JP2009217693A - Production line management device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To predict and determine normality/abnormality in a subsequent process based on a measurement value in a preceding process. <P>SOLUTION: Normality/abnormality determination of cranking torque in the subsequent process is quantitatively performed for predictive determination by means of a Mahalanobis-Taguchi system that is a statistic method using a measurement value of bore diameter in a honing process of the preceding process. A Mahalanobis distance calculation part 13 calculates a Mahalanobis distance from a measurement value of bore diameter in a case where the cranking torque becomes normal or abnormal, a determination distance setting part 14 determines a decision value for normality/abnormality, and a predictive determination part 15 compares the Mahalanobis distance of the measurement value of bore diameter of an evaluation workpiece with the decision value to determine the normality/abnormality. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a production line management apparatus that manages a production line having a plurality of processes.

  A product such as an automobile is manufactured from a large number of parts by a production line having a plurality of processes. For example, an engine, which is one of the main parts of an automobile, is manufactured through a plurality of processes in a production line. That is, the engine is manufactured through a number of processes such as a casting process of parts such as a cylinder block main body and a crankshaft, a processing process of these parts, and an assembly process of these parts. And in a plurality of processes among many processes, quality inspection is performed several times and a good engine is manufactured.

  Techniques that incorporate statistical methods have also been proposed for quality inspection of parts in production lines. For example, Patent Document 1 describes a technique for discriminating between a non-defective product and a defective product during manufacturing using the Mahalanobis distance obtained from statistical analysis. Patent Document 2 describes a technique for monitoring the tendency of a line state using time-series data of Mahalanobis distance.

JP 2006-293659 A JP 2004-165216 A

  As described above, in a production line including a large number of processes, quality inspection is performed in a plurality of processes among the numerous processes. However, if some formed product is formed in the downstream downstream process using a part that is determined to be non-defective (normal) in the upstream upstream process, the formed product is defective (abnormal) in the subsequent process inspection. ) May be determined.

  For example, in the engine production line, even when the bore diameter measurement value is within the standard in the cylinder block body machining process, an abnormality was confirmed by the cranking torque measurement in the assembly process of attaching the crankshaft to the cylinder block body. There is a case. Of course, if an abnormality is found in the formed product in the measurement of the assembly process, it goes without saying that measures such as removing the formed product from the production line are taken and thorough quality control is performed.

  However, if the occurrence of an abnormality in the post-process can be predicted from the measurement of the part in the pre-process, for example, the part can be removed from the production line at the stage of the pre-process, and the process and inspection in the post-process can be omitted. That is, in addition to thorough quality control, for example, the efficiency of the production line can be improved. Incidentally, the techniques described in Patent Documents 1 and 2 do not predict the occurrence of an abnormality in the subsequent process from the measurement of the parts in the previous process.

  The present invention has been made in such a situation, and an object of the present invention is to provide a technique for predicting and determining normal abnormality in a subsequent process from a measurement value in a previous process.

  In order to achieve the above object, a production line management apparatus according to a preferred aspect of the present invention is a production line management apparatus that manages a production line having a plurality of processes, and is a workpiece measurement value measured in a previous process. Measurement value acquisition means for acquiring a reference space setting means for setting a reference space reflecting the statistical properties of the measurement values of the non-defective workpiece based on the plurality of measurement values in the previous process for the plurality of non-defective workpieces; Based on the measurement value in the previous process for the workpiece to be evaluated obtained through the measurement value acquisition means and the reference space set in the reference space setting means, the statistical value of the measurement value and the reference space is statistically determined. A statistical distance calculating means for calculating a correct distance and the statistical distance for the measurement value of the evaluation target workpiece measured in the previous process, in the subsequent process. And having a prediction judgment means for judging to predict the normal or abnormal formation article formed by using the target work, the.

  In a desirable mode, the production line management device uses each of a plurality of measurement values in a previous process for a plurality of inspection workpieces, using a plurality of inspection workpieces used for a formed product for which normality / abnormality is determined in advance, and It further has a determination distance setting means for setting a determination distance that is a determination reference for normality / abnormality of the formed product from a statistical distance from the reference space, and the prediction determination means is calculated by the statistical distance calculation means. The statistical distance of the workpiece to be evaluated is compared with the determination distance set in the determination distance setting means, and the normality / abnormality of the formed product formed in the subsequent process is determined using the workpiece to be evaluated. It is characterized by that.

  In a desirable mode, the production line management device removes the evaluation target work from the production line when the formed product formed in the subsequent process using the evaluation target work is predicted to be abnormal. It is characterized in that it is not allowed to enter.

  In a desirable aspect, the production line is a production line for a vehicle engine, and in a machining process corresponding to the preceding process, a measurement value of a cylinder body corresponding to the workpiece is measured, and a set corresponding to the subsequent process is set. In the attaching step, a cylinder portion corresponding to the formed product is formed using a cylinder portion main body corresponding to the workpiece to be evaluated, and the prediction determination means is a cylinder corresponding to the workpiece to be evaluated measured in the machining step. Based on the statistical distance about the measured value of a part main part, the abnormalities of a cylinder part formed using the cylinder part main part in an assembling process are predicted and judged.

  In a desirable mode, the prediction determination means is formed by attaching a crank to the cylinder part body in the assembly process based on a statistical distance about a measured value of the bore diameter of the cylinder part body measured in the machining process. The normality / abnormality of the cranking torque of the cylinder portion to be determined is predicted and determined.

  In a preferred aspect, the reference space setting means sets the reference space based on a Mahalanobis-Taguchi system, and the statistical distance calculation means calculates a Mahalanobis distance as the statistical distance.

  According to the present invention, it is possible to predict and determine normality / abnormality in the subsequent process from the measurement value in the previous process. Further, in a preferred aspect of the present invention, when it is predicted that the formed product formed in the subsequent process using the evaluation target workpiece is abnormal, the evaluation target workpiece is removed from the production line and input to the subsequent process. By not doing so, for example, it becomes possible to omit useless processing and inspection in the subsequent process.

  FIG. 1 is a functional block diagram for explaining a preferred embodiment of a production line management apparatus according to the present invention. FIG. 1 shows a production line management apparatus 10 and a production line 100 managed thereby.

  The production line 100 includes a plurality of processes, and produces a finished product through a plurality of processes from a plurality of parts (workpieces). In the present embodiment, the production line 100 manufactures a vehicle engine. That is, parts such as a cylinder block main body and a crankshaft are cast in the casting line 110, the parts are processed in the processing line 120, and the parts are assembled to each other in the assembly line 130, and are completed through the test bench 140. An engine (or a partially completed product in the engine) is manufactured.

  FIG. 2 is a block diagram for explaining steps related to the manufacture of the cylinder block in the engine production line 100 shown in FIG. When the cylinder block main body cast in the casting process is put into the cylinder block main body processing step, a honing process 122 is performed after the cylinder block main body is bored, and a bore is formed in the cylinder block main body. . In the present embodiment, the bore diameter is measured in the honing process 122 (the bore diameter will be described in detail later using FIG. 5). Then, after the appearance inspection, the cylinder block body is put into the cylinder block sub-assembly process.

  In the cylinder block sub-assembly process, an engine No. that uniquely identifies the engine product with respect to the cylinder block main body. The crank (crankshaft) conveyed from the crankshaft machining step is attached to the cylinder block body, and the crank cap is tightened. Next, the piston sub-assembly conveyed from the piston sub-assembly step is attached to the cylinder block body, and the connecting rod bolts are tightened. In this way, a crank, a piston sub-assembly and the like are assembled to the cylinder block body to form a cylinder block, and cranking torque measurement 132 of the cylinder block is performed.

  In this embodiment, as will be described in detail later, normality / abnormality in cranking torque measurement 132 is predicted from the measured bore diameter in honing process 122. Therefore, in this embodiment, the cranking torque measurement 132 can be omitted. Of course, in addition to the normal / abnormal prediction of the cranking torque measurement 132, the cranking torque measurement 132 may be actually performed for confirmation.

  Returning to FIG. 1, each of the casting line 110, the processing line 120, the assembly line 130, and the test bench 140 is provided with one or a plurality of controllers 50, and each controller 50 obtains information regarding the corresponding process. To do. The acquired information includes a measured value of the bore diameter in the honing process 122 (FIG. 2) in the processing line 120, and the like. When the cranking torque measurement 132 (FIG. 2) is performed in the assembly line 130, the measured value of the cranking torque is also acquired.

  Information obtained from each process in the casting line 110 via the controller 50 is collected by the server 20. Information obtained from each process in the processing line 120 via the controller 50 is collected by the server 30, and information obtained from each process in the assembly line 130 and the test bench 140 via the controller 50 is collected from the server 50. 40 collected. The production line management device 10 collects information collected by the servers 20, 30, and 40, and the entire production line 100 is managed by the production line management device 10.

  The production line management apparatus 10 includes a data storage unit 16 that is a storage device such as a hard disk, and stores information collected from the servers 20, 30, and 40 in the data storage unit 16. When the information is stored in the data storage unit 16, the production line management device 10 forms the following quality history data, for example.

  FIG. 3 is a diagram for explaining an example of the quality history data managed by the production line management apparatus 10. The production line management device 10 has an engine No. for identifying each engine. The cylinder block ID and the crankshaft ID used for manufacturing each engine are stored in association with each other. Further, for each engine, a measured value such as a measured value of the bore diameter measured in the honing process and a measured value of the cranking torque measurement is also stored. When the cranking torque measurement is omitted, the measurement result of the cranking torque measurement in the quality history data shown in FIG. 3 is also omitted.

  Returning to FIG. 1, the production line management apparatus 10 includes a measurement value acquisition unit 11, a reference space setting unit 12, a Mahalanobis distance calculation unit 13, a determination distance setting unit 14, and a prediction determination unit 15 in addition to the data storage unit 16. ing. The production line management device 10 is realized by a computer, for example. In that case, a program for causing the computer to function as the measurement value acquisition unit 11, the reference space setting unit 12, the Mahalanobis distance calculation unit 13, the determination distance setting unit 14, and the prediction determination unit 15 is formed and stored. By causing a computer to read a storage medium such as an optical disk, the production line management apparatus 10 is realized in cooperation with the program and hardware such as a CPU in the computer.

  In the present embodiment, the production line management device 10 quantitatively determines normal / abnormality in cranking torque measurement based on the Mahalanobis-Taguchi system, which is one of statistical analysis methods, from a bore diameter measurement value in honing. Predicted and determined. Therefore, the normal / abnormal prediction will be described in detail below. In the following description, the reference numerals in FIG. 1 are used for the portions (configurations) shown in FIG.

  The main part contributing to the generation of cranking torque is a bore part in the crankshaft, piston or cylinder block in the engine. In the present embodiment, attention is paid to the fact that the measured value of the cranking torque is influenced by the measured value of the bore diameter.

  FIG. 4 is a view for explaining the bore 63 formed in the cylinder block main body 61, and shows a cross section of the cylinder block main body 61 during processing. The example shown in FIG. 4 is a cylinder block body 61 of a V-type 6-cylinder engine, and a total of six bores 63 are formed in each of the right (R) and left (L) rows. In FIG. 4, identification numbers (1) to (6) are assigned to the six bores 63. The bore 63 is processed by the honing grindstone shaft 65 of the honing machine in the honing process. When the bore 63 is formed, the bore diameter is measured in the honing process.

  FIG. 5 is a view for explaining the bore diameter measured in the honing process, and shows one of the six bores 63. Each bore 63 is formed in a substantially cylindrical shape as shown in FIG. For each bore 63, along the vertical direction thereof, that is, along the depth direction of the bore 63, the bore 63 in each of the three directions of the upper, middle, and lower portions is orthogonal to each other in the X-axis direction and the Y-axis direction. The diameter of is measured.

  In FIG. 5, the bore diameter in the X-axis direction at the top is ixu, and the bore diameter in the Y-axis direction at the top is iyu. In addition, the bore diameter in the X-axis direction in the middle part is ixc, the bore diameter in the Y-axis direction in the middle part is yyc, the bore diameter in the X-axis direction in the lower part is ixd, and the bore diameter in the Y-axis direction in the lower part is iyd. In this manner, a total of six bore diameters are measured for each bore 63. i is bore no. (I = 1 to 6).

  FIG. 6 is a diagram illustrating an example of bore diameter measurement result data. For example, the measured value of the bore diameter obtained from the processing line 120 is acquired by the measured value acquisition unit 11 of the production line management apparatus 10 via the controller 50 or the server 30 and the bore diameter for each engine (each cylinder block). These measurement values are collected and measurement result data as shown in FIG. 6 is stored in the data storage unit 16 of the production line management apparatus 10.

  In FIG. 6, for each block ID (cylinder block ID) that identifies each cylinder block, the measurement date and time when the bore diameter is measured and the measured value of the bore diameter related to the cylinder block are associated with each other. For each cylinder block, the bore diameter measurement values at six locations (see FIG. 5) are stored for each of the six bores. That is, a total of 36 measured bore diameters from 1xu to 6yd are stored for each cylinder block.

  Next, the relationship between the cranking torque and the mechanical friction loss of the bore portion in the crankshaft, piston, or cylinder block will be described. FIG. 7 is a view for explaining the mechanical friction loss in the bore 63. The piston rings 71, 73, and 75 of the piston 67 have a great influence on the cranking torque. The piston ring 71 is a top, the piston ring 73 is a second, and the piston ring 75 is an oil ring.

  When the piston 67 moves in a bore 63 provided in the cylinder block body 61, a mechanical friction loss L is generated between the cylinder block body 61 and the piston rings 71, 73, 75 due to the sliding surface pressure. . The mechanical friction loss L at this time greatly affects the cranking torque. This relationship can be expressed as, for example, the following equation.

  In the present embodiment, the cranking torque measurement in the assembly process is quantitatively predicted using the bore diameter measured in the machining process. Piston ring tension is not measured because it is difficult to measure. The value of the piston ring tension is unknown, but if the piston ring tension and the bore diameter are within the standard, it is assumed that the measured value of the cranking torque is within the standard. That is, the cranking torque value is proportional to the piston ring tension and inversely proportional to the bore diameter according to Equation (1) (see FIG. 8).

  FIG. 8 is a diagram showing the relationship among piston ring tension, bore diameter, and cranking torque. In this embodiment, the piston ring tension and the bore diameter are considered to be within the standard, and attention is paid to the correlation between the cranking torque value and the bore diameter.

  In the present embodiment, as described with reference to FIGS. 5 and 6, a total of 36 bore diameters are measured for each engine (each cylinder block) at each of the six cylinders. However, considering the relationship between the measured values of the 36 bore diameters and the cranking torque, the arrangement angle of the bores in the cylinder block of the V-type 6-cylinder engine and the phase (positional relationship) in the bores of the crankshaft and piston are taken into consideration. Even if all 36 bore diameter data are within the standard, the cranking torque may not be within the standard.

  As for the bore diameter of each bore, it has been revealed that there are multiple collinearity at the upper, middle, and lower portions in the same axial direction (X-axis direction or Y-axis direction). Therefore, for each bore, the average value of the data in the upper, middle and lower portions in the X-axis direction is taken, the average value is taken as the bore diameter in the X-axis direction, and the upper, middle and lower portions in the Y-axis direction are taken into account. The average value of the data at the locations is taken, and the average value is taken as the bore diameter in the Y-axis direction. As a result, two values of the bore diameter in the X-axis direction and the bore diameter in the Y-axis direction are obtained for each bore, and a total of 12 measured bore diameter values are obtained from the bores of 6 cylinders for each engine. .

  Then, when only a plurality of bore diameter data within the standard are selected and their average and variance are compared, the distribution shown in FIG. 9 is obtained. From FIG. 9, it was found that the distribution is larger when the cranking torque value is abnormal than when the cranking torque value is normal, both in terms of average and variance. This indicates that it is possible to determine whether the cranking torque value is normal or abnormal based on the measured bore diameter value within the standard.

  As described with reference to FIGS. 1 and 2, in the manufacturing process sequence of the production line 100, the bore diameter is measured during the honing process in the cylinder block main body process, which is the previous process, and cranking torque is measured. Is measured in a cylinder block sub-assembly process, which is a subsequent process.

  In consideration of the analysis content of predicting and determining the cranking torque value in the subsequent process from the measured bore diameter value measured in the previous process of the production line 100, in this embodiment, Mahalanobis, which is a statistical technique, is used. -Utilize Taguchi (MT) system. Therefore, the Mahalanobis-Taguchi (MT) system will be described next.

  FIG. 10 is a diagram for explaining the outline of the Mahalanobis-Taguchi (MT) system. The Mahalanobis-Taguchi (MT) system replaces the difference between the unit space (reference space) 81 and the inspection data 87, which is a measured reference data group related to non-defective products, with a Mahalanobis distance 89 which is a one-dimensional statistical distance, This is a determination method for determining pass / fail of the inspection data 87. FIG. 10 also shows an average value 85 and a variance value 83 of the reference data group.

  Unlike discriminant analysis, which is a multivariate analysis technique that determines which of a plurality of populations belong to, the Mahalanobis-Taguchi (MT) system is a method to check how close to one normal population, One feature is that analysis is possible even with a small amount of defective data.

  FIG. 11 is a flowchart showing a procedure of the Mahalanobis-Taguchi (MT) system in the present embodiment. First, reference data, which is measured data for a non-defective product (good product work), is collected (S101), and a unit space is defined by the reference space setting unit 12 of the production line management apparatus 10 based on the collected data. (S103).

  The average value (formula 2), variance (formula 3) and standard deviation (formula 4) for the measurement items (x1, x2,..., Xi,..., Xn) are generally as follows: become that way.

  The measurement item in this embodiment is an average value of bore diameters for the upper part, the middle part, and the lower part. As shown in FIG. 12, a total of 12 from 1x to 6y is obtained for each block ID, that is, for each cylinder block. It becomes the measured value. That is, the bore No. in FIG. The upper, middle, and lower X-axis bore diameter measurements of i (i = 1 to 6) are ixu, ixc, and ixd, respectively, and the upper, middle, and lower Y-axis bore diameter measurements are iyu, iyc, Assuming iyd, ix and iy (i = 1 to 6) in FIG. 12 are calculated as follows.

  Next, normalization of each measured value in FIG. 12 is performed. Here, first, T1 = 1x, T2 = 1y, T3 = 2x, T4 = 2y,..., T11 = 6x, T12 = 6y, and conversion is performed as shown in FIGS. In FIG. 13, for the 12 measurement items normalized T1, T2,..., T12, Equation 7 is used for X-axis data, and Equation 8 is used for Y-axis data.

  From Equation 7 and Equation 8, as shown in FIG. 13, the average value for each measurement item is 0 and the standard deviation is 1 for all the measurement items T1 to T12.

  Next, a correlation matrix of normalized data is obtained. If the correlation coefficient is generally expressed by equation (9), the correlation matrix is generally expressed by equation (10).

  Further, an inverse matrix of the correlation matrix is obtained. Generally, Equation 11 is obtained. In this embodiment, n = 12.

  This completes the definition of the unit space (reference space), and the Mahalanobis distance is calculated using the inspection data. For example, if the inspection data is a measured value of the bore diameter shown in FIG. The bore diameter measurement values of the upper, middle, and lower X axes of i (i = 1 to 6) are ix'u, ix'c, and ix'd, respectively, and the bore diameter measurement values of the upper, middle, and lower Y axes. Are iy'u, iy'c, and iy'd, the average values ix 'and iy' (i = 1 to 6) of the bore diameters for the upper, middle, and lower parts are calculated as follows.

The results obtained by averaging the measured values of the bore diameters shown in FIG. 14 using the formulas 12 and 13 are ix ′ and iy ′ (i = 1 to 6) shown in FIG. Further, when ix ′ and iy ′ shown in FIG. 15 are normalized by Expression 7 and Expression 8, T′1, T′2,..., T′12 normalized as shown in FIG. can get. Then, the Mahalanobis distance D ² is calculated from the normalized T ′ 1 to T ′ 12 as shown in Equation 14.

The number 14 formula is a general equation for calculating the Mahalanobis distance D ^2, the n = 12 in this embodiment.

  Further, it is verified whether the result is valid and actually usable in the production line 100. In other words, it is necessary to verify whether the cranking torque is normal or abnormal based on the measured bore diameter using the defined unit space.

FIG. 17 is a diagram illustrating a verification result regarding the normality / abnormality determination of the cranking torque. If the sought unit space, for the measurement of the bore diameter for the (reference data) is calculated Mahalanobis distance D ² (FIG. 11, S105), the measurement values of all the bore diameter (reference data), D ² < P ′ (FIG. 17, “unit space”).

On the other hand, when verified using separately collected inspection data, the Mahalanobis distance D ² (= P ₁ ) obtained from the measured bore diameter when cranking torque is normal is all P ₁ <P (FIG. 17, “At verification: normal torque”), and Mahalanobis distance D ² (= P ₂ ) obtained from the measured bore diameter when cranking torque is abnormal is P ₂ > P (FIG. 17, “Verification: Torque Abnormality”). In other words, in the example shown in FIG. 17, by setting the determination value S = P (P ₁ <P, P ₂ > P) (FIG. 11, S107), normality of cranking torque can be calculated from Mahalanobis distance D ^2. Abnormality can be determined.

  By the way, in the example of FIG. 17, the number of bore diameter measurement values when the cranking torque is abnormal in the inspection data is less frequently generated and the number of bore diameter measurement values when normal is obtained. About 1/10.

When the determination value for normality of cranking torque is determined by the determination distance setting unit 14 of the production line management device 10, a determination is made on a part (evaluation work) that is actually used for production in the production line 100. . That is, the bore diameter is measured in the honing process for the cylinder block main body, which is an evaluation work, and the measured value is collected in the measured value acquisition unit 11 (FIG. 11, S109). By Mahalanobis distance calculating unit 13, the measurement of the bore diameter of the collected evaluation work, Mahalanobis distance D ² is calculated based on the number 14 expression. The Mahalanobis distance calculated here is DX (FIG. 11, S111).

  In addition, the processing of the measurement value of the bore diameter of the evaluation workpiece is performed using Equation 12 and Equation 13 when the measurement value of the bore diameter of the evaluation workpiece is FIG. The data shown in FIG. 15 is converted, and further normalized by the equations (7) and (8) to be converted into data in the format shown in FIG.

  When the Mahalanobis distance DX for the workpiece for evaluation is calculated, the prediction determination unit 15 of the production line management apparatus 10 compares the Mahalanobis distance DX with the determination value S (FIG. 11, S113). When DX> S, it is determined as “abnormal”, and the workpiece for evaluation is removed from the line in the honing process (FIG. 11, S115). If necessary, it is desirable to inspect an evaluation workpiece determined to be abnormal.

  In the case of DX ≦ S, it is determined to be “normal”, and the evaluation work is passed from the honing process to the subsequent process (FIG. 11, S117). Then, the processing from S109 to S117 in FIG. 11 is repeatedly executed until the work is completed. In this way, normality / abnormality is determined in honing for a plurality of evaluation workpieces.

  As mentioned above, although preferred embodiment of this invention was described, according to embodiment mentioned above, since the normal abnormality of a post process can be estimated using the measurement data of a pre process, an abnormal product is made into a post process. Can be suppressed. Further, since the measurement data of the previous process is used to perform prediction determination of the subsequent process and abnormal items are removed, it is possible to suppress performing unnecessary processing and assembly in the subsequent process. In addition, the quality information collection of the quality history information system can also omit the collection of information related to the cranking torque in the subsequent process.

  As mentioned above, although preferred embodiment of this invention was described, embodiment mentioned above and its effect are only a mere illustration in all points, and do not limit the scope of the present invention. The present invention includes various modifications without departing from the essence thereof.

It is a functional block diagram for demonstrating the production line management apparatus which concerns on this invention. It is a block diagram for demonstrating the process regarding manufacture of a cylinder block. It is a figure for demonstrating an example of quality history data. It is a figure for demonstrating the bore formed in a cylinder block main body. It is a figure for demonstrating the bore diameter measured in a honing process. It is a figure which shows an example of the measurement result data of a bore diameter. It is a figure for demonstrating the mechanical friction loss in a bore | bore. It is a figure which shows the relationship between piston ring tension, a bore diameter, and cranking torque. It is a figure which shows the average and dispersion | distribution of bore diameter data. It is a figure for demonstrating the outline | summary of a Mahalanobis-Taguchi system. It is a flowchart which shows the procedure of a Mahalanobis-Taguchi system. It is a figure which shows a total of 12 measured values for each cylinder block. It is a figure which shows 12 measurement items normalized. It is a figure which shows an example of the measured value of a bore diameter. It is a figure which shows the measured value of the bore diameter by which the average process was carried out. It is a figure which shows 12 measurement items normalized. It is a figure which shows the verification result about the normality abnormality determination of cranking torque.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 Production line management apparatus, 11 Measurement value acquisition part, 12 Reference space setting part, 13 Mahalanobis distance calculation part, 14 Judgment distance setting part, 15 Prediction judgment part, 16 Data storage part, 100 Production line.

Claims (6)

A production line management device for managing a production line having a plurality of processes,
A measurement value acquisition means for acquiring a measurement value of the workpiece measured in the previous process;
A reference space setting means for setting a reference space that reflects the statistical properties of the measurement values of the non-defective workpiece, based on a plurality of measurement values in the previous process for a plurality of non-defective workpieces;
Based on the measurement value in the previous process for the workpiece to be evaluated obtained through the measurement value acquisition means and the reference space set in the reference space setting means, the statistical value of the measurement value and the reference space is statistically determined. Statistical distance calculating means for calculating a correct distance;
Prediction based on the statistical distance for the measurement value of the evaluation target workpiece measured in the previous step, and predicting and judging normality of a formed product formed using the evaluation target workpiece in the subsequent step A determination means;
Having
Production line management device characterized by that. In the production line management device according to claim 1,
Using a plurality of inspection workpieces used for a molded product for which normality / abnormality is determined in advance, formation from a statistical distance between each of a plurality of measured values and a reference space in a previous process for a plurality of inspection workpieces It further has a determination distance setting means for setting a determination distance that is a determination criterion for normality or abnormality of the product
The prediction determination unit compares a statistical distance for the evaluation target workpiece calculated by the statistical distance calculation unit with a determination distance set by the determination distance setting unit, and uses the evaluation target workpiece. Determine the normality of the formed product formed in the subsequent process
Production line management device characterized by that. In the production line management device according to claim 2,
When it is predicted that the formed product formed in the subsequent process using the evaluation target work is abnormal, the evaluation target work is not removed from the production line and is not input to the subsequent process.
Production line management device characterized by that. In the production line management device according to any one of claims 1 to 3,
The production line is a vehicle engine production line,
In the machining step corresponding to the previous step, the measured value of the cylinder body corresponding to the workpiece is measured,
In the assembly step corresponding to the post-process, a cylinder portion corresponding to the formed product is formed using a cylinder portion main body corresponding to the workpiece to be evaluated,
The prediction determination unit is formed by using the cylinder body in the assembly process based on a statistical distance regarding the measured value of the cylinder body corresponding to the evaluation object workpiece measured in the machining process. Predict and judge normality of cylinder part
Production line management device characterized by that. In the production line management device according to claim 4,
The prediction determination means is a cylinder part formed by attaching a crank to the cylinder part body in the assembly process based on a statistical distance about a measured value of the bore diameter of the cylinder part body measured in the machining process. Predicting and determining normality and abnormality of cranking torque of
Production line management device characterized by that. In the production line management device according to any one of claims 1 to 5,
The reference space setting means sets the reference space based on the Mahalanobis-Taguchi system,
The statistical distance calculation means calculates a Mahalanobis distance as the statistical distance.
Production line management device characterized by that.

Images