
(1) The pass rate estimation device calculates a deviation value of the actual value with respect to the standard value for each inspection item of the current lot using the past actual value of the rejected lot. The deviation value becomes larger as the actual value is larger than the standard value, and becomes smaller as the actual value is smaller than the standard value.
(2) The pass rate estimation apparatus compares the deviation value for each inspection item of the current lot with the deviation value of the actual value of the past lot, and calculates the similarity between the current lot and each past lot. That is, in the past lot, a lot whose deviation value of each inspection item is similar to the current lot has a high similarity to the current lot.
(3) The acceptance rate estimation apparatus reads the determination criterion, compares the similarity between the current lot and the past lot, and identifies similar past lots. That is, the past lots having the same tendency as the current lot and the deviation values are narrowed down.
(4) The pass rate estimation device reads a retest pass rate of similar past lots, and estimates the retest pass rate of the current lot using the retest pass rate.

  Thus, by estimating the re-inspection pass rate of the current lot from the similarity of the deviation values, the re-inspection pass rate of the current lot can be estimated with high accuracy. Since the production manager can theoretically determine that it is sufficient to additionally produce materials and parts corresponding to the retest failure rate by “1-retest pass rate”, the cost of extra materials and parts created is reduced. It is possible to reduce the cost of inventory management for extra materials and parts.

[Secondary evaluation method]
First, a two-dimensional evaluation method for inspecting a lot that has not passed the first inspection twice in order to reduce the defect rate while improving the quality control will be described.

  The two-dimensional evaluation method includes, for example, a “test method change method” in which the test method is changed between the first time and the second time, and an “upper / lower limit sample evaluation method”. The test method changing method is an inspection method for performing the second inspection in accordance with the product standard because the first inspection simplifies the method of measuring the actual value compared to the product standard. The upper and lower limit sample evaluation method is an inspection method in which inspection is performed using two standards, that is, a manufacturing standard based on upper and lower standard samples and a product standard that is more relaxed than that.

  FIG. 2A is an example for explaining a test method changing method. For example, in the toner inspection, inspection items such as loose apparent density, color analysis (L, a *, b *), coarse particle amount, fixing lower limit (smear), fixing upper limit (hot), and amount of loose aggregates are test methods. Inspected by the modified method. The smear property in this will be described as an example. The smear property means that image quality is deteriorated because the toner fixed on the paper is not sufficiently fixed. For example, it means that an image on one side is blurred or smeared when an image is formed on the other side with toner fixed on one side. The inspection device measures physical properties (such as viscosity) at the lower limit of fixation (temperature) for smear inspection. If the actual value is within a predetermined range with respect to the standard value, the inspection item can be passed.

  The first inspection can pass the inspection if the standard value is satisfied at a temperature narrower than the temperature range of the product standard (for example, 140 degree one point). By doing so, the inspection time can be shortened. On the other hand, lots that cannot pass the first inspection are inspected in the temperature range of the product standard to see whether the actual value satisfies the standard value. In this way, by inspecting at different temperatures within the temperature range of the product standard, the inspection accuracy can be improved, and materials and parts that satisfy the product standard can be extracted from the entire lot.

FIG. 2B is an example of a diagram illustrating the upper and lower limit sample evaluation method. This inspection is effective when inspecting inspection items for which accurate measurement is difficult. For example, the following cases.
・ Inspection items that may cause variations in actual values depending on the condition of the measuring instrument ・ Inspection items that do not affect the quality control even if the decision is made to the upper and lower limit border products of the standard In FIG. Inspection items to which the evaluation method is applied include circularity, volume resistivity, dielectric constant, surface WAX amount, Si · Ti (content), and pressure pieces. In this upper and lower limit sample evaluation method, the inspection result (actual value) by a standard sample in which the actual value is guaranteed to be smaller than the upper limit of the standard value of the product standard is set as the upper limit standard value of the manufacturing standard. Similarly, an inspection result (actual value) by a standard sample in which the actual value is guaranteed to be larger than the lower limit of the standard value of the product standard is set as the lower limit of the manufacturing standard.

  In the first inspection, whether or not the inspection is passed is determined depending on whether or not the actual value of the current lot falls within the upper and lower limits. In this way, even if there are variations due to the condition of the measuring instrument, it can be inspected by comparison (relatively) with the standard sample.

  As shown in the figure, in the manufacturing standard applied to the first inspection, the standard value of the upper and lower standard samples is stricter than the product standard to prevent the outflow of defective products (determined with a margin). . In the second inspection, the product standard relaxed from the manufacturing standard becomes the standard value, and all the lots are inspected at this time. Therefore, materials and parts that pass the second inspection are extracted.

  In the two-dimensional evaluation method, the pass rate estimation method of the present embodiment estimates a pass rate at which a lot that has not passed the first inspection passes the second inspection.

[Configuration example]
FIG. 3 shows an example of a schematic configuration diagram of the inspection system 300 of the present embodiment. The inspection system 300 includes an inspection device 200 and a pass rate estimation device 100 connected via a network 301. Although it is not always necessary to be connected via the network 301, this makes it easy to transfer the actual value from the inspection apparatus 200 to the pass rate estimation apparatus 100. Further, the function of the pass rate estimation apparatus 100 may be mounted on the inspection apparatus 200.

  The network 301 is a network that is completed in-house such as an in-house LAN or a wide area LAN (WAN), but may be an IP-VNP (Virtual Private Network), the Internet VPN, the Internet, or the like. The latter configuration is effective when, for example, the department or company for inspecting the inspection object is separated from the department or company that receives the inspection object. Note that a telephone line may be included in a part of the network 301, and wired connection or wireless connection may be used.

  When the inspection object is produced by a predetermined quantity (that is, lot), the inspection apparatus 200 samples the predetermined number or quantity, inspects the inspection item, and the result value is a pass rate estimation apparatus. To 100. Since there are generally a plurality of inspection items, there are often a plurality of inspection devices 200. The inspection items include, for example, the items described with reference to FIG. 2, but an inspection useful for inspecting the inspection object is appropriately performed.

  FIG. 4 shows an example of a hardware configuration diagram of the pass rate estimation apparatus 100. The acceptance rate estimation apparatus 100 is a form of a computer. The pass rate estimation apparatus 100 includes a CPU 101, a RAM 102, a ROM 103, a storage medium mounting unit 104, a communication device 105, an input device 106, a drawing control unit 107, and an HDD 108 that are mutually connected by a bus. The CPU 101 provides various functions by reading an OS (Operating System) and a program from the HDD 108 and executing them, and comprehensively controls processing performed by the pass rate estimation apparatus 100.

  The RAM 102 is a working memory (main storage memory) that temporarily stores data necessary for the CPU 101 to execute the program. The ROM 103 stores a program (static input data) for starting up a BIOS (Basic Input Output System) and the OS. It is remembered.

  A storage medium 110 can be attached to and detached from the storage medium mounting unit 104, and a program recorded in the storage medium 110 is read and stored in the HDD 108. Further, the storage medium mounting unit 104 can also write data stored in the HDD 108 into the storage medium 110. The storage medium 110 is, for example, a USB memory or an SD card.

  The input device 106 is a keyboard, a mouse, a trackball, or the like, and accepts various operation instructions from a manufacturing manager of the acceptance rate estimation device 100.

  The HDD 108 may be a nonvolatile memory such as an SSD, and stores various data such as an OS, a program, and a standard value. The acceptance rate estimation apparatus 100 has a program 111 for estimating the acceptance rate of the current lot.

  The communication device 105 is a NIC (Network Interface Card) for connecting to a network 301 such as the Internet, and is, for example, an Ethernet (registered trademark) card.

  The drawing control unit 107 interprets a drawing command written in the graphic memory by the CPU 101 executing the program 111, generates a screen, and draws it on the display 109.

  FIG. 5 is an example of a functional block unit of the pass rate estimation apparatus 100. The pass rate estimation apparatus 100 includes a deviation value calculation unit 11, a similarity calculation unit 12, a similarity determination unit 13, and a pass rate estimation unit 14. In addition, the pass rate estimation apparatus 100 includes, in order to calculate the deviation value, the similarity, and the pass rate, the standard values for each inspection item, the deviation value of the past lot actual value, the determination criterion, the past, Each data includes the re-inspection pass rate of lots and the actual values of past rejected lots. DBs 1 to 5 are mounted on the HDD 108. In addition, the pass rate estimation apparatus 100 should just be able to read each DB1-5 via the network 301, and the pass rate estimation apparatus 100 does not need to have DB1-5. DB1 to 5 are schematic diagrams, and DB1 to 5 may be consolidated into five or less DBs.

The details of each function in FIG. 5 will be described later. First, each function and procedure will be briefly described with reference to the overall flow. FIG. 6 is an example of a flowchart illustrating an overall flow of a procedure for the acceptance rate estimation apparatus 100 to estimate the acceptance rate of the second inspection.
S10: The deviation value calculation unit 11 calculates the deviation value of the actual value of the current lot. That is, the standard value for each inspection item is read from DB1, and the past actual value of rejected lots is read from DB5, and the deviation value of the actual value of the current lot from the standard value is calculated. The deviation value of the actual value of the past lot has already been calculated. (The deviation value of the past inspection results is recalculated each time the number of lots that failed the inspection increases.)
S20: Next, the similarity calculation unit 12 reads the deviation value of the past lot actual value from DB2.
S30: The similarity calculation unit 12 calculates the similarity by comparing the deviation value of the actual value of the past lot with the deviation value of the current lot.
S40: The similarity determination unit 13 reads a determination criterion from the DB 3, and specifies a past lot whose similarity with the current lot is a predetermined value or more.
S50: The acceptance rate estimation unit 14 reads from DB4 the reexamination acceptance rate of the past lot whose similarity with the current lot is a predetermined value or more.
S60: The pass rate estimation unit 14 weights and averages the re-inspection pass rate of the past lot, and estimates the re-inspection pass rate of the current lot.

[Calculation of deviation value]
FIG. 7 shows an example of a flowchart for explaining the procedure for calculating the deviation value, and FIGS. 8 and 9 show examples of diagrams for explaining specific examples of calculation. The procedure of FIG. 7 is repeatedly executed for each inspection item.

  As shown in FIG. 8A, there are four inspection items: “dielectric constant”, “magnetic property”, “aggregation degree”, and “fluidity”. The standard value of dielectric constant is 105, the standard value of magnetic property is 2.5, the standard value of cohesion is 26, and the standard value of fluidity is 0.055. Past actual values are obtained for lots 1 to 3, and the actual value of lot 4 is obtained as the current lot.

  Lots 1 to 4 all failed in the first inspection of the secondary evaluation method. In addition, lots 1 to 3 have already been re-inspected, and “re-inspection pass rate” is described. The “inspection date” is the first inspection date.

  FIG. 8B is an example of a diagram illustrating a comparison between the actual value and the standard value of each lot regarding the dielectric constant. The actual value of each lot deviates from the standard value, but there are some that are likely to deviate depending on the inspection item, and some that are difficult to deviate. For this reason, at first glance, even if the amount of deviation is large, there are those that pass the re-inspection, and even if the amount of deviation is small, there are those that do not pass the re-inspection. In order to digitize this tendency, the deviation value calculation unit 11 calculates a standard deviation.

  As shown in FIG. 7, the deviation value calculation unit 11 reads the past lot actual value and past actual value that were rejected in the past from the DB 5 for each inspection item (initially dielectric constant) (S101). . Further, the deviation value calculation unit reads the standard values of the inspection items one by one (initially dielectric constant) from DB1. The reason why only the lots that have been rejected in the past lot are read out is because the lots that have passed are not re-inspected, and thus are not suitable as data for estimating the re-inspection pass rate.

  First, the actual value of the dielectric constant of the current lot is 108, the actual value of the dielectric constant of lot 1 is 102, the actual value of the dielectric constant of lot 2 is 160, and the actual value of the dielectric constant of lot 3 is 110.

Next, the deviation value calculation unit 11 calculates a standard deviation (S102). The formula for calculating the standard deviation is as follows.
Standard deviation = √ {Σ (each actual value−standard value) ² / the number of past lots that did not pass the inspection by the first inspection method}
Assuming that the actual values follow a normal distribution, about 2/3 of the actual values are included between the standard value ± standard deviation.

Next, the deviation value calculation unit 11 calculates the deviation value of the actual value of the current lot (S103). The formula for calculating the deviation value is as follows.
Deviation value = {(actual value−standard value) / standard deviation} +50
That is, even for inspection items whose actual values are likely to vary, the deviation amount is divided by the standard deviation, so that the deviation value is suppressed more than the deviation amount itself. For inspection items whose actual values are less likely to vary, the standard deviation also becomes smaller, so the deviation value tends to be suppressed more than the deviation amount itself.

  The deviation value calculation unit 11 writes the deviation value for each inspection item of the current lot in the HDD 108 or the like (S104). When the deviation value calculation unit 11 obtains the deviation value of the dielectric constant, the deviation value calculation unit 11 obtains the deviation values of magnetic property, cohesion degree, and fluidity.

  FIG. 9A shows an example of the calculation result of the deviation value of each inspection item of the current lot, and FIG. 9B shows an example of a graph showing the deviation value of the dielectric constant of each lot. Since the standard value (median value) of the deviation value is 50, the deviation from 50 indicates how much the deviation value of the dielectric constant of each lot is from the standard value.

(Similarity calculation)
FIG. 10 shows an example of a flowchart for explaining the similarity calculation procedure. This similarity is a similarity between the deviation value of each inspection item of the current lot and the deviation value of each inspection item of the past lots 1 to 3. That is, the similarity calculation unit 12 calculates the following three similarities for each inspection item.
Similarity between current lot (lot 4) and lot 1 Similarity between current lot (lot 4) and lot 2 Similarity between current lot (lot 4) and lot 3 Accordingly, the procedure of FIG. 10 is repeated for the number of past lots. .

  First, the similarity calculation unit 12 reads the deviation value of the dielectric constant of the current lot from the HDD 108 and the deviation value of each inspection item of the past lots 1 to 3 from the DB 2 (S301).

  The deviation value of the dielectric constant of the lot this time is 50.48, the deviation value of the dielectric constant of the lot 1 is 49.52, the magnetic deviation value of the lot this time is 49.94, the magnetic deviation value of the lot 1 is 50.30, The deviation value is 52.64, the deviation value of the cohesion degree of lot 1 is 51.06, the deviation value of fluidity of the current lot is 48.47, and the deviation value of fluidity of lot 1 is 50.96.

Next, the similarity calculation unit 12 calculates a standard similarity (S302). The standard similarity calculation formula is as follows.
Standard similarity = √ {Σ (deviation value of each result−deviation value of current lot) ² / the number of past lots that did not pass the inspection by the first inspection method}
As is clear from the numerator “deviation value of each track record—deviation value of the current lot”, the standard similarity represents the average variation of the deviation value of the past lot with respect to the deviation value of the current lot.

Next, the similarity calculation unit 12 calculates the similarity with the current lot for each inspection item (S303). The formula for calculating the similarity is as follows. This calculation is performed for each inspection item (dielectric constant, magnetic property, cohesion degree, fluidity).
Similarity = {(deviation value of past lot−deviation value of current lot) / standard similarity} +50
In other words, the difference between the deviation value of the current lot and the deviation value of a certain lot in the past is divided by the average variation of the deviation value to obtain the similarity. Since the difference between the deviation value of the current lot and the deviation value is obtained, the similarity is the difference between the deviation value of the previous lot and the deviation value of the current lot. As shown in this equation, the degree of similarity increases or decreases around 50, so the closer to 50, the higher the degree of similarity with the current lot.

  For example, the similarity between Lot 1 and the current lot is 48.93 for dielectric constant, 50.71 for magnetic, 48.82 for cohesion, and 50.93 for fluidity. Since any similarity is divided by the standard similarity, the variability of each inspection item is removed from the similarity. That is, if the dielectric constant and the degree of aggregation are the same, it can be evaluated that the dielectric constant and the degree of aggregation are the same and similar.

  When the similarity is obtained for each inspection item of lot 1 and each inspection item of lot 4, the similarity calculation unit 12 calculates the similarity between each inspection item of lot 2 and each inspection item of lot 4. This is repeated as many times as the number of lots subject to past reinspection.

Note that the similarity may be calculated by other calculation methods. Since each lot has a deviation value for each inspection item, if the set of deviation values is regarded as a vector, the Euclidean distance in the n-dimensional space can be obtained from the set of deviation values of each lot. For example, the Euclidean distance between lot 1 and lot 4 can be expressed as follows. Note that the smaller the Euclidean distance, the stronger the degree of similarity.
Euclidean distance = √ {(50.48-49.52) ² + (49.94-50.30) ² + (52.64-51.06) ² + (48.47-50.96) ² }

[Similarity judgment]
FIG. 11 is an example of a diagram illustrating similarity determination. The similarity determination is a process for extracting a past lot similar to the current lot. The concept of similarity determination is that the similarity is within the upper and lower limits centered on 50 in all inspection items. The upper and lower limits are, for example, 48 to 52 (48.00 or more and less than 52.00).

  Since the similarities of the inspection items of lot 1 and lot 4 are (48.93, 50.71, 48.82, 50.93), all similarities are within the upper and lower limits. Since the similarity between the inspection items of Lot 2 and Lot 4 is (59.30, 41.20, 48.96, 51.57), the dielectric constant and magnetic properties are not within the upper and lower limits. Since the similarity between the inspection items of lot 3 and lot 4 is (50.36, 50.48, 49.00, 51.00), all the similarities are within the upper and lower limits.

  The similarity determination unit 13 identifies lots 1 and 3 in which all inspection items fall within the upper and lower limits as past lots for estimating the pass rate. If there is no past lot in which all inspection items fall within the upper and lower limits, the similarity determination unit 13 may specify a past lot in which all the inspection items fall within the upper and lower limits except for one. This is because even if one inspection item does not fall within the upper and lower limits, the re-inspection pass rate may be close. By doing so, it is possible to easily obtain the re-inspection pass rate of the current lot.

  Further, instead of the number of inspection items, for example, it may be determined that past lots in which 90% of all inspection items fall within the upper and lower limits are similar.

[Weighted average per inspection month]
By the similarity determination, the lots 1 and 3 having similar deviation values obtained from the actual values of the respective inspection items could be extracted. Therefore, the re-inspection pass rate of these lots 1 and 3 is highly likely to be a reference in the current lot. For example, if the re-inspection pass rate for lots 1 and 3 is high, there is a rational for predicting that the re-inspection pass rate for the current lot is also high.

  However, if the reliability of the actual value itself is lowered, there is a possibility that the premise for using the re-inspection pass rate of the past lot for estimation of the re-inspection pass rate of the current lot may be lacking. For example, even if the inspection apparatus 200 is calibrated, it is affected by seasonal changes, operator changes, inspection apparatus 200 replacement, and the like. In addition, the possibility that the raw material, production factory, management method, and the like of the inspection object will change increases with the passage of time. Therefore, the pass rate estimation unit 14 according to the present embodiment weights the retest pass rate according to the number of days elapsed from the inspection date of the similar lots 1 and 3 to the inspection date of the current lot. That is, the weighting is increased as the elapsed days are shorter. For example, the weighting is as follows. In the past, a weight smaller than 0.1 (for example, zero, 0.05) may be given uniformly in the past, and a weight smaller than 0.1 and smaller as the number of elapsed days may be given.

0 to 1 month before inspection: 1
1-3 months before the test: 0.6
3-6 months before the test: 0.1

ただし、Ｗｉは重み付け、Ｘｉは過去ロットの再検査合格率、ｎは類似する過去ロットの数、である。

However, Wi is weighting, Xi is the re-examination pass rate of the past lot, and n is the number of similar past lots.

  FIG. 12A is an example for explaining the estimation of the re-examination pass rate of the current lot. As shown in FIG. 12A, it is assumed that the inspection date of lot 1 is 3 to 6 months before the inspection date of lot 4, and the inspection date of lot 3 is 0 to 1 month before the inspection date of lot 4. . The reinspection pass rate for lot 1 is 100%, and the reinspection pass rate for lot 3 is 80%.

FIG. 12B is an example of a diagram for explaining the calculation of the reinspection pass rate of lot 4. The weight of lot 1 is 1, the reinspection pass rate is 100%, the weight of lot 3 is 0.1, and the reinspection pass rate is 80%. For this reason, the reinspection pass rate of lot 4 is as follows.
Lot 4 re-inspection pass rate = (1 x 1.00 + 0.1 x 0.8) / (1 + 0.1)
= 0.9164 (= 91.64%)
Therefore, the acceptance rate estimation unit 14 estimates that the reinspection acceptance rate at which lot 4 passes by reinspection is 91.64%. From now on, the production manager can make a production plan that 100-91.64 = 8.36%, which is not enough for lot 4 to produce products, may be created in the future lot. Since it is sufficient to reproduce the minimum necessary amount in this way, it is possible to reduce the cost due to extra materials and parts, and to suppress the cost for inventory management of extra materials and parts.

DESCRIPTION OF SYMBOLS 11 Deviation value calculation part 12 Similarity calculation part 13 Similarity determination part 14 Pass rate estimation part 100 Pass rate estimation apparatus 200 Inspection apparatus 300 Inspection system

JP 2010-2111309 A

Claims (6)

The inspection object is inspected for each lot by the first inspection method, and the possibility of passing if the lot that did not pass the inspection is inspected by the second inspection method different from the first inspection method is estimated. An acceptance rate estimation device that performs
A measurement value acquisition means for acquiring a measurement value of the current lot in one or more inspection items;
The past measurement value of the past inspection item of the past lot that did not pass the inspection by the first inspection method, and the past measurement value that stored the re-inspection pass rate that the past lot passed by the second inspection method Storage means;
Deviation amount calculation means for calculating a deviation amount of the measurement value of the current lot with respect to a standard value, standardized by the past measurement value for each inspection item;
Past deviation amount storage means for storing a past deviation amount of the past measurement value;
Difference calculating means for calculating the magnitude of the difference of the past deviation amount with respect to the deviation amount for each inspection item;
Re-inspection pass rate estimation means for reading the re-inspection pass rate of the past lot in which the difference for each inspection item falls within a predetermined value from the past measurement value storage unit and estimating the re-inspection pass rate of the current lot When,
A pass rate estimation apparatus having The past measurement value storage means stores an inspection date according to the first inspection method for each lot,
The re-inspection pass rate estimation means estimates the re-inspection pass rate of the current lot by weighting the re-inspection pass rate according to the number of days elapsed from the inspection date of the previous lot to the inspection date of the current lot.
The acceptance rate estimation apparatus according to claim 1, wherein: The deviation amount calculating means has the following formula:
A = √ {Σ (past measurement value−standard value) ² / the number of past lots that did not pass the inspection by the first inspection method}
Deviation amount = {(measured value−standard value) / A} +50
To obtain the deviation amount from
The difference calculating means has the following formula:
B = √ {Σ (deviation amount of past lot−deviation amount of current lot) ² / number of past lots that did not pass inspection by the first inspection method}
Difference size = {(deviation amount of the previous lot−deviation value of the current lot) / B} +50
Calculating the magnitude of the difference from
The pass rate estimation apparatus according to claim 1 or 2, wherein The difference calculation means obtains a Euclidean distance between the set of deviation amounts of each inspection item of the current lot and the set of deviation amounts of each inspection item of the previous lot,
The re-inspection pass rate estimation means reads the re-inspection pass rate of the past lot that the Euclidean distance from the current lot falls within a predetermined value, and estimates the re-inspection pass rate of the current lot.
The pass rate estimation apparatus according to claim 1 or 2, wherein The inspection object is inspected for each lot by the first inspection method, and the possibility of passing if the lot that did not pass the inspection is inspected by the second inspection method different from the first inspection method is estimated. A pass rate estimation method to
A step of acquiring a measurement value of the current lot in one or more inspection items,
A step of calculating a deviation amount with respect to a standard value of the measurement value of the current lot for each inspection item;
The past measurement value of the past inspection item of the past lot that did not pass the inspection by the first inspection method, and the past measurement value that stored the re-inspection pass rate that the past lot passed by the second inspection method A deviation amount calculation means reads out the past measurement value from the storage means and calculates a deviation amount of the measurement value of the current lot with respect to a standard value, which is standardized by the past measurement value for each inspection item; ,
From the past deviation amount storage means storing the past deviation amount of the past measurement value, the reinspection pass rate estimation means reads the retest pass rate of the past lot in which the difference for each inspection item falls within a predetermined value. , Estimating the retest pass rate for this lot,
A pass rate estimation method having The inspection object is inspected for each lot by the first inspection method, and the possibility of passing if the lot that did not pass the inspection is inspected by the second inspection method different from the first inspection method is estimated. A measurement value acquisition step for acquiring a measurement value of the current lot in one or more inspection items;
A deviation amount calculating step for calculating a deviation amount with respect to a standard value of the measurement value of the current lot for each inspection item;
The past measurement value of the past inspection item of the past lot that did not pass the inspection by the first inspection method, and the past measurement value that stored the re-inspection pass rate that the past lot passed by the second inspection method A deviation amount calculation step of calculating the deviation amount of the measurement value of the current lot with respect to the standard value, which is standardized by the past measurement value for each inspection item, from the storage unit, and
From the past deviation amount storage means that stores the past deviation amount of the past measurement value, the retest pass rate of the past lot in which the difference for each inspection item falls within a predetermined value is read, and the retest pass of the current lot is passed. A retest pass rate estimation step to estimate the rate;
A program that executes

Images