WO2021106288A1 - 品質計測方法及び品質計測装置 - Google Patents

品質計測方法及び品質計測装置 Download PDF

Info

Publication number
WO2021106288A1
WO2021106288A1 PCT/JP2020/031469 JP2020031469W WO2021106288A1 WO 2021106288 A1 WO2021106288 A1 WO 2021106288A1 JP 2020031469 W JP2020031469 W JP 2020031469W WO 2021106288 A1 WO2021106288 A1 WO 2021106288A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
quality measurement
quality
measurement method
prediction model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/JP2020/031469
Other languages
English (en)
French (fr)
Japanese (ja)
Inventor
尚大 平
坂本　雅基
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Kao Corp
Original Assignee
Kao Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Kao Corp filed Critical Kao Corp
Priority to CN202080082342.6A priority Critical patent/CN114760969B/zh
Publication of WO2021106288A1 publication Critical patent/WO2021106288A1/ja
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61F—FILTERS IMPLANTABLE INTO BLOOD VESSELS; PROSTHESES; DEVICES PROVIDING PATENCY TO, OR PREVENTING COLLAPSING OF, TUBULAR STRUCTURES OF THE BODY, e.g. STENTS; ORTHOPAEDIC, NURSING OR CONTRACEPTIVE DEVICES; FOMENTATION; TREATMENT OR PROTECTION OF EYES OR EARS; BANDAGES, DRESSINGS OR ABSORBENT PADS; FIRST-AID KITS
    • A61F13/00—Bandages or dressings; Absorbent pads
    • A61F13/15—Absorbent pads, e.g. sanitary towels, swabs or tampons for external or internal application to the body; Supporting or fastening means therefor; Tampon applicators
    • A61F13/84—Accessories, not otherwise provided for, for absorbent pads
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00—Program-control systems
    • G05B19/02—Program-control systems electric
    • G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]

Definitions

  • the present invention relates to a quality measuring method and a quality measuring device.
  • Patent Document 1 describes a method for calculating the amount of change due to a change in the manufacturing process. In this method, for the purpose of improving the productivity of the manufacturing plant, the amount of change due to the change in the manufacturing process such as the control function and the operation method is accurately calculated.
  • Patent Document 2 discloses a system and a process that can be configured to correlate manufacturing parameters and performance feedback parameters with individual absorbent articles manufactured by a processing apparatus.
  • Product performance data obtained from product testing can be used as a tool for manufacturers to manufacture and / or process and adjust future processing equipment.
  • Patent Document 3 describes a manufacturing method for manufacturing a product through a plurality of manufacturing processes. This method is also used in the production of absorbent articles, and can improve the machine operating rate and the non-defective rate of products.
  • the present invention is a quality measurement method in a process of manufacturing a product through a plurality of manufacturing processes, and is acquired by in-line measurement based on a step of constructing a prediction model between quality data and process data and the prediction model.
  • a quality measurement method including a step of calculating a quality measurement value of a manufactured product from the process data.
  • the present invention is a quality measuring device used for a device that manufactures a product through a plurality of manufacturing processes, the quality measuring device having an analysis server, and the analysis server for quality data and process data.
  • a quality measurement device in which a prediction model is constructed between the two, and the analysis server calculates a quality measurement value from the process data by in-line measurement based on the prediction model.
  • FIG. 5 is a schematic configuration diagram showing a preferable example of a manufacturing apparatus using the quality measurement method according to the present invention together with various sensors used in the quality measurement method. It is a drawing which showed a concrete example of the process data which is the object of time-series processing.
  • A shows an example of a method of associating process data with each other in a time-series processing process
  • (B) shows an example of a storage state of process data after time-series processing.
  • FIG. 1 It is a block diagram which showed the preferable embodiment of the quality measuring apparatus of this invention.
  • (A) shows a form in which the data collection CPU is separated from the analysis server, and
  • (B) shows a form in which the data collection CPU and the analysis server are integrated. It is a graph which compared the quality measurement value of the inlet side seal strength calculated in Examples 1 to 3 with the actual measurement value.
  • the present invention relates to a quality measurement method and a quality measurement device capable of measuring quality that is difficult to measure by in-line measurement without performing sampling inspection.
  • quality measurement method and quality measurement device of the present invention quality that is difficult to measure by in-line measurement can be measured without performing sampling inspection.
  • a preferred embodiment of the quality measurement method of the present invention will be described below with reference to FIGS. 1 to 4 by taking the quality measurement method of the absorbent article as an example. First, the types of data used in the quality measurement method of the present invention will be described.
  • Process data is data acquired by in-line measurement in a plurality of processes for manufacturing a product. Process data is roughly classified into material data acquired from the material of the product and equipment data acquired from the equipment. Material data and equipment data are data that are acquired in time series by in-line measurement in the manufacturing process and can change with time. Specific examples of the material data include sheet raw fabric diameter, sheet raw fabric storage humidity, sheet temperature, sheet tension, sheet thickness and the like. Specific examples of the device data include oil pressure, pattern roll temperature, anvil roll temperature, roll rotation speed, motor shaft rotation speed, motor shaft load, dancer pressure, conveyor transfer speed, and the like. Further, the process data may include data that can be appropriately operated as a set value by the operator of the production line (hereinafter, referred to as “operation data”).
  • operation data data that can be appropriately operated as a set value by the operator of the production line
  • the operation data include oil pressure, pattern roll temperature, anvil roll temperature, dancer pressure, conveyor transfer speed, etc. in the case of device data, and sheet raw material storage humidity, etc. in the case of material data.
  • the process data preferably includes sheet tension, sheet thickness, or both as material data.
  • Quality data is data that is difficult to measure by in-line measurement. For example, it is measured and acquired by sampling inspection such as destructive inspection (offline measurement).
  • quality data sanitary evaluated by seal strength, liquid remaining amount, wetback amount, absorption rate, basis weight distribution, etc. of diaper side seal, napkin embossed seal, eye mask ear hook seal, individual packaging outer circumference seal, etc.
  • Sensory evaluation such as paint quality, touch, texture, appearance, etc. evaluated by the thermal performance, viscosity, solid content, etc. of the heating element evaluated by the absorption performance, duration, maximum temperature, temperature rise rate, etc. of the product, and moisture permeability
  • the moisture permeability of the film can be mentioned.
  • the seal strength includes the seal strength of the portion where two or more sheets are joined.
  • the seal strength is generally measured by a destructive inspection performed while peeling off the bonded seal using a tensile tester. Therefore, the seal strength corresponds to the quality data.
  • Various methods can be used to bond the sheets. Specific examples thereof include heat sealing, ultrasonic sealing, laser welding, and hot melt bonding.
  • the quality measurement method of the present invention is used in a method of manufacturing a product through a plurality of manufacturing steps.
  • the products in the present specification mean not only finished products (final products) but also intermediate products and semi-finished products in the middle of the manufacturing process.
  • the product to be produced is not particularly limited as long as it exhibits the effects of the present invention.
  • a product manufactured using a sheet can be mentioned, and a composite sheet used for a diaper or the like is typically mentioned.
  • the sheet include those made of various materials, and include, for example, a fiber sheet such as a non-woven fabric.
  • the manufactured product is not limited to a single-wafered product in which individual products are separated, and may be a continuous product such as a roll-up paper. If the product is continuous, the process data and quality data may be a moving average of the individual data.
  • the moving average method is not particularly limited as long as the production method of the present invention is effective. For example, when performing a moving average of ten values consecutive in FIG. 3 (A), the moving average value corresponding to the data X 10 is 10 pieces of data immediately before the successive including data X 10 (i.e., the data It is the value obtained by dividing the sum of X 1 to X 10) by 10.
  • the moving average value is the value obtained by summing all the continuous data and dividing by 10.
  • the moving average value corresponding to the data X 3 is a value obtained by dividing the sum of the four values of the data X 0 to X 3 by 10.
  • the fiber sheet described above is easily deformed by an external force, and its shape is difficult to stabilize. Therefore, the fiber sheet can be deformed in the manufacturing process depending on the conditions. Therefore, in the quality measurement method of the present invention, it is preferable to acquire process data related to the fiber sheet in chronological order. At that time, it is preferable to acquire process data before and after the processing of the fiber sheet.
  • the process data of the sheet tension and the sheet thickness are preferable as a material for better estimating the quality data (for example, the quality of the seal between the fiber sheets) regarding the fiber sheet when manufacturing the composite sheet.
  • the quality measurement method of the present invention is based on a process of constructing a prediction model between quality data and process data (hereinafter referred to as "prediction model construction process") and process data acquired by in-line measurement based on the prediction model. It has a process of calculating the quality measurement value of the manufactured product (hereinafter, referred to as "in-line quality measurement process").
  • the quality measurement value is data of the same type as the above-mentioned quality data calculated by in-line measurement. Therefore, in the in-line quality measurement process, the product quality measurement value is estimated by the in-line measurement, instead of measuring the product quality data by the offline measurement.
  • the prediction model construction process will be described in order.
  • the prediction model construction process typically includes a process data acquisition process 61, a time series processing process 62, a data combination process 63, an end determination process 64, a data analysis process 71 to 73, and generalization performance confirmation. It consists of steps 81 to 83 and the like.
  • the prediction model construction process is not limited to the form shown in FIG.
  • a matrix test, a factor fixation test, an experimental design method, or the like can be used. It is preferable to use design of experiments from the viewpoint of reducing the number of combinations of values of process data to be acquired.
  • the orthogonal array used in the design of experiments is not particularly limited. For example, an L 12 orthogonal array, an L 18 orthogonal array, an L 36 orthogonal array, or the like can be used.
  • the rate of fluctuation of the predicted quality measurement value can be changed. For example, if the level assigned to the orthogonal array is increased, the rate of fluctuation of the predicted quality measurement value becomes smaller. On the contrary, if the level assigned to the orthogonal array is reduced, the rate of fluctuation of the predicted quality measurement value becomes large.
  • Process data acquisition process 61 process data for constructing a prediction model is acquired from the production line by in-line measurement.
  • material data and device data are acquired together as process data. From the viewpoint of acquiring process data directly related to the manufactured product, it is preferable to acquire the process data during the operation of the production line.
  • a radiation thermometer can be used to acquire temperature process data.
  • a contact type displacement meter or a non-contact type displacement meter using a laser, ultrasonic waves, capacitance or the like can be used.
  • the process data can be acquired before and after machining by providing various sensors before and after the machining part of the manufacturing apparatus.
  • the manufactured product is unstable such as a fiber sheet whose thickness etc. can change in the manufacturing process
  • acquiring process data before and after processing is particularly important from the viewpoint of accurately estimating quality data. It is effective.
  • the time-series processing step 62 is a step of correcting the time delay between the process data and storing it in the analysis server or the like.
  • a large number of sensors for acquiring process data may be provided during the process of manufacturing a product. Then, the process data acquired at the same time by the sensor provided upstream in the machine flow direction and the sensor provided downstream becomes the data observed for different products. Therefore, in the present invention, it is preferable to perform time-series processing on the process data in order to make each process data correspond to the same product.
  • the time delay time lag
  • the data is stored by correcting the time delay between the sensors that acquire the process data according to the speed of the machine flow.
  • individual products and process data can be associated and organized, and data analysis becomes possible.
  • the process data and the quality data can be associated with each other.
  • FIG. 2 A specific method of time-series processing will be described with reference to the manufacturing apparatus 11 shown in FIG. Various sensors used in the quality measurement method of the present invention are arranged in the manufacturing apparatus 11.
  • a temperature sensor 33B for acquiring the process data of the seal unit temperature a sheet raw fabric sensor 34A for acquiring the process data of the upper shaft sheet raw fabric diameter, and the like can be mentioned.
  • each distance is made to correspond to the number of sheets cut (the number of products).
  • the temperature sensor 33B is the 34th from the motor 24E.
  • the sheet raw material sensor 34A is the 40th sheet from the motor 24E.
  • the number of products from the motor 24E (see FIG. 2) to the temperature sensor 33B (see FIG. 2) is 34.
  • the data Y 6 acquired by the temperature sensor 33B at the timing retroactively for 34 sheets according to the speed of the machine flow is the process data of the seal unit temperature corresponding to the product in which the data Z 40 is observed. It becomes.
  • the data X 0 that goes back in time for 40 sheets is the process data of the upper shaft sheet original diameter corresponding to the product in which the data Z 40 was observed. Become. In this way, the process data can be made to correspond to each product and the time delay can be corrected.
  • the data combining step 63 is a step of associating and combining the process data of each product with the quality data of each product.
  • the method of joining is not particularly limited, but quality data may be stored and organized in the same row or column of spreadsheet software for each product in the same manner as process data. Since the prediction model construction process has the data combination process 63, the quality data can be associated with various process data. For example, material data, equipment data, and operation data can be associated with and combined with quality data. From the viewpoint of speeding up data analysis, it is preferable to use the process data in which the time delay is corrected in the time series processing step 62 in the data combining step 63. For one product, the corresponding process data and quality data are stored in an analysis server or the like as a group of data sets. That is, as many data sets as the number of products can be acquired.
  • the end determination step 64 is a step of determining whether or not the number of data sets acquired through the process data acquisition step 61 to the data combination step 63 is sufficient for use in data analysis. When the number of data sets is sufficient, the process proceeds to the data analysis step 71 described later. On the other hand, when the number of data sets is not sufficient, the process data acquisition step 61 to the data combination step 63 are repeated while appropriately changing the combination of the value of the process data until a sufficient number of data sets can be acquired. At this time, it is preferable to efficiently acquire the data set by the above-mentioned design of experiments.
  • the data analysis step 71 is a step of performing the first data analysis after acquiring the data set.
  • the acquired data set is divided into two.
  • One of the two divisions is used as training data, and the other is used as test data.
  • the training data is mainly data for constructing a model that can be a candidate for a prediction model (hereinafter, referred to as "prediction model candidate").
  • the test data is data for verifying the prediction model candidates constructed mainly from the training data and determining the prediction model.
  • various analyzes are performed on the training data and the test data, and a plurality of prediction model candidates are constructed. For the constructed multiple prediction model candidates, the scores described later are calculated.
  • the division ratio of the training data and the test data may be arbitrarily determined, but it is generally divided at a ratio of 8: 2 or 5: 5.
  • cross-validation may be used to divide the training data and the test data.
  • the method of determining and selecting the prediction model candidate is not particularly limited. For example, there is a method of adopting a prediction model whose analysis result using training data and test data is above a certain level as a prediction model candidate.
  • the analysis method performed on the training data and the test data is not particularly limited as long as it exhibits the effect of the present invention.
  • various linear regressions and various non-linear regressions can be mentioned.
  • Specific examples of linear regression include multiple regression analysis, PLS regression analysis, lasso regression analysis, ridge regression analysis, elastic net regression analysis, sgd regression analysis, support vector regression analysis, and the like.
  • Specific examples of non-linear regression include non-linear support vector regression analysis, kernel ridge regression analysis, random forest regression analysis, MLP regression analysis, and the like.
  • hyperparameters By adjusting the hyperparameters, the accuracy of the prediction model is improved.
  • the method for adjusting hyperparameters is also not particularly limited as long as it produces the effects of the present invention.
  • the generalization performance confirmation step 81 is a step of confirming the generalization performance of the analysis result of the data analysis step 71.
  • the generalized performance means the performance that the constructed prediction model can be used for quality measurement in in-line measurement.
  • a prediction model (optimal prediction model) can be efficiently selected by making a judgment in two stages of analysis of training data and analysis of test data.
  • R 2 coefficient of determination
  • R correlation coefficient
  • MSE Mean Square Error
  • RMSE Root Mean Square Error
  • MAE Mean Absolute Error
  • AIC Akaike' Information criteria
  • BIC Bayesian Information Criterion
  • the generalization performance confirmation step 81 among the prediction model candidates whose training data score and test data score are above a certain level, the one with the highest test data score is determined as the prediction model. If there is a prediction model candidate whose training data score and test data score have reached a certain value, the one with the highest test data score is determined as the prediction model, and the prediction model construction process is completed. If not, the process proceeds to the data analysis step 72 described later.
  • the data analysis step 72 is a second data analysis step performed after the data analysis step 71.
  • a part of the data set used in the data analysis step 71 is deleted, or a part of the process data constituting the data set is deleted. Is deleted (the types of process data are reduced), and the data analysis step 72 is performed. That is, the number of data used in the data analysis step 72 is smaller than the number of data used in the data analysis step 71. By reducing the number of data used for analysis, the calculation time can be shortened.
  • the specific method of the data analysis step 72 is the same as that of the data analysis step 71.
  • the data set to be analyzed is divided into two, one is used as training data and the other is used as test data.
  • Various analyzes are performed on the training data and the test data in the same manner as in the data analysis step 71, and a plurality of prediction model candidates are constructed. For the constructed multiple prediction model candidates, the training data score and the test data score are calculated.
  • the generalization performance confirmation step 82 is a step of confirming the generalization performance of the analysis result of the data analysis step 72.
  • the specific method in the generalization performance confirmation step 82 is the same as that in the generalization performance confirmation step 81. If there is a prediction model candidate whose training data score and test data score have reached a certain value, the one with the highest test data score is determined as the prediction model, and the prediction model construction process is completed. If not, the process proceeds to the data analysis step 73 described later.
  • the data analysis step 73 is a third and subsequent data analysis steps performed after the data analysis step 72.
  • the data analysis step 73 In order to use a data set different from the data analysis so far, after appropriately adding or deleting data to the data sets used in the data analysis steps 71 and 72, Perform data analysis. That is, the number of data used in the data analysis step 73 may be larger or smaller than the number of data used in the data analysis steps 71 and 72.
  • additional data is acquired as appropriate. Specifically, a series of operations by the process data acquisition step 61, the time series processing step 62, and the data combination step 63 are performed in the same manner as described above.
  • the specific method of the data analysis step 73 is the same as that of the data analysis steps 71 and 72. That is, the data set to be analyzed is divided into two, one is used as training data and the other is used as test data. Various analyzes are performed on the training data and the test data in the same manner as in the data analysis steps 71 and 72, and prediction model candidates are constructed. The training data score and the test data score are calculated for the constructed multiple prediction model candidates.
  • the generalization performance confirmation step 83 is a step of confirming the generalization performance of the analysis result of the data analysis step 73.
  • the specific determination method in the generalization performance confirmation step 83 is the same as that in the generalization performance confirmation steps 81 and 82. If there is a prediction model candidate whose training data score and test data score have reached a certain value, the one with the highest test data score is determined as the prediction model, and the prediction model construction process is completed. If not, the data analysis step 73 and the generalization performance confirmation step 83 are repeated until a prediction model in which the training data score and the test data score reach a certain numerical value can be constructed.
  • the process moves to the in-line quality measurement process.
  • the in-line quality measurement process will be described.
  • the in-line quality measurement process is a process of estimating a quality measurement value based on a prediction model in a process of manufacturing a product through a plurality of manufacturing processes.
  • the quality measurement value is a value calculated by applying process data to the prediction model, and corresponds to an estimated value of quality data.
  • the in-line quality measurement process typically includes a process data acquisition process 91, a time series processing process 92, a quality measurement value estimation process 93, a quality measurement value transmission process 94, a quality determination process 95, and the like. ..
  • the in-line quality measurement step is not limited to the form shown in FIG. 4 as long as the effect of the present invention is exhibited.
  • process data acquisition and time series processing process In the process data acquisition process 91 and the time-series processing process 92 in the in-line quality measurement process, process data acquisition and time delay correction are performed in the same manner as in the process data acquisition process 61 and the time-series processing process 62 in the prediction model construction process. I do. At that time, it is preferable to manufacture the product with the same manufacturing device as the manufacturing device used in the prediction model construction process. By using the same device, the prediction model can be applied to the in-line quality measurement process.
  • the quality measurement value estimation step 93 is a step of estimating the quality measurement value by combining the process data with the prediction model. Specifically, the process data acquired in the process data acquisition process 91 is applied to the prediction model determined in the prediction model construction process to calculate the quality measurement value. In this way, values that are difficult to measure in-line can be estimated.
  • the quality measurement value transmission step 94 is a step of transmitting the quality measurement value calculated in the quality measurement value estimation step 93 to a CPU (Central Processing Unit, for example, a CPU 13 on the manufacturing apparatus side described later) directly connected to the manufacturing apparatus 11.
  • a CPU Central Processing Unit, for example, a CPU 13 on the manufacturing apparatus side described later
  • the quality measurement value transmission step 94 may not be necessary. The form of data exchange will be described later.
  • the quality determination step 95 is a step of determining whether or not the manufactured product has a certain level of quality or not by using the quality measurement value as an index. Therefore, in the present invention, it is preferable to determine the quality of the quality measurement value by means such as providing the quality determination step 95. By confirming whether the quality measurement value is within a certain numerical range, it is possible to confirm whether the product has a certain quality or higher without performing a sampling inspection of the product.
  • the quality measurement value can be kept within a certain numerical range by appropriately changing the setting value (operation data, etc.) of the manufacturing equipment. In this way, it becomes possible to receive the results of online measurement and provide appropriate feedback to stabilize the quality of the product above a certain level and improve the productivity of the product. As a result, the automatic operation of the production line becomes easy.
  • the manufacturing apparatus issues a control command based on the quality measurement value. For example, when the quality measurement value deviates significantly from a certain numerical range, the manufacturing apparatus can be instructed to stop the manufacturing line, so that the manufacturing line can be stopped quickly and automatically when a defective product occurs. In addition, it is possible to notify the operator of an abnormality at an early stage when the production line is stopped. By taking early action when an abnormality occurs, the quality of the product can be kept constant. On the contrary, while the normal quality measurement value is calculated, the operation continuation command of the production line can be issued, and the automatic operation of the production line becomes possible.
  • the manufacturing apparatus 11 shown in FIG. 2 is an apparatus for manufacturing a composite sheet 41 by sticking two sheets 42 and 43 together by a seal bond.
  • the manufacturing apparatus 11 is composed of a seal unit 21, a pattern roll 22, an anvil roll 23, motors 24A to 24E, conveyors 25A to 25H, and the like.
  • the manufacturing apparatus 11 manufactures a product, that is, a composite sheet 41, through a plurality of manufacturing steps.
  • Specific examples of the manufacturing method using the manufacturing apparatus 11 include a step of feeding out the sheets 42 and 43 from the sheet raw fabrics 44 and 45, a step of transporting the fed sheets 42 and 43 by conveyors 25A to 25H, and a step of transporting the fed sheets 42 and 42.
  • Examples thereof include a step of laminating 43 inside the seal unit 21 to form a composite sheet 41, a step of cutting the composite sheet 41 with a cutter for each individual product, and the like.
  • the manufacturing apparatus 11 can acquire process data by providing various sensors. Thickness sensors 31A to 31C, tension sensors 32A and 32B, temperature sensors 33A to 33C, sheet raw fabric sensors 34A and 34B, camera 35 and the like are arranged in the manufacturing apparatus 11 as sensors for acquiring process data by in-line measurement. Will be done. The number and types of sensors are not particularly limited as long as the effects of the present invention can be obtained. In the composite sheet manufacturing apparatus 11, a fiber sheet whose shape tends to be unstable is processed. Therefore, it is preferable that paired sensors (for example, tension sensors 32A and 32B) are arranged on the upstream side and the downstream side of the seal unit 21 for sealing.
  • paired sensors for example, tension sensors 32A and 32B
  • another thickness sensor is arranged on the downstream side of the seal unit 21 in addition to the thickness sensors 31A to 31C.
  • the process data acquired by various sensors is transmitted to the quality measurement device to perform the above-mentioned time series processing, data combination, data analysis, and the like.
  • the quality measurement method of the present invention can be used in various product manufacturing devices in addition to the composite sheet manufacturing device 11.
  • an apparatus for producing a paint by mixing iron powder, activated carbon and water, which are raw materials for the paint (hereinafter, also referred to as “paint manufacturing apparatus”) can be mentioned.
  • the paint manufacturing apparatus is composed of a compounding tank, a stirring blade, a motor, and the like, and manufactures paint through a plurality of manufacturing processes.
  • Specific examples of the manufacturing method using this paint manufacturing apparatus include a step of adding iron powder to the blending tank, a step of adding activated charcoal to the blending tank, a step of adding water to the blending tank, and a step of stirring the raw materials with a stirring blade.
  • the process including the step of applying the manufactured paint to the base material sheet and the like. By these steps, for example, a sheet product having suitable heat generation characteristics can be produced.
  • process data such as iron powder addition amount, activated charcoal addition amount, water addition amount, iron powder particle size, activated charcoal particle size, stirring blade rotation speed, stirring blade stop time, etc. Therefore, quality measurement values such as the viscosity and solid content of the paint can be calculated.
  • the quality measuring device used in the quality measuring method of the present invention will be described with reference to FIG.
  • the quality measuring device is not limited to the form shown in FIG. 5 as long as the effect of the present invention is exhibited.
  • the quality measuring device 12 shown in FIG. 5A has a manufacturing device side CPU 13, a data collecting CPU 14, and an analysis server 15.
  • the quality measuring device 12 arranges various sensors (not shown) in the manufacturing device 11 as means for performing in-line measurement.
  • the manufacturing apparatus side CPU 13 acquires process data 1 from various sensors arranged in the manufacturing apparatus 11 at any time. Further, the manufacturing apparatus side CPU 13 issues a control command 4 of the manufacturing line to the manufacturing apparatus 11 based on the quality measurement value 3.
  • the threshold value of the quality measurement value 3 is set by the CPU 13 on the manufacturing apparatus side, and when the quality measurement value 3 reaches outside the threshold range, the CPU 13 on the manufacturing apparatus side issues a product discharge command or equipment to the manufacturing apparatus 11. Send a stop command, etc.
  • the manufacturing apparatus side CPU 13 also exchanges data with the data collecting CPU 14. Specifically, the data collection CPU 14 acquires the process data 1 from the manufacturing apparatus side CPU 13. Further, the data collection CPU 14 also exchanges data with the analysis server 15 and transfers the process data 1 and the like to the analysis server 15.
  • the analysis server 15 process data acquisition process 61, time series processing process 62, data combination process 63, end determination process 64, data analysis process 71 to 73, generalization performance confirmation process 81 to 83, process data acquisition process 91, time.
  • the serial processing step 92 and the quality measurement value estimation step 93 are executed.
  • the quality data 2 used for combining with the process data is measured offline by a sampling inspection device 16 such as a tensile tester and stored in the analysis server 15.
  • the analysis server 15 builds a prediction model between the quality data 2 and the process data 1.
  • the analysis server 15 calculates the quality measurement value 3 from the process data 1 by the in-line measurement based on the prediction model.
  • the quality measurement value 3 is transferred from the analysis server 15 to the data collection CPU 14, and further transferred to the manufacturing apparatus side CPU 13.
  • FIG. 5 (B) is different from the form shown in FIG. 5 (A), and does not have the data acquisition CPU 14. That is, in the form shown in FIG. 5B, the analysis server 15 also plays the role of the data collection CPU 14. Other points are the same as those shown in FIG. 5 (A).
  • the present invention further discloses the following quality measurement method, product manufacturing method, and quality measurement device.
  • a quality measurement method in the process of manufacturing a product through multiple manufacturing processes.
  • the process of building a forecast model between quality data and process data A quality measurement method including a step of calculating a quality measurement value of a manufactured product from the process data acquired by in-line measurement based on the prediction model.
  • ⁇ 2> The quality measurement method according to ⁇ 1>, wherein the design of experiments method is used for constructing the prediction model.
  • ⁇ 3> The quality measurement method according to ⁇ 1> or ⁇ 2>, wherein a plurality of prediction model candidates are constructed in the construction of the prediction model, and the prediction model candidate having the highest generalization performance is determined as the prediction model.
  • ⁇ 4> The quality measurement method according to any one of ⁇ 1> to ⁇ 3>, wherein the process data is processed in time series in the construction of the prediction model.
  • ⁇ 5> The quality measurement method according to any one of ⁇ 1> to ⁇ 4>, wherein the quality data is measured by sampling inspection.
  • ⁇ 6> The quality measurement method according to any one of ⁇ 1> to ⁇ 5>, which determines the quality of the quality measurement value.
  • ⁇ 7> The quality measurement method according to any one of ⁇ 1> to ⁇ 6>, wherein a control command is issued based on the quality measurement value.
  • ⁇ 8> The quality measurement method according to any one of ⁇ 1> to ⁇ 7>, wherein the product is a composite sheet.
  • ⁇ 9> The quality measurement method according to any one of ⁇ 1> to ⁇ 8>, wherein the process data includes sheet tension, sheet thickness, or both.
  • ⁇ 10> The quality measurement method according to any one of ⁇ 1> to ⁇ 9>, wherein the quality measurement value is the seal strength of a portion where two or more sheets are joined.
  • ⁇ 11> The quality measuring method according to ⁇ 10>, wherein the bonding is performed by heat sealing, ultrasonic sealing, laser welding, or hot melt bonding.
  • ⁇ 12> The quality measurement method according to any one of ⁇ 1> to ⁇ 11>, wherein the CPU on the manufacturing apparatus side acquires the process data from a sensor arranged in the manufacturing apparatus.
  • ⁇ 13> The quality measurement method according to ⁇ 12>, wherein the CPU on the manufacturing apparatus side issues a control command for the manufacturing line to the manufacturing apparatus based on the quality measurement value.
  • ⁇ 14> Any one of ⁇ 1> to ⁇ 13>, wherein the process for constructing the prediction model includes a process data acquisition process, a time series processing process, a data combination process, an end determination process, a data analysis process, and a generalization performance confirmation process.
  • ⁇ 15> The quality measurement method according to ⁇ 14>, wherein in the process data acquisition step, material data and device data are acquired together as the process data.
  • the time-series processing step is a step of correcting a time delay between the process data and storing the data in the analysis server.
  • the quality measurement method according to any one of ⁇ 14> to ⁇ 16>, wherein the data combining step is a step of associating and combining the process data of each product with the quality data of each product.
  • the end determination step is a step of determining whether or not the number of data acquired through the process data acquisition step, the time series processing step, and the data combination step is sufficient for use in data analysis.
  • the quality measurement method according to any one of ⁇ 14> to ⁇ 17>.
  • ⁇ 19> The quality measurement method according to any one of ⁇ 14> to ⁇ 18>, wherein in the generalization performance confirmation step, the prediction model is determined in two stages of analysis of learning data and analysis of test data.
  • ⁇ 20> The quality measurement method according to any one of ⁇ 1> to ⁇ 19>, wherein when the step of constructing the prediction model is completed, the step of calculating the quality measurement value is started.
  • ⁇ 21> The quality measurement method according to any one of ⁇ 3> to ⁇ 20>, wherein in the step of calculating the quality measurement value, the quality measurement value is estimated based on the prediction model.
  • the quality measurement method described in. ⁇ 23> The quality measurement method according to ⁇ 22>, wherein in the process data acquisition step and the time series processing step, the process data is acquired and the time delay is corrected.
  • the quality measurement value estimation step is a step of estimating the quality measurement value by combining the process data with the prediction model.
  • the quality measurement value transmission step is a step of transmitting the quality measurement value to a CPU directly connected to a manufacturing apparatus.
  • the quality determination step is a step of determining whether or not the manufactured product has a certain level of quality or not using the quality measurement value as an index. Described quality measurement method.
  • a method for manufacturing a product which comprises the quality measurement method according to any one of ⁇ 1> to ⁇ 26>.
  • a quality measuring device used for equipment that manufactures products through multiple manufacturing processes has an analysis server and The analysis server builds a prediction model between quality data and process data, A quality measurement device in which the analysis server calculates a quality measurement value from the process data by in-line measurement based on the prediction model.
  • the quality measuring device which has a CPU on the control device side and a CPU for collecting data.
  • the analysis server executes a process data acquisition process, a time series processing process, a data combination process, an end determination process, a data analysis process, a generalization performance confirmation process, and a quality measurement value estimation process, ⁇ 28> or ⁇ 29>.
  • the composite sheet 41 was manufactured using the manufacturing apparatus 11 shown in FIG. 2, various sensors, and the quality measuring apparatus 12 shown in FIG. 5 (A), and the seal strength was estimated as in Examples 1 to 3 below.
  • the operation data were oil pressure, pattern roll temperature, anvil roll temperature, dancer pressure, sheet raw material storage humidity, and conveyor transfer speed.
  • Process data other than operation data includes motor shaft rotation speed at 5 locations, motor shaft load at 5 locations, entry side seat temperature, exit side seat temperature, seal unit temperature, upper shaft seat original fabric diameter, and lower shaft seat original. The reverse diameter, the entry side seat tension, the exit side seat tension, and the seat thickness at three locations were used. That is, a total of 26 types of process data including operation data were used.
  • the quality data were the inlet side seal strength and the outlet side seal strength.
  • the “entry side” refers to the downstream side of the manufacturing apparatus 11 in the machine flow direction
  • the “exit side” refers to the upstream side of the manufacturing apparatus 11 in the machine flow direction.
  • the prediction model candidates were constructed and the prediction model was determined.
  • the manufacturing apparatus 11 shown in FIG. 2 was used to attach the sheets for the absorbent article while acquiring the process data to manufacture the composite sheet.
  • the acquired process data was processed in time series to correct the time delay between the sensors.
  • all composite sheets were sampled and inspected, and quality data was measured.
  • a tensile tester was used to measure the inlet-side seal strength and the outlet-side seal strength, which are quality data.
  • 40 sets of data sets were acquired at one time. The above operation was repeated 18 times, and a total of 720 sets of data sets were acquired. During repeated use of the experimental design, it was assigned process data L 18 orthogonal table.
  • the 720 sets of data sets were divided into 360 sets of training data and 360 sets of test data, and linear analysis and non-linear analysis were performed on each of them to construct a plurality of prediction model candidates.
  • the operation of acquiring the data set was repeated three more times, and 120 sets of data sets were additionally acquired.
  • a total of 840 sets of data sets were divided into 420 sets of training data and 420 sets of test data, and the first data analysis was performed for each.
  • linear analysis and non-linear analysis were performed, and a plurality of prediction model candidates were constructed.
  • the lasso regression was performed assuming that the R 2 values of both the training data and the test data exceeded 0.6, so the model by the lasso regression has sufficient generalization performance. I decided. Further, the model according to lasso regression for the R 2 of the test data score the highest in comparison with other predictive model candidates, to determine the model by lasso regression prediction model.
  • Example 1 According to the flow chart shown in FIG. 4, the sheets for absorbent articles were laminated under the conditions of the pattern roll temperature of 125 ° C. and the anvil roll temperature of 170 ° C. in the manufacturing apparatus used for acquiring the data set, and a total of 100 composite sheets were attached. Manufactured. The operation data was not changed during manufacturing. The process data acquired at the time of bonding was processed in time series together with the operation data to correct the time delay between the sensors. The lasso regression determined in the prediction model was applied to the process data after the time-series processing, and the values of the inlet-side seal strength and the outlet-side seal strength were estimated as quality measurement values for each of the 100 composite sheets. The average value of 100 quality measurement values was calculated for each of the inlet side seal strength and the outlet side seal strength.
  • Example 2 Except that the pattern roll temperature was 120 ° C. and the anvil roll temperature was 160 ° C., 100 quality measurement values were estimated for each of the inlet side seal strength and the outlet side seal strength in the same manner as in Example 1, and the average value was calculated. I asked.
  • Example 3 Except that the pattern roll temperature was 115 ° C. and the anvil roll temperature was 150 ° C., 100 quality measurement values were estimated for each of the inlet side seal strength and the outlet side seal strength in the same manner as in Example 1, and the average value was calculated. I asked.
  • the difference between the average value of the quality measurement value of the entry side seal strength and the average value of the measured value of the entry side seal strength is suppressed within 0.3 N / 15 mm. It was found that the inlet seal strength can be estimated with high accuracy by using the prediction model. Further, in Example 1 in which the composite sheet was manufactured by setting the pattern roll temperature and the anvil roll temperature to relatively high temperatures, the inlet-side seal strength could be maintained at a relatively large value. On the other hand, as in Examples 2 and 3, the inner seal strength could be gradually reduced by gradually lowering the pattern roll temperature and the anvil roll temperature. In this way, it was found that the inlet seal strength of the manufactured composite sheet can be adjusted by manipulating the pattern roll temperature and the anvil roll temperature.

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Epidemiology (AREA)
  • Vascular Medicine (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Manufacturing & Machinery (AREA)
  • General Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Quality & Reliability (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • General Factory Administration (AREA)
  • Absorbent Articles And Supports Therefor (AREA)
PCT/JP2020/031469 2019-11-28 2020-08-20 品質計測方法及び品質計測装置 Ceased WO2021106288A1 (ja)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202080082342.6A CN114760969B (zh) 2019-11-28 2020-08-20 质量测量方法和质量测量装置

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2019215774A JP7084907B2 (ja) 2019-11-28 2019-11-28 品質計測方法及び品質計測装置
JP2019-215774 2019-11-28

Publications (1)

Publication Number Publication Date
WO2021106288A1 true WO2021106288A1 (ja) 2021-06-03

Family

ID=76087826

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2020/031469 Ceased WO2021106288A1 (ja) 2019-11-28 2020-08-20 品質計測方法及び品質計測装置

Country Status (3)

Country Link
JP (1) JP7084907B2 (https=)
CN (1) CN114760969B (https=)
WO (1) WO2021106288A1 (https=)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2023028245A (ja) * 2021-08-19 2023-03-03 Ｊｆｅスチール株式会社 操業条件提示方法および操業条件提示装置

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114777714B (zh) * 2022-03-31 2024-03-01 广东康怡卫生用品有限公司 一种卫生用品厚度测量方法
CN115169745A (zh) * 2022-08-12 2022-10-11 芜湖云一新材料科技有限公司 生产质量预测方法、系统及计算机可读介质
KR20260026561A (ko) * 2023-07-05 2026-02-26 제이에프이 스틸 가부시키가이샤 프로세스 데이터의 편집 방법, 프로세스의 이상 검지 방법, 프로세스 데이터의 편집 장치 및 프로세스의 이상 검지 장치

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08117827A (ja) * 1994-10-24 1996-05-14 Mitsubishi Electric Corp 圧延装置
JP2006024195A (ja) * 2004-06-03 2006-01-26 National Cheng Kung Univ 生産工程の品質予測システムおよびその方法
JP2007258731A (ja) * 2007-04-23 2007-10-04 Canon System Solutions Inc プロセスと品質との関係についてのモデル作成装置及びモデル作成方法
US20130339919A1 (en) * 2012-06-18 2013-12-19 International Business Machines Corporation Method and Apparatus for Hierarchical Wafer Quality Predictive Modeling
JP2016538652A (ja) * 2013-09-03 2016-12-08 ザ プロクター アンド ギャンブル カンパニー 吸収性製品加工ラインの目標製造パラメータを調整するためのシステム及び方法
JP2017111827A (ja) * 2015-12-18 2017-06-22 フィッシャー−ローズマウント システムズ，インコーポレイテッド 継続的プロセス検証（ｃｐｖ）用の分析／統計モデリングを使用するための方法及び装置、製品
JP2018129030A (ja) * 2016-11-25 2018-08-16 花王株式会社 製品の製造方法
JP2019074969A (ja) * 2017-10-17 2019-05-16 新日鐵住金株式会社 品質予測装置及び品質予測方法
CN109816191A (zh) * 2017-11-20 2019-05-28 财团法人资讯工业策进会 多工作站系统的质量预测方法及其系统

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000252179A (ja) * 1999-03-04 2000-09-14 Hitachi Ltd 半導体製造プロセス安定化支援システム
JP4296160B2 (ja) 2005-03-29 2009-07-15 株式会社東芝 回路基板の品質解析システム及び品質解析方法
JP2009099745A (ja) * 2007-10-16 2009-05-07 Toshiba Corp 半導体装置の生産管理装置及び生産管理方法
US8145338B2 (en) * 2009-06-02 2012-03-27 The Procter & Gamble Company Systems and methods for detecting and rejecting defective absorbent articles from a converting line
JP2011242923A (ja) 2010-05-17 2011-12-01 Fuji Electric Co Ltd モデルの入力変数調整装置
JP5014500B1 (ja) 2011-04-04 2012-08-29 シャープ株式会社 異常要因特定方法および装置、上記異常要因特定方法をコンピュータに実行させるためのプログラム、並びに上記プログラムを記録したコンピュータ読み取り可能な記録媒体
US9110452B2 (en) * 2011-09-19 2015-08-18 Fisher-Rosemount Systems, Inc. Inferential process modeling, quality prediction and fault detection using multi-stage data segregation
GB201512819D0 (en) * 2015-07-21 2015-09-02 Scaife Martin Customised fast moving consumer goods production system
CN105136449B (zh) * 2015-08-24 2018-05-18 哈尔滨工程大学 一种基于磨损机理的车用传动装置的磨损随机过程试验预测方法
JP6746139B2 (ja) * 2016-09-08 2020-08-26 公立大学法人会津大学 携帯端末を用いた察知エージェントシステム、察知エージェントシステムにおける機械学習方法、及びこれを実施するためのプログラム
CN109472057B (zh) * 2018-10-16 2023-05-23 浙江大学 基于跨工步隐式参数记忆的产品加工质量预测装置及方法
CN109659933B (zh) * 2018-12-20 2022-03-18 浙江工业大学 一种基于深度学习模型的含分布式电源配电网电能质量预测方法
CN109711714B (zh) * 2018-12-24 2023-01-10 浙江大学 基于并联长短期记忆网络的制造装配产品质量预测方法
CN109483816A (zh) * 2018-12-27 2019-03-19 东莞市誉铭新精密技术股份有限公司 一种手机塑料壳体注塑工艺及注塑装置

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08117827A (ja) * 1994-10-24 1996-05-14 Mitsubishi Electric Corp 圧延装置
JP2006024195A (ja) * 2004-06-03 2006-01-26 National Cheng Kung Univ 生産工程の品質予測システムおよびその方法
JP2007258731A (ja) * 2007-04-23 2007-10-04 Canon System Solutions Inc プロセスと品質との関係についてのモデル作成装置及びモデル作成方法
US20130339919A1 (en) * 2012-06-18 2013-12-19 International Business Machines Corporation Method and Apparatus for Hierarchical Wafer Quality Predictive Modeling
JP2016538652A (ja) * 2013-09-03 2016-12-08 ザ プロクター アンド ギャンブル カンパニー 吸収性製品加工ラインの目標製造パラメータを調整するためのシステム及び方法
JP2017111827A (ja) * 2015-12-18 2017-06-22 フィッシャー−ローズマウント システムズ，インコーポレイテッド 継続的プロセス検証（ｃｐｖ）用の分析／統計モデリングを使用するための方法及び装置、製品
JP2018129030A (ja) * 2016-11-25 2018-08-16 花王株式会社 製品の製造方法
JP2019074969A (ja) * 2017-10-17 2019-05-16 新日鐵住金株式会社 品質予測装置及び品質予測方法
CN109816191A (zh) * 2017-11-20 2019-05-28 财团法人资讯工业策进会 多工作站系统的质量预测方法及其系统

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2023028245A (ja) * 2021-08-19 2023-03-03 Ｊｆｅスチール株式会社 操業条件提示方法および操業条件提示装置
JP7493128B2 (ja) 2021-08-19 2024-05-31 Ｊｆｅスチール株式会社 操業条件提示方法および操業条件提示装置

Also Published As

Publication number Publication date
JP7084907B2 (ja) 2022-06-15
CN114760969A (zh) 2022-07-15
JP2021086457A (ja) 2021-06-03
CN114760969B (zh) 2023-04-25

Similar Documents

Publication Publication Date Title
JP7084907B2 (ja) 品質計測方法及び品質計測装置
US11628502B2 (en) Method of feedback controlling 3D printing process in real-time and 3D printing system for the same
KR102580590B1 (ko) 제조 공정을 모니터링하기 위한 방법, 체계적인 의존성을 간접적으로 유도하기 위한 방법, 품질 적합화 방법, 제조 공정을 시작하기 위한 방법, 압출 제품을 제조하기 위한 방법 및 압출 제품을 제조하기 위한 시스템
EP2286935B1 (en) Steel plate quality assurance system and method
AT513245B1 (de) Planheitsmessung und Messung der Eigenspannungen für ein metallisches Flachprodukt
US6600961B2 (en) Intelligent control method for injection machine
JP5685208B2 (ja) 薄板用熱間圧延機の制御装置および薄板用熱間圧延機の制御方法
JP2021536063A (ja) 金属製品の製造をモニタリングするための方法および電子デバイス、関連するコンピュータプログラムならびに設備
CN110402172A (zh) 用于运行辊式矫直机的方法和辊式矫直机
KR101877341B1 (ko) 교정 롤 시스템의 교정 롤 정렬 방법
US10095199B2 (en) Energy consumption predicting device for rolling line
CN113780852B (zh) 一种板带轧制过程质量缺陷的诊断方法
JP2025004220A (ja) 解析装置、解析方法および解析プログラム
Drossel et al. Modular system to measure and control the force distribution in deep drawing processes to ensure part quality and process reliability
JP2021086457A5 (https=)
US20150023387A1 (en) Steel plate quality assurance system and equipment thereof
JP7135962B2 (ja) 鋼板の仕上出側温度制御方法、鋼板の仕上出側温度制御装置、及び鋼板の製造方法
Brik et al. Assessing process capability index using sampling plan in the presence of measurement system errors
CN102601549B (zh) 焊剂填充率的判定装置、判定方法、判定系统
CN115121626A (zh) 一种基于误差补偿的热轧带钢瞬态热辊型预报方法
KR100660231B1 (ko) 연속압연공정의 강판의 최적 목표형상 설정방법
Jin et al. Quality prediction and control in rolling processes using logistic regression
CN105642676B (zh) 一种利用侧导板辊缝值进行宽度模型控制的方法
EP4596130A1 (en) Method for predicting width of rough rolled material, method for controlling width of rough rolled material, method for producing hot-rolled steel sheet, and method for generating width prediction model for rough rolled material
Mayr et al. Towards an intelligent straightening system for flat enameled copper wire: problem statement, review of related work, and basic concept

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20891455

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20891455

Country of ref document: EP

Kind code of ref document: A1