WO2022270506A1 - 塗膜の性状の変動量の予測方法及び予測システム、塗布物の製造条件の変動量の予測方法及び予測システム、塗布物の製造方法 - Google Patents
塗膜の性状の変動量の予測方法及び予測システム、塗布物の製造条件の変動量の予測方法及び予測システム、塗布物の製造方法 Download PDFInfo
- Publication number
- WO2022270506A1 WO2022270506A1 PCT/JP2022/024739 JP2022024739W WO2022270506A1 WO 2022270506 A1 WO2022270506 A1 WO 2022270506A1 JP 2022024739 W JP2022024739 W JP 2022024739W WO 2022270506 A1 WO2022270506 A1 WO 2022270506A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- amount
- variation
- coating film
- manufacturing conditions
- manufacturing
- Prior art date
Links
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 358
- 238000000576 coating method Methods 0.000 title claims abstract description 276
- 239000011248 coating agent Substances 0.000 title claims abstract description 272
- 238000000034 method Methods 0.000 title claims abstract description 148
- 230000008859 change Effects 0.000 title claims abstract description 48
- 238000013473 artificial intelligence Methods 0.000 claims abstract description 109
- 238000010801 machine learning Methods 0.000 claims abstract description 39
- 239000003973 paint Substances 0.000 claims description 93
- 230000008569 process Effects 0.000 claims description 56
- 239000000463 material Substances 0.000 claims description 34
- 238000005259 measurement Methods 0.000 claims description 31
- 238000004422 calculation algorithm Methods 0.000 claims description 19
- 230000002093 peripheral effect Effects 0.000 claims description 11
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 239000007788 liquid Substances 0.000 claims description 6
- 230000005856 abnormality Effects 0.000 claims description 5
- 230000000007 visual effect Effects 0.000 claims description 5
- 238000007637 random forest analysis Methods 0.000 claims description 4
- 238000003066 decision tree Methods 0.000 claims description 3
- 238000010438 heat treatment Methods 0.000 claims description 3
- 238000010238 partial least squares regression Methods 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 2
- 230000010411 postconditioning Effects 0.000 claims 1
- 239000008199 coating composition Substances 0.000 description 14
- 230000000052 comparative effect Effects 0.000 description 12
- VYPSYNLAJGMNEJ-UHFFFAOYSA-N silicon dioxide Inorganic materials O=[Si]=O VYPSYNLAJGMNEJ-UHFFFAOYSA-N 0.000 description 11
- 238000000691 measurement method Methods 0.000 description 10
- HJOVHMDZYOCNQW-UHFFFAOYSA-N isophorone Chemical compound CC1=CC(=O)CC(C)(C)C1 HJOVHMDZYOCNQW-UHFFFAOYSA-N 0.000 description 8
- 238000002360 preparation method Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 239000000203 mixture Substances 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000010422 painting Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 239000011347 resin Substances 0.000 description 6
- 229920005989 resin Polymers 0.000 description 6
- 238000001228 spectrum Methods 0.000 description 6
- 239000003607 modifier Substances 0.000 description 5
- 238000010606 normalization Methods 0.000 description 5
- 239000000049 pigment Substances 0.000 description 5
- 238000007781 pre-processing Methods 0.000 description 5
- 238000003860 storage Methods 0.000 description 5
- UQSXHKLRYXJYBZ-UHFFFAOYSA-N Iron oxide Chemical compound [Fe]=O UQSXHKLRYXJYBZ-UHFFFAOYSA-N 0.000 description 4
- 239000003086 colorant Substances 0.000 description 4
- JHIVVAPYMSGYDF-UHFFFAOYSA-N cyclohexanone Chemical compound O=C1CCCCC1 JHIVVAPYMSGYDF-UHFFFAOYSA-N 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 239000002994 raw material Substances 0.000 description 4
- 239000000377 silicon dioxide Substances 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 239000004925 Acrylic resin Substances 0.000 description 3
- 229920000178 Acrylic resin Polymers 0.000 description 3
- GWEVSGVZZGPLCZ-UHFFFAOYSA-N Titan oxide Chemical compound O=[Ti]=O GWEVSGVZZGPLCZ-UHFFFAOYSA-N 0.000 description 3
- 239000006229 carbon black Substances 0.000 description 3
- 238000010227 cup method (microbiological evaluation) Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 230000001678 irradiating effect Effects 0.000 description 3
- 238000002156 mixing Methods 0.000 description 3
- 239000003960 organic solvent Substances 0.000 description 3
- 238000007665 sagging Methods 0.000 description 3
- 239000007787 solid Substances 0.000 description 3
- 239000002904 solvent Substances 0.000 description 3
- 239000000654 additive Substances 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 238000001035 drying Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000009472 formulation Methods 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 235000012239 silicon dioxide Nutrition 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- OGIDPMRJRNCKJF-UHFFFAOYSA-N titanium oxide Inorganic materials [Ti]=O OGIDPMRJRNCKJF-UHFFFAOYSA-N 0.000 description 2
- 239000004034 viscosity adjusting agent Substances 0.000 description 2
- 239000012463 white pigment Substances 0.000 description 2
- BQCIDUSAKPWEOX-UHFFFAOYSA-N 1,1-Difluoroethene Chemical compound FC(F)=C BQCIDUSAKPWEOX-UHFFFAOYSA-N 0.000 description 1
- YCKRFDGAMUMZLT-UHFFFAOYSA-N Fluorine atom Chemical compound [F] YCKRFDGAMUMZLT-UHFFFAOYSA-N 0.000 description 1
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 229910000831 Steel Inorganic materials 0.000 description 1
- RTAQQCXQSZGOHL-UHFFFAOYSA-N Titanium Chemical compound [Ti] RTAQQCXQSZGOHL-UHFFFAOYSA-N 0.000 description 1
- 235000010724 Wisteria floribunda Nutrition 0.000 description 1
- 229910000611 Zinc aluminium Inorganic materials 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 229910045601 alloy Inorganic materials 0.000 description 1
- 239000000956 alloy Substances 0.000 description 1
- HXFVOUUOTHJFPX-UHFFFAOYSA-N alumane;zinc Chemical compound [AlH3].[Zn] HXFVOUUOTHJFPX-UHFFFAOYSA-N 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 239000011324 bead Substances 0.000 description 1
- 229910052791 calcium Inorganic materials 0.000 description 1
- 239000011362 coarse particle Substances 0.000 description 1
- 238000009500 colour coating Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 239000002612 dispersion medium Substances 0.000 description 1
- 238000004070 electrodeposition Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000011737 fluorine Substances 0.000 description 1
- 229910052731 fluorine Inorganic materials 0.000 description 1
- -1 for example Substances 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- YOBAEOGBNPPUQV-UHFFFAOYSA-N iron;trihydrate Chemical compound O.O.O.[Fe].[Fe] YOBAEOGBNPPUQV-UHFFFAOYSA-N 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 239000006224 matting agent Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
- 239000001054 red pigment Substances 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 238000005488 sandblasting Methods 0.000 description 1
- 238000005507 spraying Methods 0.000 description 1
- 238000010186 staining Methods 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 235000013759 synthetic iron oxide Nutrition 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000002834 transmittance Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
- 239000001052 yellow pigment Substances 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B05—SPRAYING OR ATOMISING IN GENERAL; APPLYING FLUENT MATERIALS TO SURFACES, IN GENERAL
- B05C—APPARATUS FOR APPLYING FLUENT MATERIALS TO SURFACES, IN GENERAL
- B05C11/00—Component parts, details or accessories not specifically provided for in groups B05C1/00 - B05C9/00
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B05—SPRAYING OR ATOMISING IN GENERAL; APPLYING FLUENT MATERIALS TO SURFACES, IN GENERAL
- B05C—APPARATUS FOR APPLYING FLUENT MATERIALS TO SURFACES, IN GENERAL
- B05C11/00—Component parts, details or accessories not specifically provided for in groups B05C1/00 - B05C9/00
- B05C11/10—Storage, supply or control of liquid or other fluent material; Recovery of excess liquid or other fluent material
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B05—SPRAYING OR ATOMISING IN GENERAL; APPLYING FLUENT MATERIALS TO SURFACES, IN GENERAL
- B05C—APPARATUS FOR APPLYING FLUENT MATERIALS TO SURFACES, IN GENERAL
- B05C5/00—Apparatus in which liquid or other fluent material is projected, poured or allowed to flow on to the surface of the work
- B05C5/02—Apparatus in which liquid or other fluent material is projected, poured or allowed to flow on to the surface of the work the liquid or other fluent material being discharged through an outlet orifice by pressure, e.g. from an outlet device in contact or almost in contact, with the work
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B05—SPRAYING OR ATOMISING IN GENERAL; APPLYING FLUENT MATERIALS TO SURFACES, IN GENERAL
- B05D—PROCESSES FOR APPLYING FLUENT MATERIALS TO SURFACES, IN GENERAL
- B05D3/00—Pretreatment of surfaces to which liquids or other fluent materials are to be applied; After-treatment of applied coatings, e.g. intermediate treating of an applied coating preparatory to subsequent applications of liquids or other fluent materials
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
Definitions
- the present invention relates to a method and system for predicting the amount of variation in properties of a coating film, a method and system for predicting the amount of variation in manufacturing conditions of a coated product, and a method for manufacturing a coated product.
- the standard conditions and the previous conditions are, for example, recorded past data such as a work management table, etc., and by referring to it, it is possible to approach the target coating film properties to some extent. It has been done by a method that involves work. In this method, the coating manager must manually adjust the manufacturing conditions such as coating pressure, line speed, dispensing pressure, voltage, and temperature. By repeating this operation until the target film thickness and the color difference and gloss difference from the target coated plate become equal to or less than a certain value, a coating film having predetermined coating film properties can be obtained.
- the present invention provides a method and system for predicting the amount of change in the properties of the coating film, which can quickly predict the amount of change in the properties of the coating film, and the manufacturing conditions can be adjusted regardless of the range of standard values for quality. It is an object of the present invention to provide a method and system for predicting the amount of variation in manufacturing conditions of a coated product, and a method for manufacturing a coated product that can produce a coated product having a desired coating film quality.
- the gist and configuration of the present invention are as follows.
- the target A method for predicting the optimum amount of variation in the manufacturing conditions to achieve an artificial intelligence model acquisition step of acquiring a predetermined artificial intelligence model that includes the amount of variation in manufacturing conditions as an input and the amount of variation in the properties of the coating film as an output;
- a property variation amount prediction step of predicting the amount of variation in the properties of the coating film by inputting the amount of variation in the manufacturing conditions in the artificial intelligence model by a computer, In the property variation amount prediction step, a plurality of variation amounts of the manufacturing conditions are input, and a plurality of coating film property variation amounts corresponding to the variation amounts of the respective manufacturing conditions are predicted, A manufacturing condition variation amount, wherein the optimal variation amount of the manufacturing condition is determined as the optimal variation amount of the manufacturing condition for achieving the target based on the predicted variation amount of the properties of the plurality of coating films.
- (3) further comprising a relational data input step of inputting into the computer relational data indicating the relation between the amount of variation in the manufacturing conditions and the amount of variation in properties of the coating film;
- the artificial intelligence model obtaining step is performed by machine learning using the relational data input in the relational data input step as teacher data by the computer.
- the properties of the coating film include any one or more of color, gloss, viscosity, film thickness, smoothness, visual appearance, flip-flop properties, and coating film abnormality of the coating film, above (1)
- the manufacturing conditions are (a) any one or more of viscosity and heating residue of the paint, (b) in the step of applying paint to the object, any one or more of the peripheral speed of the roll, the application pressure to the object, and the flow rate of the paint; (c) any one or more of the baking temperature and baking time in the baking process, and (d) any one or more of the temperature, humidity and paint liquid temperature of the production line,
- the method according to any one of claims 1 to 4 comprising any one or more of (a) to (d) of
- the artificial intelligence model acquisition step is a step of creating the artificial intelligence model by machine learning, and the machine learning is a neural network, an ensemble tree composed of a plurality of decision trees, or partial least squares regression.
- the method according to any one of (1) to (5) above, which uses a predictive algorithm based on the method.
- another optimum amount of variation in the manufacturing condition is determined as the amount of variation in another optimal manufacturing condition for achieving the target.
- a manufacturing condition variation amount prediction sub-step a post-condition-adjusted manufacturing sub-process of adjusting the manufacturing conditions by the predicted amount of variation in the different manufacturing conditions, applying the paint to the object, and obtaining the coated object;
- a property measurement sub-step of measuring the property of the coating film of the obtained coating When the property of the coating film measured in the property measurement step deviates from the desired property by a predetermined threshold value or more, the artificial intelligence model acquisition substep, the property variation prediction substep, and the manufacturing condition variation amount.
- a system for predicting the amount of variation in the properties of a coating film obtained by applying a coating material to an object an artificial intelligence model acquisition unit that acquires a predetermined artificial intelligence model that includes a variation amount of manufacturing conditions as an input and a variation amount of the property of the coating film as an output;
- a computer comprising a property prediction unit that calculates and predicts the amount of change in the properties of the coating film by inputting the amount of change in the manufacturing conditions.
- the target A system that predicts the optimum amount of variation in the manufacturing conditions to achieve an artificial intelligence model acquisition unit that acquires a predetermined artificial intelligence model that includes the amount of variation in the manufacturing conditions as an input and the amount of variation in the properties of the coating film as an output;
- a property prediction unit that calculates and predicts a plurality of fluctuation amounts of properties of the coating film by inputting a plurality of fluctuation amounts of the manufacturing conditions;
- a manufacturing condition variation amount wherein the optimal variation amount of the manufacturing condition is determined as the optimal variation amount of the manufacturing condition for achieving the target based on the predicted variation amount of the properties of the plurality of coating films.
- a system comprising: a computer comprising a predictor;
- the manufacturing conditions have process manufacturing conditions in each of a plurality of processes, monitoring each of the manufacturing conditions when the manufacturing conditions are adjusted by the amount of variation in the manufacturing conditions predicted by the method described in (2) above; When a process manufacturing condition adjusted by a predetermined value or more from the variation amount is observed among the manufacturing conditions, In the artificial intelligence model, after fixing the monitored process manufacturing conditions, by inputting the fluctuation amounts of the other plurality of the manufacturing conditions, the fluctuation amounts of the properties of the plurality of coating films are calculated and predicted. a property re-prediction step; production condition fluctuation, wherein the optimum production condition fluctuation amount is re-determined as the optimum production condition fluctuation amount for achieving the target based on the predicted fluctuation amounts of the properties of the plurality of coating films.
- a fixed representative absolute value of the coating film properties of the paint before adjustment is added to the amount of variation in the coating film properties before and after adjustment obtained by measurement. Therefore, a mode of using a representative absolute value after variation for learning can also be included as one mode of the above method, for example.
- add it to the fixed representative absolute value of the coating film properties of the paint before adjustment Indirectly obtaining the absolute value of the coating film property can also be included, for example, as one aspect of the above method.
- a method and system for predicting the amount of change in the properties of a coating film which can quickly predict the amount of change in the properties of the coating film, and the manufacturing conditions can be adjusted regardless of the range of standard values for quality. It is possible to provide a method and system for predicting the amount of variation in manufacturing conditions of a coated product, and a method for manufacturing a coated product that can produce a coated product having a desired coating film quality.
- 4 is a flow chart of a method for predicting the amount of change in properties of a coating film according to one embodiment of the present invention.
- 4 is a flow chart of a method for predicting the amount of variation in manufacturing conditions according to one embodiment of the present invention
- 1 is a flow chart of a method for manufacturing an applied article according to an embodiment of the present invention
- 4 is a flow chart showing a sub-flow of a method for manufacturing an applied article according to one embodiment of the present invention.
- 1 is a schematic diagram of a CCL
- FIG. 1 is a block diagram of a prediction system for a coating film property variation amount according to an embodiment of the present invention
- FIG. 1 is a block diagram of a prediction system for the amount of variation in manufacturing conditions according to one embodiment of the present invention
- the acceptability of coating film properties varies depending on the application.
- paint film properties related to color in automobile refinishing, the boundary between the normally painted part of a broken car (equivalent to the target painted board) and the part to be painted with newly adjusted paint for repair after repair. Since gradation can be applied, the degree of color matching is not strictly required to be about ⁇ E ⁇ 0.5.
- coil coating which is used for pre-coating (coating before assembly), allows for color differences due to differences in production lots, as painted plates from different production lots may be assembled side by side during product manufacturing. Many products have extremely small widths, for example, about ⁇ E ⁇ 0.1.
- This color difference level is an area that changes depending on the type of color difference meter and the condition, so the accuracy of the absolute value of color is insufficient. At the same timing, it is necessary to measure the color difference from the standard plate for each production and for each toning, and make adjustments so that the color difference is within a predetermined range from the standard plate.
- the adjustment work as described above depends on the experience of the painting supervisor in many respects, and the process is complicated, so automation of the adjustment work is desired.
- a slight deviation in the prediction accuracy tends to result in a wrong direction of correction.
- the b * value of the target color is 3.0
- the true b * value of the color of the paint before adjustment is 2.9
- the b of the color of the paint before adjustment predicted from the manufacturing conditions If the value is 3.1 (that is, the error is 0.2), the temperature of the drying oven should actually be raised to increase the b * value by 0.1, but the absolute value is predicted to be over 0.1 (yellowish is strong), the temperature of the drying oven must be lowered, and in the absolute value prediction, the color may rather move away from the target color due to the adjustment, and there were cases where fine adjustment was impossible.
- the present invention provides a prediction method and prediction system for the amount of change in the properties of a coating film, a method and a prediction system for the amount of change in the manufacturing conditions of a coated product, which have high prediction accuracy while achieving efficiency by a machine learning technique.
- a prediction system a method for producing a coated product, and a method for producing a coating film, which can bring the properties of the coating film closer to the desired one with high accuracy while achieving efficiency by means of machine learning.
- FIG. 1 is a flowchart of a method for predicting the amount of change in properties of a coating film according to one embodiment of the present invention.
- An embodiment of a method for estimating the amount of variation in the properties of a coating film obtained by applying a coating material to an object will be described below with reference to FIG.
- the method for predicting the amount of change in the property of the coating film of the present embodiment can be executed using, for example, a prediction system for the amount of change in the property of the coating film according to one embodiment of the present invention, which will be described later. .
- the properties of the coating film to be predicted preferably include at least the color of the paint, and also preferably include the gloss and film thickness of the paint in addition to the color.
- the predicted coating film properties include, for example, hiding rate, flip-flop property, smoothness, visual appearance, electrical resistance, contact angle, staining property, solar reflectance, and ultraviolet transmittance. , weather resistance, viscoelasticity, and coating film abnormality.
- the coating pressure is adjusted to change the film thickness, both the color and the gloss change, and when a gloss modifier is added to adjust the gloss, the color also changes. is preferably predicted and adjusted at the same time.
- L * value, a * value, and b * value in L * a * b * color space can be used.
- the color can be measured using a known color measurement method.
- CM-512m3 commercially available from Konica Minolta Co., Ltd. is used, and when the light receiving part perpendicular to the coating film is 0 °, The L * value, a * value, and b * value measured by irradiating the light source at angles of 25°, 45°, and 75° can be measured.
- it can be measured using X-Rite MA68II (manufactured by X-Rite).
- the measurement angle can be adjusted as appropriate according to the purpose or equipment to be used. Any other index can be used. Further, for example, it is reflection spectrum data, and an arbitrary index such as an index in which the reflection spectrum intensity is colored every 5 nm from 380 nm to 780 nm can be used. Gloss is not particularly limited, but gloss can be used as an index. Gloss can be measured using a known gloss measurement method, for example, the 60 ° gloss of the coating film formed on the test plate is measured with a specular gloss meter (gloss meter VG 7000 (Nippon Denshoku Industries Co., Ltd. )) in accordance with JIS K 5600-4-7 (specular glossiness).
- specular gloss meter gloss meter VG 7000 (Nippon Denshoku Industries Co., Ltd. )
- Viscosity can be measured using a known viscosity measurement method, for example, according to JIS K 5600-2-2 (flow cup method). It is preferable to use a wave scan value as an index for smoothness.
- the wave scan values are du (wavelength 0.1 mm or less), Wa (wavelength 0.1 to 0.3 mm), Wb (wavelength 0.3 to 1.0 mm), Wc (wavelength 1.0 to 3.0 mm), Any one of Wd (wavelength 3.0 to 10.0 mm), We (wavelength 10.0 to 30.0 mm), Lw (wavelength 1.2 to 12 mm), and Sw (wavelength 0.3 to 1.2 mm) preferably one or more.
- the wave scan value means that the smaller the value, the less unevenness of the wavelength on the surface, and the better the appearance quality of the coating film.
- the flip-flop property can be measured using a gonio-color difference meter or the like.
- coating film defects include unevenness, blisters, cracks, sagging, pinholes, frames, and the like. These can also be measured by known methods.
- the paint to be used for painting can be one commonly used for that painting, but when preparing and preparing paint, for example, pigments, resins, additives, solvents and / or water are added to SG It can be prepared by adding a resin, a solvent, and an additive to a primary color paint prepared by dispersing with a mill or the like to prepare many types of pigments, dispersing them, and then repeating the adjustment. .
- relationship data indicating the relationship between the amount of variation in the manufacturing conditions and the amount of variation in the properties of the coating film is input into the computer (step S101: relationship data input step).
- Relevant data can be prepared from past data or the like.
- the above manufacturing conditions are manufacturing conditions related to obtaining a coating film by applying paint to an object.
- the manufacturing conditions can be one or more properties of the paint, and specifically preferably one or more of the viscosity and the heat residue of the paint.
- the manufacturing conditions can be the conditions of each step in the process of applying paint to the object. Specifically, in the process of applying paint to the object, if it is roll coating , line speed, roll peripheral speed, coating pressure and paint flow rate, applicator roll condition, material, hardness, pick-up roll type, etc. Roll peripheral speed, coating pressure on the target, and paint flow rate Any one or more is preferable.
- the roll having a peripheral speed is not limited to an application roll, a backup roll, a pickup roll, a metering roll, or the like, but an application roll is particularly preferable.
- the manufacturing conditions can also be the conditions of a process such as a baking process that is performed between processes. and baking time.
- the manufacturing conditions can be environmental conditions in the manufacturing line, for example, any one or more of temperature, humidity, and coating liquid temperature in the manufacturing line.
- the discharge pressure, discharge amount, type of gun, baking temperature, baking time, temperature and humidity in the baking process can be mentioned.
- coating voltage, liquid temperature, energization time, coating direction (horizontal or vertical), etc. can be mentioned.
- the manufacturing conditions are (a) one or more of the viscosity of the paint and the heat residue, (b) in the process of applying the paint to the object, the roll peripheral speed, the object Any one or more of the coating pressure to the object and the flow rate of the paint, (c) any one or more of the baking temperature and baking time in the baking process, and (d) the temperature, humidity and paint liquid temperature of the production line and any one or more of (a) to (d).
- the relational data is preferably updated constantly, timely, or periodically in order to enable more accurate prediction.
- processing such as processing such as data normalization/standardization and generation of new data, and data filtering processing for removing inappropriate data is performed on the prepared relational data. For example, regarding the difference between the actual measurement value and the predicted value of paint film quality, if the standard deviation is ⁇ , the data that exceeds the average value ⁇ 2 ⁇ will be carefully examined, and data that is suspected of input errors or description errors will be examined. can be deleted.
- Such data cleansing and normalization prevent over-learning in machine learning, which will be described later, and enable more accurate prediction of paint properties.
- performing data cleansing is not essential, and this step can be omitted.
- Data cleansing can be commonly applicable to multiple or all algorithms (e.g., removing data such as outliers deemed inappropriate by any machine learning algorithm), or it can actually be (e.g., removing error-prone data in a particular algorithm).
- the later-described artificial intelligence model creation process can also be created by re-learning the data after this data cleansing.
- a predetermined artificial intelligence model is created that includes the amount of variation in the manufacturing conditions as an input and outputs the amount of variation in the property of the coating film (step S102: artificial intelligence model creation process).
- the artificial intelligence model creation step is performed by machine learning with a computer using the relational data input in the relational data input step (step S101) as teacher data.
- the relational data may be only the fluctuation amount of the manufacturing conditions, or may be the numerical value and the fluctuation amount of the manufacturing conditions. These can use measured values or set values as data.
- the measured values of the properties of the coating film can be used, and the difference between the measured values of the properties of the coating film before and after changing the manufacturing conditions can be calculated and used as the amount of variation.
- the artificial intelligence model is created in this step, but the artificial intelligence model does not necessarily have to be created, and the created artificial intelligence model can be obtained from the outside.
- an artificial intelligence model can be acquired by a communication unit of a computer, or it can be acquired by a person by transfer or the like (artificial intelligence model acquisition step).
- Machine learning algorithms can use any known algorithm, such as decision trees, linear regression, partial least squares regression, Lasso regression, ridge regression, polynomial regression, Gaussian process regression, support vector machines, random forests. , gradient boosting, K nearest neighbors, neural networks, Bayesian estimation, or ensemble learning prediction algorithms of these can be used. It is particularly preferable that machine learning uses a prediction algorithm based on a neural network.
- the neural network may use convolutional neural networks and deep learning with 3 to several hundred layers with dropout.
- Machine learning frameworks such as TensorFlow, Keras, Caffe, PyTorch, Chainer, and Scikit-learn may also be used to create artificial intelligence.
- machine learning can be supervised learning (including reinforcement learning).
- the computer uses the relational data input in the relational data input step (step S102) as learning data (supervised data in this example) for machine learning (supervised There is learning).
- the algorithm is preferably determined by the following method.
- multiple AI models are created using various machine learning algorithms, and the accuracy of each AI model is confirmed by comparing the predicted output value and the answer using data for which the answer is known in advance. do.
- the optimal hyperparameters are determined by, for example, creating a model through cross-validation or using Bayesian optimization. By re-learning all the data with the determined hyperparameters, an artificial intelligence model can be created in which the amount of variation in manufacturing conditions is used as input and the amount of variation in coating properties after adjustment is used as output.
- preprocessing is performed on the prepared relational data.
- predetermined operations such as normalization/standardization of data, taking logarithm, and multiplying by exponentiation (exponentiation processing) are added, and processing such as generation of new data is performed. and data filtering to remove inappropriate data.
- various machine learning algorithms can be used in this embodiment, and the above preprocessing can be commonly applied to a plurality of or all algorithms (for example, any machine data such as outliers that are considered inappropriate even by the learning algorithm), or it can be applied exclusively to the actual machine learning algorithm (e.g., it can cause errors in a particular algorithm). data that are easy to understand).
- zero-point relationship data indicating the relationship that the amount of change in coating properties due to adjustment is 0 when the amount of change in manufacturing conditions is 0 (addition of zero points).
- the data used for machine learning is, for example, the data of records of adjustments made by line managers, it will not include the data stating that "manufacturing conditions are not changed", and the results obtained by machine learning will not change the manufacturing conditions. If not, there is a concern that the point where the amount of change in the properties of the coating film is 0 will not be passed through, and the accuracy of prediction will be reduced. Therefore, by adding the above zero-point relation data, such a problem can be avoided and the prediction accuracy can be further improved. Here, it can be changed to zero, and a value close to zero may be used.
- the data when recording continuous adjustment work as data, the data may be amplified by adding them together as one adjustment record (combination).
- the coating pressure was increased by 50 Kgf in the first recording, and the film thickness decreased by 1.2 ⁇ m.
- the application pressure is increased by 150 Kgf from the beginning, and the data is added as the data in which the film thickness is lowered by 3.8 ⁇ m.
- the accuracy of prediction can be further improved from both the effect of learning data that fluctuates greatly and the effect of increasing the amount of data.
- Inferred data can also be added between data as preprocessing.
- an approximated straight line or an approximated curve can be obtained from data, and points on the approximated straight line or the approximated curve can be used as estimated data. Thereby, the number of data can be increased to further improve the accuracy of prediction.
- past data may be connected by a line or curve, and an arbitrary number of data may be amplified on the line.
- preprocessing and normalization enable prediction of coating properties with higher accuracy in machine learning, which will be described later.
- pretreatment is not essential, and this step can be omitted.
- a predetermined artificial intelligence model can be created using the preprocessed data in the same manner as described above.
- a computer predicts the amount of change in the properties of the coating film by inputting the amount of change in the manufacturing conditions in the artificial intelligence model (step S103: property change amount prediction step).
- the relationship data input includes not only the amount of variation in the manufacturing conditions but also the numerical value of the manufacturing condition, by inputting the numerical value and the amount of variation of the manufacturing condition into the artificial intelligence model, the properties of the coating film can be obtained. It is preferable to predict the amount of variation.
- step S103 it is preferable to narrow down the range of manufacturing conditions to be input in advance by referring to past data.
- the computer can calculate the amount of variation in the properties of the coating film.
- the predicted value of the amount of variation in the properties of the coating film to be calculated may be a single value, or may be composed of a plurality of candidate groups.
- the calculated predicted value of the coating film property variation consists of a plurality of candidate groups, it is preferable to further include a step of appropriately selecting a single value from among them using a predetermined criterion.
- the predetermined criteria may vary.
- the predicted value of the amount of variation in the properties of the coating film with respect to the amount of variation in the manufacturing conditions can be quickly obtained by computer calculation.
- FIG. 2 is a flow chart of a method for predicting the amount of variation in manufacturing conditions according to one embodiment of the present invention.
- the manufacturing conditions are varied and adjusted to obtain the coated material having the target properties of the coating film of the coated material.
- An embodiment of a method for predicting the optimum amount of variation in manufacturing conditions to achieve a target in manufacturing will be described by way of example.
- the method of predicting the amount of variation in manufacturing conditions according to the present embodiment can be executed, for example, using a system for predicting the amount of variation in manufacturing conditions according to an embodiment of the present invention, which will be described later.
- FIG. 1 is a flow chart of a method for predicting the amount of variation in manufacturing conditions according to one embodiment of the present invention.
- the properties of the coating film are any one or more of color, gloss, viscosity, film thickness, smoothness, visual appearance, flip-flop property, and coating film abnormality. is preferably included.
- colors for example, L * value, a * value, and b * value in L * a * b * color space (JIS Z8781-4 (2013)) can be used.
- the color can be measured using a known color measurement method.
- CM-512m3 commercially available from Konica Minolta Co., Ltd.
- the L * value, a * value, and b * value measured by irradiating the light source at angles of 25°, 45°, and 75° can be measured.
- it can be measured using X-Rite MA68II (manufactured by X-Rite).
- the measurement angle can be adjusted as appropriate according to the purpose or equipment to be used. Any other index can be used. Further, for example, it is reflection spectrum data, and an arbitrary index such as an index in which the reflection spectrum intensity is colored every 5 nm from 380 nm to 780 nm can be used. Gloss is not particularly limited, but gloss can be used as an index.
- Gloss can be measured using a known gloss measurement method, for example, the 60 ° gloss of the coating film formed on the test plate is measured with a specular gloss meter (gloss meter VG 7000 (Nippon Denshoku Industries Co., Ltd. )) in accordance with JIS K 5600-4-7 (specular glossiness). Viscosity can be measured using a known viscosity measurement method, for example, according to JIS K 5600-2-2 (flow cup method). It is preferable to use a wave scan value as an index for smoothness.
- the wave scan values are du (wavelength 0.1 mm or less), Wa (wavelength 0.1 to 0.3 mm), Wb (wavelength 0.3 to 1.0 mm), Wc (wavelength 1.0 to 3.0 mm), Any one of Wd (wavelength 3.0 to 10.0 mm), We (wavelength 10.0 to 30.0 mm), Lw (wavelength 1.2 to 12 mm), and Sw (wavelength 0.3 to 1.2 mm) preferably one or more.
- the wave scan value means that the smaller the value, the less unevenness of the wavelength on the surface, and the better the appearance quality of the coating film.
- the flip-flop property can be measured using a gonio-color difference meter or the like.
- coating film defects include unevenness, blisters, cracks, sagging, pinholes, frames, and the like. These can also be measured by known methods.
- the manufacturing conditions are as follows: (a) one or more of the viscosity of the paint and the heat residue; (b) the roll peripheral speed in the step of applying the paint to the object; Any one or more of the coating pressure to the object and the flow rate of the paint, (c) any one or more of the baking temperature and baking time in the baking process, and (d) any one of the temperature and humidity of the production line One or more, preferably any one or more of (a) to (d). Further details of the above items are the same as those of the embodiment of FIG. 1, and therefore will not be described again.
- step S201 relationship data input step
- step S202 artificial intelligence model creation step
- the amount of change in the properties of the coating film is then predicted by inputting the amount of change in the manufacturing conditions in the artificial intelligence model by a computer (step S203: property change volume prediction process).
- step S203 property change volume prediction process
- the relationship data input includes not only the amount of variation in the manufacturing conditions but also the numerical value of the manufacturing condition
- the numerical value and the amount of variation of the manufacturing condition are input to the artificial intelligence model. By doing so, it is preferable to predict the amount of variation in the properties of the coating film.
- the property variation amount prediction step it is preferable to narrow down the range of manufacturing conditions to be input in advance by referring to past data or the like.
- step S203 in the embodiment of FIG. to predict.
- the variation amount of the optimum manufacturing conditions is determined as the variation amount of the optimum manufacturing conditions for achieving the target (step S204 : production condition fluctuation amount prediction step).
- the coating film property that has the smallest difference from the target property among the predicted variation amounts of the coating film properties The amount of variation in the manufacturing conditions corresponding to the amount of variation in is determined as the amount of variation in the optimum manufacturing conditions for achieving the target.
- the manufacturing condition variation amount prediction method of the present embodiment it is possible to obtain an optimum manufacturing condition variation amount prediction value when the properties of the coating film are to be varied by a predetermined variation amount. Therefore, it is sufficient to adjust the manufacturing conditions by the predicted value. Since this adjustment can be performed regardless of the range of standard values for quality, there is no need to adjust the manufacturing conditions even though there is a deviation between the predicted value and the measured value as described above. The problem of not being able to obtain a coated product having the desired coating film quality as a result of failure to appropriately adjust the manufacturing conditions due to the determination of .
- step S204 it is preferable to input a plurality of fluctuation amounts of the manufacturing conditions, which are obtained by combining numerical values of quantities. According to this, prediction can be performed efficiently.
- step S204 it is also preferable to determine, from among the variations in the manufacturing conditions predicted in the manufacturing condition variation amount prediction step (step S204), those within the allowable range as the variation amounts in the manufacturing conditions for achieving the target. According to this, it is possible to predict the amount of variation in the manufacturing conditions after setting the allowable range in advance. Get results efficiently.
- the above numerical value generation range can be obtained based on a certain range from preset standard values or actual values for manufacturing conditions. This makes it possible to obtain predicted values with small deviations from standard values and actual values.
- the variation amounts of the plurality of manufacturing conditions input in the manufacturing condition variation amount prediction step (step S204) consist of candidates equal to or more than the index of the number of manufacturing conditions.
- the allowable variation in manufacturing conditions for coating pressure is -200 Kgf to +250 Kgf, create a variation in candidate manufacturing conditions in 10 steps at intervals of 50 Kgf, and use the other three variations in manufacturing conditions.
- the difference in color is severe and more finely tuned colors are often required, as per the index of the number of each raw material + 1 of the variation amount of 10 manufacturing conditions (10000 in the case of 4 raw materials), especially coil coating etc.
- the number of variations in manufacturing conditions + 2 (100000 if the variation is 4) is a candidate. This is because prediction corresponding to such a fine tone color can be performed by setting .
- a numerical value for the amount of variation in each manufacturing condition is generated, a candidate for the amount of variation in the manufacturing condition is generated by randomly combining the generated numerical values, and the amount of variation in the manufacturing condition for achieving the target is calculated for the candidate.
- Prediction is preferred. For example, when fine adjustment of coating film properties such as film thickness is required, and when the range of variations in manufacturing conditions is large, there are 1 trillion possible combinations (for example, 10 stages with 12 raw materials, or 100 stages with 6 raw materials). ), which makes it difficult and impractical to obtain the predicted value. Therefore, by using random combinations as described above, the number of combinations can be reduced to 10,000,000, and the calculation time for obtaining the predicted values can be shortened.
- the brute-force combinations may also be, for example, 100,000, 1,000,000, or 100,000,000.
- the numerical value itself may be randomly generated within the numerical value generation range.
- the amount of variation in the manufacturing conditions for achieving the target by performing a predetermined calculation on the amount of variation in the manufacturing conditions obtained using the method.
- a predetermined calculation is performed on the predicted variation amount of manufacturing conditions, for example, multiplying the predicted value by a value such as 70% is exemplified.
- the gap target value it can be set at, for example, 70% of the target ⁇ L*.
- the gap target value is set at 70% of ⁇ L* from the target, three adjustments will result in a difference from the target value of 3 ⁇ 0.9 ⁇ 0.27 ⁇ 0.08. can be 0.1 or less.
- it may be set at 700% of the target ⁇ L* and the resulting predicted value multiplied by 10%.
- the target fluctuation value can be set larger than the difference from the target.
- FIG. 3 is a flow chart of a method for manufacturing an applied article according to one embodiment of the present invention.
- the manufacturing conditions are varied and adjusted to obtain the coated material having the target properties of the coating film of the coated material.
- An embodiment of a manufacturing method is illustrated.
- the properties of the coating film are any of color, gloss, viscosity, film thickness, smoothness, visual appearance, flip-flop properties, and coating film abnormality. It preferably contains one or more.
- L * value, a * value, and b * value in L * a * b * color space can be used.
- the color can be measured using a known color measurement method.
- CM-512m3 commercially available from Konica Minolta Co., Ltd. is used, and when the light receiving part perpendicular to the coating film is 0 °, The L * value, a * value, and b * value measured by irradiating the light source at angles of 25°, 45°, and 75° can be measured.
- it can be measured using X-Rite MA68II (manufactured by X-Rite).
- the measurement angle can be adjusted as appropriate according to the purpose or equipment to be used. Any other index can be used. Further, for example, it is reflection spectrum data, and an arbitrary index such as an index in which the reflection spectrum intensity is colored every 5 nm from 380 nm to 780 nm can be used. Gloss is not particularly limited, but gloss can be used as an index. Gloss can be measured using a known gloss measurement method, for example, the 60 ° gloss of the coating film formed on the test plate is measured with a specular gloss meter (gloss meter VG 7000 (Nippon Denshoku Industries Co., Ltd. )) in accordance with JIS K 5600-4-7 (specular glossiness).
- specular gloss meter gloss meter VG 7000 (Nippon Denshoku Industries Co., Ltd. )
- Viscosity can be measured using a known viscosity measurement method, for example, according to JIS K 5600-2-2 (flow cup method). It is preferable to use a wave scan value as an index for smoothness.
- the wave scan values are du (wavelength 0.1 mm or less), Wa (wavelength 0.1 to 0.3 mm), Wb (wavelength 0.3 to 1.0 mm), Wc (wavelength 1.0 to 3.0 mm), Any one of Wd (wavelength 3.0 to 10.0 mm), We (wavelength 10.0 to 30.0 mm), Lw (wavelength 1.2 to 12 mm), and Sw (wavelength 0.3 to 1.2 mm) preferably one or more.
- the wave scan value means that the smaller the value, the less unevenness of the wavelength on the surface, and the better the appearance quality of the coating film.
- the flip-flop property can be measured using a gonio-color difference meter or the like.
- coating film defects include unevenness, blisters, cracks, sagging, pinholes, frames, and the like. These can also be measured by known methods.
- FIG. 1 As in the embodiment of FIG. 1
- the manufacturing conditions are as follows: (a) one or more of the viscosity of the paint and the heat residue; (b) the roll peripheral speed in the step of applying the paint to the object; Any one or more of the coating pressure to the object and the flow rate of the paint, (c) any one or more of the baking temperature and baking time in the baking process, and (d) any one of the temperature and humidity of the production line One or more, preferably any one or more of (a) to (d). Further details of the above items are the same as those in the embodiment of FIGS. 1 and 2, and therefore will not be described again.
- relationship data indicating the relationship between the amount of variation in the manufacturing conditions and the amount of variation in the property of the coating film is input into the computer (step S301: relationship data input step).
- a predetermined artificial intelligence model is created that includes the amount of variation in the manufacturing conditions as an input and the amount of variation in the property of the coating as an output (step S302: artificial intelligence model creation step).
- the computer predicts the amount of variation in the properties of the coating film by inputting the amount of variation in the manufacturing conditions in the artificial intelligence model (step S303: property variation amount prediction step).
- step S304 amount of variation in manufacturing conditions prediction process.
- the manufacturing conditions are then adjusted by the amount of variation in the manufacturing conditions predicted, and the paint is applied to the object to obtain a coating film (Step S305: Manufacturing after condition adjustment). process).
- Step S305 Manufacturing after condition adjustment
- step S306 property measurement step. Measurement can be performed using any measuring means corresponding to the properties to be measured.
- the amount of variation in the manufacturing conditions corresponding to the amount of variation in the properties required to obtain the desired coating film properties is predicted, and the manufacturing conditions are adjusted based on the prediction results. Since the coating film can be formed again after adjustment, a coated article having desired coating film quality can be produced.
- the property variation amount prediction step when the properties of the coating film measured in the property measurement step deviate from the desired properties by a predetermined threshold value or more, the property variation amount prediction step, the manufacturing condition variation amount prediction step, the manufacturing process after condition adjustment, and the property It is preferred to repeat the measuring step. As a result, even if the prediction at one time is incorrect, it is possible to manufacture a coated product having the desired coating film quality through a plurality of adjustments.
- relationship data indicating the relationship between the amount of variation in the manufacturing conditions adjusted in the manufacturing process after condition adjustment and the amount of variation in the properties of the coating film measured in the property measurement process is input to a computer, and the relationship data is It is preferable to further include a relational data update step of updating the .
- This will allow us to update the relational data and create more data-based artificial intelligence models. Updates can be made as soon as new data become available, or they can be made periodically or in a timely manner. However, since over-learning may be caused, this step is not essential and can be omitted.
- FIG. 4 is a flow chart showing a sub-flow of the method for manufacturing an applied article according to one embodiment of the present invention.
- the coating film properties measured in the property measurement step deviate from the desired properties by a predetermined threshold value or more for a predetermined number of times, it is preferable to perform the following subflow.
- this subflow first, by computer machine learning, another artificial intelligence model is created that includes the amount of variation in the manufacturing conditions as an input and the amount of variation in the properties of the coating as an output ( Step S307: artificial intelligence model creation sub-process).
- step S308 property variation amount prediction sub-step.
- the amount of variation in a plurality of manufacturing conditions selected in advance as candidates is input, and the amount of variation in the properties of a plurality of coating films corresponding to the amount of variation in each manufacturing condition is predicted.
- the amount of variation in another optimum manufacturing condition is determined as the amount of variation in another optimum manufacturing condition for achieving the target (step S309: manufacturing condition fluctuation amount prediction sub-process).
- step S310 manufacturing sub-process after condition adjustment
- step S311 property measurement sub-process
- Each sub-step can be performed in the same manner as the corresponding steps of the present process shown in FIG. 3, except that a different artificial intelligence model is used to obtain a different prediction result.
- this sub-flow even if the accuracy of the artificial intelligence model created first is low, it is possible to manufacture a coated product having a coating film with desired properties.
- the flow can be terminated when the properties of the coating film measured in the property measurement step deviate from the desired properties by less than a predetermined threshold. can. Note that in the present disclosure, it is not always necessary to perform subflows.
- FIG. 5 is a schematic diagram of a CCL (Color Coating Line) as an example.
- the step of preparing a paint (step (a)), the coating step using a roll (step (b)), the baking step using an oven (step (c)), and the winding It consists of a plurality of steps including a removing step (step (d)).
- Each process has process manufacturing conditions.
- the process manufacturing conditions are, for example, one or more of the viscosity and heating residue of the paint in the step (a), and in the step (b), for example, the peripheral speed of the roll, the application pressure to the object, and the paint flow rate, for example, one or more of baking temperature and baking time in step (c), and in step (d), for example, production line temperature, humidity and coating liquid temperature.
- the manufacturing conditions for each step are monitored when the manufacturing conditions are adjusted by the amount of variation in the manufacturing conditions predicted in the manufacturing condition variation amount prediction step in the flow of FIG.
- the measured value or (after adjustment) set value is used, and if it is a baking temperature or baking time, the measured value or (after adjustment ) setpoints can be used to measure temperature and humidity under the environmental conditions of the production line. Then, when a process manufacturing condition adjusted to deviate from the planned amount of variation by a predetermined value or more is observed among the process manufacturing conditions, the following flow is performed. First, in the created artificial intelligence model, after fixing the process manufacturing conditions in which the adjusted process manufacturing conditions were observed to deviate from the planned amount of variation by a predetermined value or more, input the amount of variation in other multiple manufacturing conditions.
- the amount of variation in properties of a plurality of coating films is calculated and predicted (property re-prediction step).
- the amount of variation in the optimum manufacturing conditions is re-determined as the amount of variation in the optimum manufacturing conditions to achieve the target (Re-prediction of variation in manufacturing conditions process).
- the amount of variation in the manufacturing conditions in the process subsequent to the current process is adjusted based on the result of the process of re-predicting the amount of variation in manufacturing conditions.
- the amount of variation in the optimum manufacturing conditions in step (c) and/or step (d) is re-determined as the amount of variation in optimum manufacturing conditions to achieve the target.
- the amount of variation in the manufacturing conditions in the optimum step (c) and / or step (d) for obtaining a coating film with desired properties can be re-predicted.
- the manufacturing conditions in step (c) and/or step (d) are adjusted. Since it is necessary to evaluate the properties of the paint film after a series of processes have been completed and the paint has dried, it is necessary to adjust again in the event that the adjustment has deviated.
- FIG. 6 is a block diagram of a prediction system for the amount of change in properties of a coating film according to one embodiment of the present invention.
- This system is a system that predicts the amount of variation in the properties of a paint film obtained by applying paint to an object.
- a coating film property variation prediction system 10 of the present embodiment includes a computer 11 .
- the computer 11 has a machine learning function.
- the computer 11 also includes an artificial intelligence model creation unit 12 and a property prediction unit 13 .
- the artificial intelligence model creation unit 12 has a function of machine learning, and creates a predetermined artificial intelligence model by machine learning, which includes the amount of variation in the manufacturing conditions as an input and outputs the amount of variation in the property of the coating film. is.
- the property prediction unit 13 calculates and predicts the amount of change in the property of the coating film by inputting the amount of change in the manufacturing conditions in the artificial intelligence model.
- the artificial intelligence model creation unit 12 and the property prediction unit 13 can be processors.
- the computer 11 receives the above-described relational data.
- the computer 11 has a storage unit 14 (memory) for storing relational data and a communication unit 15 for transmitting and receiving the relational data.
- the communication unit 15 can transmit and receive not only relational data but also other data.
- the artificial intelligence model creation unit 12 preferably has a function of performing machine learning using the input relational data as teacher data.
- the computer 11 further includes a functional unit that performs data cleansing on relational data.
- the prediction system 10 include a display section (display) for displaying the prediction result. The details of the manufacturing conditions, properties of the coating film, machine learning, etc. are the same as in the embodiment of the method shown in FIG.
- the predicted value of the amount of variation in the properties of the coating film with respect to the amount of variation in the manufacturing conditions can be quickly obtained by computer calculation.
- FIG. 7 is a block diagram of a prediction system for the amount of variation in manufacturing conditions according to one embodiment of the present invention.
- this system adjusts the manufacturing conditions by varying the manufacturing conditions.
- This is a system that predicts the amount of variation in optimal manufacturing conditions to achieve a target.
- the manufacturing condition prediction system 20 of the present embodiment is composed of a computer 21 .
- the computer 21 has a function of machine learning.
- the computer 21 also has an artificial intelligence model creation unit 22 , a property prediction unit 23 , a storage unit 24 and a communication unit 25 . Since these are the same as those described for the artificial intelligence model creation unit 12, property prediction unit 13, storage unit 14, and communication unit 15 in the embodiment shown in FIG. 6, detailed description thereof will be omitted.
- the property prediction unit 23 is configured to receive a plurality of variations in manufacturing conditions and predict variations in the properties of a plurality of coating films corresponding to the respective variations in manufacturing conditions.
- the computer 21 further includes a manufacturing condition prediction unit 26 that determines the amount of change in the optimum manufacturing conditions as the optimum manufacturing conditions for achieving the target based on the predicted amount of change in the properties of the plurality of coating films.
- the manufacturing condition prediction unit 26 can be a processor.
- the manufacturing condition prediction unit 26 corresponds to the amount of change in the properties of the coating film that has the smallest difference from the amount of change in the properties of the target coating film among the predicted amounts of change in the properties of the coating film.
- the amount of variation in manufacturing conditions can be determined as the amount of variation in optimum manufacturing conditions for achieving a target.
- the computer 21 further has a functional unit that performs data cleansing on the relational data.
- the prediction system 20 include a display unit 27 (display) for displaying the prediction result.
- display unit 27 display
- this prediction system 20 it is possible to obtain the predicted value of the optimal manufacturing condition variation amount when it is desired to vary the properties of the coating film by a predetermined variation amount. Therefore, it is sufficient to adjust the manufacturing conditions by the predicted value. Since this adjustment can be performed regardless of the range of standard values for quality, there is no need to adjust the manufacturing conditions even though there is a deviation between the predicted value and the measured value as described above. The problem of not being able to obtain a coated product having the desired coating film quality as a result of failure to appropriately adjust the manufacturing conditions due to the determination of .
- ⁇ Preparation example of white primary color paint 20 parts by mass of an acrylic resin as a resin and 35 parts by mass of isophorone as an organic solvent are added, and the resin is uniformly dissolved using a disper, followed by mixing 46 parts by mass of titanium oxide as a white pigment, followed by a sand mill (dispersion medium: glass beads ) was used to disperse the pigment coarse particles until the maximum particle size was 10 ⁇ m or less to prepare a white primary color paint.
- ⁇ Gloss modifier> As the gloss modifier, commercially available silica (silicon dioxide) used as a matting agent was used. ⁇ Viscosity modifier> Isophorone used as a solvent was used as a viscosity modifier.
- ⁇ Preparation example of coating composition 1> After adding 5 parts by mass of acrylic resin, 25 parts by mass of fluororesin, 35 parts by mass of isophorone and 35 parts by mass of cyclohexanone and uniformly dissolving the resin using a disper, 25 parts by mass of white primary color paint, 2 parts by mass of black primary color paint Parts, 69 parts by mass of yellow primary color paint, 4 parts by mass of red primary color paint and 2 parts by mass of gloss modifier 1 are added and uniformly mixed using a disper to obtain coating composition 1 (solid content concentration: 47% by mass). prepared.
- Coating compositions 2 to 4 were prepared in the same manner as in the preparation example of coating composition 1 except that the types and amounts of each material were changed as shown in Table 2 below. Table 2 shows the formulation of each coating composition.
- Acrylic resin Paraloid B44 (manufactured by Rohm & Haas), solid content concentration: 100% by mass
- Fluorine resin KYNAR500 (manufactured by ARKEMA), solid content concentration: 100% by mass
- White pigment TI-PURE R-706 (titanium oxide, manufactured by DuPont)
- Black pigment 1 Mitsubishi carbon black MA-100 (carbon black, manufactured by Mitsubishi Chemical Corporation)
- Black pigment 2 SUNBLACK X15 (carbon black, manufactured by Shiraishi Calcium Co., Ltd.)
- TAROX synthetic iron oxide LL-XLO yellow iron oxide, manufactured by Titan Kogyo Co., Ltd.
- Red pigment TODA COLOR 140ED (iron oxide, manufactured by Toda Kogyo Co., Ltd.)
- Gloss modifier 1 GASIL HP395 (synthetic silica, manufactured by INEOS SILICAS) ⁇
- Coating film production method The following steps are carried out in a single coating line.
- Fine Tough G Primer (Primer 1) is applied as an undercoat to the material (zinc-aluminum alloy plated steel sheet) with a roll coater (standard film thickness: 5 ⁇ m), and then baked for 60 seconds at a maximum temperature of 210°C. was performed to form an undercoat film.
- the coating composition 1 is applied with a roll coater (standard film thickness: 15 ⁇ m), baked at a maximum material temperature of 250 ° C. for 60 seconds, and immediately cooled to remove the coating film of the coating composition before adjustment. formed.
- the film thickness of the undercoat paint or paint composition was calculated from the weight before and after removing the paint film from the cut-out coated plate by sandblasting and the specific gravity of the paint film.
- Color Difference Measurement Method Lab Scan XE manufactured by HunterLab Co. was used to measure the color of the coating film, and the difference from the intended color of the coated plate was defined as the color difference.
- Artificial intelligence creation example ⁇ Data input for creating an artificial intelligence model
- the application roll (AP roll) speed fluctuation amount, coating pressure (backup roll and The application pressure of the AP roll) variation amount and primary color correction addition amount data were input into a computer.
- the film thickness, color, and gloss the film thickness, color, gloss, and paint viscosity before adjusting the coating parameters and the film thickness variation, color variation, and gloss variation after adjusting the coating parameters were used.
- Preprocessing such as data cleansing, zero point introduction, combination, exponentiation processing, and normalization was performed in advance.
- the film thickness, color, gloss, paint viscosity before adjustment, AP roll speed fluctuation amount, coating pressure fluctuation amount, primary color paint addition amount are used as explanatory variables in the computer, and the film thickness fluctuation amount, color fluctuation amount, after adjusting the coating parameters,
- a predetermined artificial intelligence model was created with the amount of gloss variation as the objective variable. Using gradient boosting, random forest, support vector machine, and neural network cross-validation as machine learning algorithms, the optimal hyperparameters are determined, and all data are retrained with the determined hyperparameters to determine the manufacturing conditions.
- a predetermined artificial intelligence model was created in which the data of the amount of variation in was used as an explanatory variable, and the amount of variation in the properties of the coating film after adjustment was used as an objective variable.
- ⁇ As for the delta gloss value the difference from the standard plate shall be 0.5 or less when the gloss value is 10 or less, 1 or less when the gloss value is 10 to 20, and 3 or less when the gloss value is 30 or more. was passed.
- Acceptance criteria for the number of adjustments can be automated, and when the difference between the film thickness of the coating film properties before adjustment and the target film thickness is within 5 ⁇ m and within ⁇ E2, the number of times of adjustment of the coating film properties is 3. When the acceptance criteria were reached within the time, it was regarded as a pass.
- Example 2> The difference from the target film thickness is about 2 ⁇ m, and the coating composition 1 having a coating film property of about ⁇ E 2.5 and the manufacturing conditions, the amount of viscosity variation, the amount of coating pressure variation, the amount of primary color paint added, A permissible range is set for the amount of variation in the baking temperature and the amount of variation in the temperature of the production line, and after generating 10 million different numbers within that range, randomly combine them to determine the film thickness before adjustment, color, gloss, Along with the paint viscosity, it is input to five artificial intelligence models that have been trained by a neural network, and after adjusting the coating parameters, the film thickness variation, L* value variation, a* value variation, b* value variation, and gloss We obtained 10 million types of candidate data for variation.
- the coating supervisor adjusted various coating conditions and adjusted the film thickness, color and gloss until the film passed the test.
- the average number of adjustments performed for coating compositions 2 to 4 was 3.5.
- Examples 1 and 2 were passed after an average of 1.5 times and 2.0 times of adjustment, satisfying the above acceptance criteria. On the other hand, in Comparative Example 1, even an experienced painting supervisor had to make three or more adjustments. In Comparative Example 2, when the film thickness was adjusted using the artificial intelligence model according to the comparative artificial intelligence creation example, the coating parameters could not be adjusted. Even if there is a discrepancy between the predicted value and the measured value, if the predicted value is within the range of the target film thickness, it is judged that there is no need for adjustment, and the coating parameter is changed. This is because the phenomenon often occurred that when the absolute value prediction was approaching the target, the property rather moved away from the target property due to the adjustment.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Application Of Or Painting With Fluid Materials (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
(1)対象物に塗料を塗布して得られた塗膜の性状の変動量を予測する方法であって、
製造条件の変動量を入力として含み、前記塗膜の性状の変動量を出力とする、所定の人工知能モデルを取得する、人工知能モデル取得工程と、
コンピュータにより、前記人工知能モデルにおいて、前記製造条件の変動量を入力することによって、前記塗膜の性状の変動量を予測する、性状変動量予測工程と、を含むことを特徴とする、塗膜の性状の変動量の予測方法。
製造条件の変動量を入力として含み、前記塗膜の性状の変動量を出力とする、所定の人工知能モデルを取得する、人工知能モデル取得工程と、
コンピュータにより、前記人工知能モデルにおいて、前記製造条件の変動量を入力することによって、前記塗膜の性状の変動量を予測する、性状変動量予測工程と、を含み、
前記性状変動量予測工程では、複数の前記製造条件の変動量を入力し、各々の前記製造条件の変動量に対応する複数の前記塗膜の性状の変動量を予測し、
予測した前記複数の前記塗膜の性状の変動量に基づいて、最適な前記製造条件の変動量を、前記目標を達成するために最適な前記製造条件の変動量として決定する、製造条件変動量予測工程をさらに含むことを特徴とする、製造条件の変動量の予測方法。
前記人工知能モデル取得工程は、前記コンピュータにより、前記関係データ入力工程において入力された前記関係データを教師データとして機械学習することにより行われる、上記(1)又は(2)に記載の方法。
(a)塗料の粘度及び加熱残分のいずれか1つ以上、
(b)前記対象物に塗料を塗布する工程における、ロール周速、前記対象物への塗着圧、及び塗料の流量のいずれか1つ以上、
(c)焼き付け工程における焼き付け温度及び焼き付け時間のいずれか1つ以上、及び
(d)製造ラインの温度、湿度及び塗料液温のいずれか1つ以上、
の(a)~(d)うちのいずれか1つ以上を含む、請求項1~4のいずれか一項に記載の方法。
得られた前記塗布物の前記塗膜の前記性状を測定する、性状測定工程と、をさらに含み、
前記性状測定工程において測定された前記塗膜の前記性状が、所望の性状から所定の閾値以上ずれた場合に、前記性状変動量予測工程、前記製造条件変動量予測工程、前記条件調整後製造工程、及び前記性状測定工程を繰り返す、塗布物の製造方法。
前記製造条件の変動量を入力として含み、前記塗膜の性状の変動量を出力とする、別の人工知能モデルを取得する、人工知能モデル取得サブ工程と、
前記コンピュータにより、前記別の人工知能モデルにおいて、前記製造条件の変動量を入力することによって、前記塗膜の性状の変動量を予測する、性状変動量予測サブ工程と、を含み、
前記性状変動量予測サブ工程では、予め候補として選定した複数の前記製造条件の変動量を入力し、各々の前記製造条件の変動量に対応する複数の前記塗膜の性状の変動量を予測し、
予測した前記複数の前記塗膜の性状の変動量に基づいて、別の最適な前記製造条件の変動量を、前記目標を達成するために最適な別の前記製造条件の変動量として決定する、製造条件変動量予測サブ工程をさらに含み、
予測した前記別の製造条件の変動量の分だけ前記製造条件を調整して、前記対象物に前記塗料を塗布して前記塗布物を得る、条件調整後製造サブ工程と、
得られた前記塗布物の前記塗膜の前記性状を測定する、性状測定サブ工程と、をさらに含み、
前記性状測定工程において測定された前記塗膜の前記性状が、所望の性状から所定の閾値以上ずれた場合に、前記人工知能モデル取得サブ工程、前記性状変動量予測サブ工程、前記製造条件変動量予測サブ工程、前記条件調整後製造サブ工程、及び前記性状測定サブ工程を繰り返す、上記(10)に記載の塗布物の製造方法。
製造条件の変動量を入力として含み、前記塗膜の性状の変動量を出力とする、所定の人工知能モデルを取得する、人工知能モデル取得部と、
前記人工知能モデルにおいて、前記製造条件の変動量を入力することによって、前記塗膜の性状の変動量を算出して予測する、性状予測部と、を備えたコンピュータを備えていることを特徴とする、システム。
前記製造条件の変動量を入力として含み、前記塗膜の性状の変動量を出力とする、所定の人工知能モデルを取得する、人工知能モデル取得部と、
前記人工知能モデルにおいて、複数の前記製造条件の変動量を入力することによって、複数の前記塗膜の性状の変動量を算出して予測する、性状予測部と、
予測した前記複数の前記塗膜の性状の変動量に基づいて、最適な前記製造条件の変動量を、前記目標を達成するために最適な前記製造条件の変動量として決定する、製造条件変動量予測部と、を備えたコンピュータを備えていることを特徴とする、システム。
上記(2)に記載の方法により予測した製造条件の変動量の分だけ前記製造条件を調整した際の、各前記製造条件をモニタリングし、
各前記製造条件のうち、前記変動量から所定の値以上ずれて調整された工程製造条件が観測された場合に、
前記人工知能モデルにおいて、モニタリングされた前記工程製造条件を固定した上で、他の複数の前記製造条件の変動量を入力することによって、複数の前記塗膜の性状の変動量を算出して予測する、性状再予測工程と、
予測した前記複数の前記塗膜の性状の変動量に基づいて、最適な前記製造条件の変動量を、前記目標を達成するために最適な前記製造条件の変動量として再決定する、製造条件変動量再予測工程と、をさらに含み、
現工程より後の工程における前記製造条件の変動量を、前記製造条件変動量再予測工程の結果に基づいて調整し直す、上記(11)に記載の製造方法。
図1は、本発明の一実施形態にかかる塗膜の性状の変動量の予測方法のフローチャートである。以下、図1を参照して、対象物に塗料を塗布して得られた塗膜の性状の変動量を予測する方法の一実施形態を例示説明する。なお、本実施形態の塗膜の性状の変動量の予測方法は、一例としては、後述の本発明の一実施形態にかかる塗膜の性状の変動量の予測システムを用いて実行することができる。ここで、予測する塗膜の性状は、少なくとも塗料の色彩を含むことが好ましく、また、色彩に加え、塗料の光沢及び膜厚をさらに含むことも好ましい。なお、予測する塗膜性状は、色彩、光沢、膜厚以外には、例えば、隠ぺい率、フリップフロップ性、平滑性、目視外観、電気抵抗、接触角、汚染性、日射反射率、紫外線透過率、耐候性、粘弾性、塗膜異常、のいずれか1つ以上を含むことが好ましい。特に、膜厚を変化させるため塗着圧を調整すると色彩も光沢も変化し、光沢を調整しようと光沢調整剤を入れると色彩も変化するように互いに影響を及ぼすので、色彩、光沢、膜厚を同時に予測して調整することが好ましい。色彩は、例えばL*a*b*色空間におけるL*値、a*値、b*値(JIS Z8781-4(2013年))を用いることができる。色彩は、既知の色彩測定方法を用いて測定することができ、一例として、コニカミノルタ株式会社から市販のCM-512m3を用いて、塗膜に垂直にある受光部を0°とした場合に、25°、45°、75°となる角度から光源を照射して測定されるL*値、a*値、b*値を測定することができる。あるいは、X-Rite MA68II(エックスライト社製)を用いて測定することができる。測定角度は、目的又は使用する機器に応じて適宜調整することができる。その他任意の指標を用いることができる。さらに例えば、反射スペクトルデータであり、380nm~780nmの5nm毎の反射スペクトル強度を色彩とした指標等、任意の指標を用いることもできる。光沢は、特には限定されないが、グロスを指標として用いることができる。グロスは、既知の光沢測定方法を用いて測定することができ、一例として、試験板に形成した塗膜の60°光沢度を、鏡面光沢度計(光沢計VG 7000(日本電色工業社製))を用い、JIS K 5600-4-7(鏡面光沢度)に準拠して測定することができる。粘度は、既知の粘度測定方法を用いることができ、一例としては、JIS K 5600-2-2(フローカップ法)に準拠して測定することができる。平滑性は、ウェーブスキャン値を指標として用いることが好ましい。ウェーブスキャン値は、du(波長0.1mm以下)、Wa(波長0.1~0.3mm)、Wb(波長0.3~1.0mm)、Wc(波長1.0~3.0mm)、Wd(波長3.0~10.0mm)、We(波長10.0~30.0mm)、Lw(波長1.2~12mm)、及びSw(波長0.3~1.2mm)のいずれか1つ以上であることが好ましい。なお、ウェーブスキャン値は、値が小さいほど表面における当該波長の凹凸が少なく、塗膜の外観品質が良いことを意味する。フリップフロップ性は、変角色差計等を用いて計測することができる。塗膜異常は、具体的には、ムラ、フクレ、割れ、タレ、ピンホール、額縁等である。これらについても、それぞれ既知の手法で測定することができる。
また、関係データ入力工程(ステップS101)後に、関係データに対してデータクレンジングを行う工程をさらに含むことが好ましい。ここでは、準備した上記関係データに対して例えば、データの正規化/標準化や新たなデータの生成といった加工処理や、不適切なデータを取り除くデータフィルタリング処理などの処理を行う。例えば、塗膜品質の実測値と予測値との差異について、標準偏差をσとしたときの平均値±2σを超えたデータに対して精査をし、入力ミスや記載ミスなどが疑われるデータを削除することができる。このようなデータクレンジングや正規化により、後述の機械学習における過学習を防止して、より精度の高い塗料性状の予測を可能にする。ただし、本開示において、データクレンジングを行うことは必須ではなく、この工程を省略することもできる。データクレンジングは、複数ないし全てのアルゴリズムに共通に適用できるものとすることができ(例えばいずれの機械学習アルゴリズムでも不適切と判断されると考えられる異常値のようなデータを取り除く)、あるいは、実際に用いる機械学習アルゴリズムにもっぱら適用できるものとすることもできる(例えば特定のアルゴリズムにおいてエラーを生じさせやすいデータを取り除く)。後述の人工知能モデル作成工程は、このデータクレンジング後のデータを再学習させて作成することもできる。
機械学習は、ニューラルネットワークによる予測アルゴリズムを用いていることが特に好ましい。ニューラルネットワークには、畳み込みニューラルネットワーク及びドロップアウトを行いながら3~数百の層を形成するディープラーニングを使用してもよい。
また、人工知能を作成するにあたって、TensorFlow、Keras、Caffe、PyTorch、Chainer、Scikit-learnなどの機械学習フレームワークを使用してもよい。
上述のように、本実施形態では、様々な機械学習のアルゴリズムを用いることができるが、上記の前処理は、複数ないし全てのアルゴリズムに共通に適用できるものとすることができ(例えばいずれの機械学習アルゴリズムでも不適切と判断されると考えられる異常値のようなデータを取り除く)、あるいは、実際に用いる機械学習アルゴリズムにもっぱら適用できるものとすることもできる(例えば特定のアルゴリズムにおいてエラーを生じさせやすいデータを取り除く)。
上記と同様の方法で前処理後のデータを用いて所定の人工知能モデルを作成することができる。
図2は、本発明の一実施形態にかかる製造条件の変動量の予測方法のフローチャートである。以下、図2を参照して、対象物に塗料を塗布して塗布物を製造するに際し、製造条件を変動させて調整することにより、塗布物の塗膜の目標となる性状を有する塗布物を製造する場合に、目標を達成するために最適な製造条件の変動量を予測する方法の一実施形態を例示説明する。なお、本実施形態の製造条件の変動量の予測方法は、一例としては、後述の本発明の一実施形態にかかる製造条件の変動量の予測システムを用いて実行することができる。ここで、図1の実施形態と同様に、塗膜の性状は、塗膜の色彩、光沢、粘度、膜厚、平滑性、目視外観、フリップフロップ性、及び塗膜異常のいずれか1つ以上を含むことが好ましい。色彩は、例えばL*a*b*色空間におけるL*値、a*値、b*値(JIS Z8781-4(2013年))を用いることができる。色彩は、既知の色彩測定方法を用いて測定することができ、一例として、コニカミノルタ株式会社から市販のCM-512m3を用いて、塗膜に垂直にある受光部を0°とした場合に、25°、45°、75°となる角度から光源を照射して測定されるL*値、a*値、b*値を測定することができる。あるいは、X-Rite MA68II(エックスライト社製)を用いて測定することができる。測定角度は、目的又は使用する機器に応じて適宜調整することができる。その他任意の指標を用いることができる。さらに例えば、反射スペクトルデータであり、380nm~780nmの5nm毎の反射スペクトル強度を色彩とした指標等、任意の指標を用いることもできる。光沢は、特には限定されないが、グロスを指標として用いることができる。グロスは、既知の光沢測定方法を用いて測定することができ、一例として、試験板に形成した塗膜の60°光沢度を、鏡面光沢度計(光沢計VG 7000(日本電色工業社製))を用い、JIS K 5600-4-7(鏡面光沢度)に準拠して測定することができる。粘度は、既知の粘度測定方法を用いることができ、一例としては、JIS K 5600-2-2(フローカップ法)に準拠して測定することができる。平滑性は、ウェーブスキャン値を指標として用いることが好ましい。ウェーブスキャン値は、du(波長0.1mm以下)、Wa(波長0.1~0.3mm)、Wb(波長0.3~1.0mm)、Wc(波長1.0~3.0mm)、Wd(波長3.0~10.0mm)、We(波長10.0~30.0mm)、Lw(波長1.2~12mm)、及びSw(波長0.3~1.2mm)のいずれか1つ以上であることが好ましい。なお、ウェーブスキャン値は、値が小さいほど表面における当該波長の凹凸が少なく、塗膜の外観品質が良いことを意味する。フリップフロップ性は、変角色差計等を用いて計測することができる。塗膜異常は、具体的には、ムラ、フクレ、割れ、タレ、ピンホール、額縁等である。これらについても、それぞれ既知の手法で測定することができる。また、製造条件についても、図1の実施形態と同様に、(a)塗料の粘度及び加熱残分のいずれか1つ以上、(b)対象物に塗料を塗布する工程における、ロール周速、対象物への塗着圧、及び塗料の流量のいずれか1つ以上、(c)焼き付け工程における焼き付け温度及び焼き付け時間のいずれか1つ以上、及び(d)製造ラインの温度及び湿度のいずれか1つ以上、の(a)~(d)うちのいずれか1つ以上を含むものとすることが好ましい。
以上の事項のさらなる詳細については、図1の実施形態と同様であるため、再度の説明を省略する。
そこで、製造条件に対し、それぞれ変動量の数値生成範囲を設定する工程をさらに含み、製造条件変動量予測工程(ステップS204)では、設定された数値生成範囲内で生成された各製造条件の変動量の数値を組み合わせてなる複数の前記製造条件の変動量を入力することが好ましい。これによれば、予測を効率的に行うことができる。
上記態様において、数値そのものを数値生成範囲内でランダムに生成してもよい
予測した製造条件の変動量に所定の演算を行う場合としては、例えば、予測値に70%等の値を乗じることが例示される。
ギャップ目標値を用いる場合の第1の例としては、例えば目標のΔL*の70%の点に設定することができる。例えば、ΔL*=3、ギャップ目標値を目標とのΔL*の70%の点に設定すると、3回の調整により、3→0.9→0.27→0.08と目標値からの差を0.1以下とすることができる。ギャップ目標値を用いる場合の第2の例としては、目標のΔL*の700%の点に設定して、得られた予測値に10%を乗じてもよい。
図3は、本発明の一実施形態にかかる塗布物の製造方法のフローチャートである。以下、図3を参照して、対象物に塗料を塗布して塗布物を製造するに際し、製造条件を変動させて調整することにより、塗布物の塗膜の目標となる性状を有する塗布物を製造する方法の一実施形態を例示説明する。ここで、図1、図2の実施形態と同様に、塗膜の性状は、塗膜の色彩、光沢、粘度、膜厚、平滑性、目視外観、フリップフロップ性、及び塗膜異常のいずれか1つ以上を含むことが好ましい。色彩は、例えばL*a*b*色空間におけるL*値、a*値、b*値(JIS Z8781-4(2013年))を用いることができる。色彩は、既知の色彩測定方法を用いて測定することができ、一例として、コニカミノルタ株式会社から市販のCM-512m3を用いて、塗膜に垂直にある受光部を0°とした場合に、25°、45°、75°となる角度から光源を照射して測定されるL*値、a*値、b*値を測定することができる。あるいは、X-Rite MA68II(エックスライト社製)を用いて測定することができる。測定角度は、目的又は使用する機器に応じて適宜調整することができる。その他任意の指標を用いることができる。さらに例えば、反射スペクトルデータであり、380nm~780nmの5nm毎の反射スペクトル強度を色彩とした指標等、任意の指標を用いることもできる。光沢は、特には限定されないが、グロスを指標として用いることができる。グロスは、既知の光沢測定方法を用いて測定することができ、一例として、試験板に形成した塗膜の60°光沢度を、鏡面光沢度計(光沢計VG 7000(日本電色工業社製))を用い、JIS K 5600-4-7(鏡面光沢度)に準拠して測定することができる。粘度は、既知の粘度測定方法を用いることができ、一例としては、JIS K 5600-2-2(フローカップ法)に準拠して測定することができる。平滑性は、ウェーブスキャン値を指標として用いることが好ましい。ウェーブスキャン値は、du(波長0.1mm以下)、Wa(波長0.1~0.3mm)、Wb(波長0.3~1.0mm)、Wc(波長1.0~3.0mm)、Wd(波長3.0~10.0mm)、We(波長10.0~30.0mm)、Lw(波長1.2~12mm)、及びSw(波長0.3~1.2mm)のいずれか1つ以上であることが好ましい。なお、ウェーブスキャン値は、値が小さいほど表面における当該波長の凹凸が少なく、塗膜の外観品質が良いことを意味する。フリップフロップ性は、変角色差計等を用いて計測することができる。塗膜異常は、具体的には、ムラ、フクレ、割れ、タレ、ピンホール、額縁等である。これらについても、それぞれ既知の手法で測定することができる。また、製造条件についても、図1の実施形態と同様に、(a)塗料の粘度及び加熱残分のいずれか1つ以上、(b)対象物に塗料を塗布する工程における、ロール周速、対象物への塗着圧、及び塗料の流量のいずれか1つ以上、(c)焼き付け工程における焼き付け温度及び焼き付け時間のいずれか1つ以上、及び(d)製造ラインの温度及び湿度のいずれか1つ以上、の(a)~(d)うちのいずれか1つ以上を含むものとすることが好ましい。
以上の事項のさらなる詳細については、図1、図2の実施形態と同様であるため、再度の説明を省略する。
本実施形態では、性状測定工程において測定された塗膜の性状が、所望の性状から所定の閾値以上ずれることが、所定の回数繰り返された場合には、以下のサブフローを行うことが好ましい。
図4に示すように、本サブフローでは、まず、コンピュータによる機械学習によって、製造条件の変動量を入力として含み、塗膜の性状の変動量を出力とする、別の人工知能モデルを作成する(ステップS307:人工知能モデル作成サブ工程)。次いで、コンピュータにより、別の人工知能モデルにおいて、製造条件の変動量を入力することによって、前記塗膜の性状の変動量を予測する(ステップS308:性状変動量予測サブ工程)。性状変動量予測サブ工程では、予め候補として選定した複数の製造条件の変動量を入力し、各々の製造条件の変動量に対応する複数の塗膜の性状の変動量を予測する。次いで、予測した複数の塗膜の性状の変動量に基づいて、別の最適な製造条件の変動量を、目標を達成するために最適な別の製造条件の変動量として決定する(ステップS309:製造条件変動量予測サブ工程)。次いで、予測した別の製造条件の変動量の分だけ製造条件を調整して、対象物に塗料を塗布して塗布物を得る(ステップS310:条件調整後製造サブ工程)。次いで、得られた塗布物の塗膜の性状を測定する(ステップS311:性状測定サブ工程)。そして、図4に示すように、性状測定工程において測定された塗膜の性状が、所望の性状から所定の閾値以上ずれた場合に、人工知能モデル作成サブ工程、性状変動量予測サブ工程、製造条件変動量予測サブ工程、条件調整後製造サブ工程、及び性状測定サブ工程を繰り返す。
このサブフローにより、仮に最初に作成した人工知能モデルの精度が低いような場合であっても、所望の性状を有する塗膜を有する塗布物を製造し得る。
図3に示す本工程又は図4に示すサブ工程において、性状測定工程において測定された塗膜の性状が、所望の性状から所定の閾値未満のずれとなった時点で本フローを終了することができる。なお、本開示において、必ずしもサブフローを行う必要はない。
図5に模式的に示す例では、塗料を準備する工程(工程(ア))、ロールを用いた塗装工程(工程(イ))、オーブンを用いた焼き付け工程(工程(ウ))、及び巻き取り工程(工程(エ))からなる複数の工程からなる。各々の工程は、工程製造条件を有する。工程製造条件は、工程(ア)では、例えば塗料の粘度及び加熱残分のいずれか1つ以上であり、工程(イ)では、例えばロール周速、対象物への塗着圧、及び塗料の流量のいずれか1つ以上であり、工程(ウ)では、例えば焼き付け温度及び焼き付け時間のいずれか1つ以上であり、また、工程(エ)では、例えば、製造ラインの温度、湿度及び塗料液温のいずれか1つ以上である。
ここで、図3のフローにおける製造条件変動量予測工程において予測した製造条件の変動量の分だけ製造条件を調整した際の、各工程製造条件をモニタリングする。例えば、塗料であれば、上記各性状や特性を測定し、製造条件であれば、測定値又は(調整後の)設定値を用い、焼き付け温度や焼き付け時間であれば、測定値又は(調整後の)設定値を用い、製造ラインの環境条件であれば、温度及び湿度を測定することができる。
そして、各工程製造条件のうち、予定した変動量から所定の値以上ずれて調整された工程製造条件が観測された場合には、以下のフローを行う。
まず、作成した人工知能モデルにおいて、予定した変動量から所定の値以上ずれて調整された工程製造条件が観測された工程製造条件を固定した上で、他の複数の製造条件の変動量を入力することによって、複数の塗膜の性状の変動量を算出して予測する(性状再予測工程)。
次いで、予測した複数の塗膜の性状の変動量に基づいて、最適な製造条件の変動量を、目標を達成するために最適な製造条件の変動量として再決定する(製造条件変動量再予測工程)。
現工程より後の工程における製造条件の変動量を、製造条件変動量再予測工程の結果に基づいて調整する。
以上のことを、図5の例で説明すると、一例としては、現在ロールを用いた塗装工程(工程(イ))を行っている場合に、工程(ア)での塗料の性状や特性の調整がずれたことが観測された場合に、そのことを前提とした上で、現工程より後の工程である工程(ウ)及び/又は工程(エ)の条件の最適な変動量を再予測することで、工程(ア)での調整のずれも考慮した適切な調整ができる場合がある。
この例では、作成した人工知能モデルを用いて、既に行っており調整がずれたことが観測された工程(ア)における塗料の性状や特性の値を固定値として代入することを前提とした上で、工程(イ)~工程(エ)の製造条件の変動量を複数入力して、複数の塗膜の性状の変動量を算出して予測し、予測した複数の塗膜の性状の変動量に基づいて、工程(ウ)及び/又は工程(エ)での最適な製造条件の変動量を、目標を達成するために最適な製造条件の変動量として再決定する。これにより、工程(ア)での調整のずれを前提とした上で、所望の性状の塗膜を得るのに最適な工程(ウ)及び/又は工程(エ)での製造条件の変動量を再予測することができる。そして、再予測した結果に基づいて、工程(ウ)及び/又は工程(エ)での製造条件を調整する。
塗膜は、一連の工程の終了及び塗料の乾燥を待ってから、その性状の評価を行う必要があるため、調整がずれてしまった場合に、再度の調整を行う必要があると判断するまでに、時間を要してしまい、その間、品質の基準を満たさない塗膜を有する塗布物が製造されてしまうおそれがある。
これに対し、このような手法によれば、一連の工程の終了を待つことなく、途中の工程で製造条件の再予測を行うことができるため、品質の基準を満たさない塗膜を有する塗布物が製造されてしまうのを極力避けることができる。
図6は、本発明の一実施形態にかかる塗膜の性状の変動量の予測システムのブロック図である。本システムは、対象物に塗料を塗布して得られた塗膜の性状の変動量を予測するシステムである。図6に示すように、本実施形態の塗膜の性状の変動量の予測システム10は、コンピュータ11を備えている。コンピュータ11は、機械学習する機能を有する。また、コンピュータ11は、人工知能モデル作成部12及び性状予測部13を備える。人工知能モデル作成部12は、機械学習する機能を有し、製造条件の変動量を入力として含み、塗膜の性状の変動量を出力とする、所定の人工知能モデルを機械学習により作成するものである。性状予測部13は、人工知能モデルにおいて、製造条件の変動量を入力することによって、塗膜の性状の変動量を算出して予測するものである。人工知能モデル作成部12及び性状予測部13は、プロセッサとすることができる。
製造条件、塗膜の性状、機械学習等についての詳細は、図1に示した方法の実施形態と同様であるので、再度の説明を省略する。
図7は、本発明の一実施形態にかかる製造条件の変動量の予測システムのブロック図である。本システムは、対象物に塗料を塗布して塗布物を製造するに際し、製造条件を変動させて調整することにより、塗布物の塗膜の目標となる性状を有する塗布物を製造する場合に、目標を達成するために最適な製造条件の変動量を予測するシステムである。図7に示すように、本実施形態の製造条件の予測システム20は、コンピュータ21で構成されている。コンピュータ21は、機械学習する機能を有する。また、コンピュータ21は、人工知能モデル作成部22、性状予測部23、記憶部24、及び通信部25を有する。これらについては、図6に示した実施形態での人工知能モデル作成部12、性状予測部13、記憶部14、及び通信部15で説明したのと同様であるため、詳細な説明を省略する。
製造条件、塗膜の性状、機械学習等についての詳細は、図2に示した方法の実施形態と同様であるので、再度の説明を省略する。
<白色原色塗料の調製例>
樹脂としてアクリル樹脂 20質量部、有機溶剤としてイソホロン 35質量部を加えて、ディスパーを用いて樹脂を均一に溶解した後、白色顔料として酸化チタン 46質量部を混合し、サンドミル(分散媒体:ガラスビーズ)を用いて、顔料粗粒の最大粒子径が10μm以下になるまで分散し、白色原色塗料を調製した。
<その他の原色塗料の調製例>
各材料種及び量を以下の表1のとおりに変更する以外は、前記白色原色塗料の調製例と同様の方法で、黒色原色塗料1、黒色原色塗料2、黄色原色塗料及び赤色原色塗料を調製した。それぞれの原色塗料の配合を表1に示す。
光沢調整剤は、つや消し剤として用いる市販品のシリカ(二酸化ケイ素)を用いた。
<粘度調整剤>
粘度調整剤は、溶剤として用いるイソホロンを用いた。
アクリル樹脂 5質量部、フッ素樹脂 25質量部、イソホロン 35質量部及びシクロヘキサノン 35質量部を加えて、ディスパーを用いて樹脂を均一に溶解した後、白色原色塗料 25質量部、黒色原色塗料1 2質量部、黄色原色塗料 69質量部、赤色原色塗料 4質量部及び光沢調整剤1 2質量部を加えて、ディスパーを用いて均一に混合し、塗料組成物1(固形分濃度:47質量%)を調製した。
各材料種及び量を以下の表2のとおりに変更する以外は、前記塗料組成物1の調製例と同様に調製し、塗料組成物2~4を調製した。それぞれの塗料組成物の配合を表2に示す。
・アクリル樹脂:パラロイドB44(Rohm & Haas社製)、固形分濃度:100質量%
・フッ素樹脂:KYNAR500(ARKEMA社製)、固形分濃度:100質量%
・白色顔料:TI-PURE R-706(酸化チタン、デュポン社製)
・黒色顔料1:三菱カーボンブラック MA-100(カーボンブラック、三菱化学社製)
・黒色顔料2:SUNBLACK X15(カーボンブラック、白石カルシウム社製)
・黄色顔料:TAROX 合成酸化鉄 LL-XLO(黄色酸化鉄、チタン工業社製)
・赤色顔料:TODA COLOR 140ED(酸化鉄、戸田工業社製)
・光沢調整剤1:GASIL HP395(合成シリカ、INEOS SILICAS社製)
・光沢調整剤2:サイリシア435(二酸化ケイ素、富士シリシア化学社製)
・有機溶剤:イソホロン(ARKEMA社製)
・有機溶剤:シクロヘキサノン(宇部興産社製)
以下の工程は一つの塗装ラインの中で、一気通貫で実施されるものである。素材(亜鉛-アルミニウム合金めっき鋼板)に下塗り塗料として、ファインタフG プライマー(プライマー1)を、ロールコーター塗装(基準膜厚:5μm)した後、その到達最高温度210℃ となる条件で60秒間焼付けを行い、下塗り塗膜を形成した。次に、塗料組成物1 を、ロールコーター塗装(基準膜厚:15μm)した後、素材最高到達温度250℃ で60秒間焼付けた後、ただちに冷却させることで、調整前塗料組成物の塗膜を形成した。
下塗り塗料や塗料組成物の膜厚は、切り出した塗装板の塗膜をサンドブラストで除去した前後の重量と塗膜比重から算出した。
色差の測定方法
塗膜の色彩はLab Scan XE(HunterLab社製)を使用して、目的とする塗装板の色彩との差を色差とした。
・人工知能モデルの作成におけるデータの入力
塗装ラインにおいて塗膜の膜厚、色彩、光沢を調整する塗装パラメータとして、アプリケーションロール(APロール)スピード変動量、塗着圧(バックアップロールとAPロールの塗着圧力)変動量、原色補正添加量の各データをコンピュータに入力した。
また、膜厚、色彩、光沢については塗装パラメータ調整前の膜厚、色彩、光沢、塗料粘度および塗装パラメータ調整後の膜厚変動量、色彩変動量、光沢変動量を用いた。
・人工知能モデルの作成
事前にデータクレンジング、ゼロ点導入、コンビネーション、累乗処理、正規化などの前処理を実施した。コンピュータに調整前膜厚、色彩、光沢、塗料粘度、およびAPロールスピード変動量、塗着圧変動量、原色塗料添加量を説明変数とし、塗装パラメータ調整後の膜厚変動量、色彩変動量、光沢変動量を目的変数とする、所定の人工知能モデルを作成した。機械学習のアルゴリズムとして勾配ブースティング、ランダムフォレスト、サポートベクターマシーン、ニューラルネットワークについて交差検証法を用いて、最適なハイパーパラメータを決定し、決定したハイパーパラメータで全データを再学習させることにより、製造条件の変動量のデータを説明変数とし、調整後塗膜性状の変動量を目的変数とする、所定の人工知能モデルを作成した。
同様にして色彩のL*、a*、b*、光沢の変動値を予測する人工知能モデルを作成し、合計5つの人工知能モデルを得た。
比較人工知能作成例
データの入力と人工知能モデルの作成における説明変数である、APロールスピード変動量、塗着圧変動量、原色塗料添加量を、APロールスピード絶対値、塗着圧絶対値、原色配合比とし、説明変数に調整前膜厚、色彩、光沢、塗料粘度を使用せずに人工知能モデルを作成する以外は、人工知能作成例と同様に作成し、比較人工知能モデル1を得た。
同様にして色彩のL*、a*、b*、光沢の絶対値を予測する比較人工知能モデルを作成し、合計5つの人工知能モデルを得た。
以下の表3に人工知能作成例を示す。
塗膜性状合格判定基準
下記を満たすことを合格判定の基準とした。
標準板と調整塗料の塗装板の膜厚、L*値、a*値、b*値の差をΔFT、ΔL*、Δa*、Δb*とした。
・ΔFT1μm以下で、かつ色彩が下記3点の条件すべてを満たすことを合格判定の基準とした。
・ΔE=√(ΔL*2 +Δa*2 +Δb*2)の値が0.1以下であり、かつΔL*、Δa*、Δb*各々が0.1以下であること。
・Δ光沢値は、光沢値が10以下は標準板との差異が0.5以下、10~20は1以下、30以上は3以下であること。を合格とした。
調整回数の合格判定基準
自動化が可能で、かつ、調整前の塗膜性状の膜厚の目標とする膜厚に対しての差異が5μm以内、ΔE2以内の時、塗膜性状の調整回数が3回以内で合格基準に達した場合、合格とした。
<実施例1>
目標とする膜厚に対しての差異が約3μm、約ΔE=1の塗膜性状を有する塗料組成物1および製造条件において、粘度変動量、塗着圧変動量、APロール周速変動量に対し許容範囲を設定し、その範囲内で各々1000万通りの数字を生成したのちに、ランダムに組み合わせて、調整前膜厚と併せて、勾配ブースティングで学習済みの5つの人工知能モデルに入力し、塗装パラメータ調整後の膜厚変動量、L*値変動量、a*値変動量、b*値変動量、光沢変動量の1000万通りの候補データを得た。
標準板とのΔFTの100%、ΔL*値、Δa*値、Δb*値の90%、Δ光沢80%の点を変動量のギャップ目標値とし、ギャップ目標値を満たすΔL*値、Δa*値、Δb*値、光沢の変動量を同時に満たす製造条件を1000万通りの中から一つ取得し調整を行った。
合格しなかったので、2回目の調整をギャップ目標値100%で同様に実施したところ合格し、調整回数は2回で合格であった。
このような調整を塗料組成物2~4でも実施した平均調整回数は2.0回であった。
目標とする膜厚に対しての差異が約2μm、約ΔE=2.5の塗膜性状を有する塗料組成物1および製造条件において、粘度変動量、塗着圧変動量、原色塗料添加量、焼き付け温度変動量、製造ラインの温度の変動量に対し許容範囲を設定し、その範囲内で各々1000万通りの数字を生成したのちに、ランダムに組み合わせて、調整前膜厚、色彩、光沢、塗料粘度と併せて、ニューラルネットワークで学習済みの5つの人工知能モデルに入力し、塗装パラメータ調整後の膜厚変動量、L*値変動量、a*値変動量、b*値変動量、光沢変動量の1000万通りの候補データを得た。
標準板とのΔFTの100%、ΔL*値、Δa*値、Δb*値の90%、Δ光沢90%の点を変動量のギャップ目標値とし、ギャップ目標値を満たすΔL*値、Δa*値、Δb*値、光沢の変動量を同時に満たす製造条件を1000万通りの中から一つ取得し、調整を行ったところ合格した。
このような調整を塗料組成物2~4でも実施した平均調整回数は1.5回であった。
目標とする膜厚に対しての差異が約2μm、約ΔE=2.5の塗膜性状を有する塗料組成物1および製造条件において、人工知能モデルを使用せずに、経験年数3年以上の塗装管理者が、各種塗装条件を調整して膜厚、色彩および光沢調整を合格するまで実施したところ調整回数は3回で合格した。
このような調整を塗料組成物2~4でも実施した平均調整回数は3.5回であった。
目標とする膜厚に対しての差異が約3μm、約ΔE=1の塗膜性状を有する塗料組成物1および製造条件において、粘度絶対値、塗着圧絶対値、APロールスピード絶対値に対し許容範囲を設定し、その範囲内で各々1000万通りの数字を生成したのちに、ランダムに組み合わせて、学習済みの5つの比較人工知能モデルに入力し、塗装パラメータ調整後の膜厚絶対値、L*値絶対値、a*値絶対値、b*値絶対値、光沢絶対値の1000万通りの候補データを得た。
標準板とのΔFTの100%、ΔL*値、Δa*値、Δb*値の90%、Δ光沢80%の点の変動量を各々の絶対値に加算してギャップ目標値とし、ギャップ目標値を同時に満たす、L*絶対値、a*絶対値、b*絶対値、光沢絶対値を持つ、APロール周速絶対値、塗着圧絶対値、原色配合比の組み合わせを1000万通りの中から一つ取得し、調整を行った。合格しなかったので、2回目の調整をギャップ目標値100%で同様に実施したが合格せず、2回目と同様にその後3回(計5回)調整しても合格しなかった。
このような調整を4つの塗料組成物2~4でも実施したが、いずれも5回以内では調整できなかった。
11:コンピュータ、
12:人工知能モデル作成部、
13:性状予測部、
14:記憶部、
15:通信部、
20:予測システム、
21:コンピュータ、
22:人工知能モデル作成部、
23:性状予測部、
24:記憶部、
25:通信部、
26:製造条件予測部、
27:表示部
Claims (15)
- 対象物に塗料を塗布して得られた塗膜の性状の変動量を予測する方法であって、
製造条件の変動量を入力として含み、前記塗膜の性状の変動量を出力とする、所定の人工知能モデルを取得する、人工知能モデル取得工程と、
コンピュータにより、前記人工知能モデルにおいて、前記製造条件の変動量を入力することによって、前記塗膜の性状の変動量を予測する、性状変動量予測工程と、を含むことを特徴とする、塗膜の性状の変動量の予測方法。 - 対象物に塗料を塗布して塗布物を製造するに際し、製造条件を変動させて調整することにより、塗布物の塗膜の目標となる性状を有する塗布物を製造する場合に、目標を達成するために最適な前記製造条件の変動量を予測する方法であって、
製造条件の変動量を入力として含み、前記塗膜の性状の変動量を出力とする、所定の人工知能モデルを取得する、人工知能モデル取得工程と、
コンピュータにより、前記人工知能モデルにおいて、前記製造条件の変動量を入力することによって、前記塗膜の性状の変動量を予測する、性状変動量予測工程と、を含み、
前記性状変動量予測工程では、複数の前記製造条件の変動量を入力し、各々の前記製造条件の変動量に対応する複数の前記塗膜の性状の変動量を予測し、
予測した前記複数の前記塗膜の性状の変動量に基づいて、最適な前記製造条件の変動量を、前記目標を達成するために最適な前記製造条件の変動量として決定する、製造条件変動量予測工程をさらに含むことを特徴とする、製造条件の変動量の予測方法。 - 前記製造条件の変動量と前記塗膜の性状の変動量との関係を示す関係データを前記コンピュータに入力する、関係データ入力工程をさらに含み、
前記人工知能モデル取得工程は、前記コンピュータにより、前記関係データ入力工程において入力された前記関係データを教師データとして機械学習することにより行われる、請求項1又は2に記載の方法。 - 前記塗膜の性状は、前記塗膜の色彩、光沢、粘度、膜厚、平滑性、目視外観、フリップフロップ性、及び塗膜異常のいずれか1つ以上を含む、請求項1~3のいずれか一項に記載の方法。
- 前記製造条件は、
(a)塗料の粘度及び加熱残分のいずれか1つ以上、
(b)前記対象物に塗料を塗布する工程における、ロール周速、前記対象物への塗着圧、及び塗料の流量のいずれか1つ以上、
(c)焼き付け工程における焼き付け温度及び焼き付け時間のいずれか1つ以上、及び
(d)製造ラインの温度、湿度及び塗料液温のいずれか1つ以上、
の(a)~(d)うちのいずれか1つ以上を含む、請求項1~4のいずれか一項に記載の方法。 - 前記人工知能モデル取得工程は、機械学習によって前記人工知能モデルを作成する工程であり、
前記機械学習は、ニューラルネットワーク、複数の決定木から構成されるアンサンブルツリー、又は部分的最小二乗回帰法による予測アルゴリズムを用いている、請求項1~5のいずれか一項に記載の方法。 - 前記機械学習は、ニューラルネットワーク、ランダムフォレスト、又は勾配ブースティング法を用いた、請求項6に記載の方法。
- 前記関係データ入力工程後に、前記関係データに対してデータクレンジングを行う工程をさらに含む、請求項1~7のいずれか一項に記載の方法。
- 前記製造条件変動量予測工程において、予測した前記複数の前記塗膜の性状の変動量の中から前記目標となる性状との差が最も小さくなる前記塗膜の性状の変動量に対応する前記製造条件の変動量を、前記目標を達成するために最適な前記製造条件の変動量として決定する、請求項2に記載の方法。
- 請求項2に記載の方法により予測した前記製造条件の変動量の分だけ前記製造条件を調整して前記対象物に前記塗料を塗布し、前記塗膜を得る、条件調整後製造工程と、
得られた前記塗布物の前記塗膜の前記性状を測定する、性状測定工程と、をさらに含み、
前記性状測定工程において測定された前記塗膜の前記性状が、所望の性状から所定の閾値以上ずれた場合に、前記性状変動量予測工程、前記製造条件変動量予測工程、前記条件調整後製造工程、及び前記性状測定工程を繰り返す、塗布物の製造方法。 - 前記性状測定工程において測定された前記塗膜の前記性状が、所望の性状から所定の閾値以上ずれることが、所定の回数繰り返された場合に、
前記製造条件の変動量を入力として含み、前記塗膜の性状の変動量を出力とする、別の人工知能モデルを取得する、人工知能モデル取得サブ工程と、
前記コンピュータにより、前記別の人工知能モデルにおいて、前記製造条件の変動量を入力することによって、前記塗膜の性状の変動量を予測する、性状変動量予測サブ工程と、を含み、
前記性状変動量予測サブ工程では、予め候補として選定した複数の前記製造条件の変動量を入力し、各々の前記製造条件の変動量に対応する複数の前記塗膜の性状の変動量を予測し、
予測した前記複数の前記塗膜の性状の変動量に基づいて、別の最適な前記製造条件の変動量を、前記目標を達成するために最適な別の前記製造条件の変動量として決定する、製造条件変動量予測サブ工程をさらに含み、
予測した前記別の製造条件の変動量の分だけ前記製造条件を調整して、前記対象物に前記塗料を塗布して前記塗布物を得る、条件調整後製造サブ工程と、
得られた前記塗布物の前記塗膜の前記性状を測定する、性状測定サブ工程と、をさらに含み、
前記性状測定工程において測定された前記塗膜の前記性状が、所望の性状から所定の閾値以上ずれた場合に、前記人工知能モデル取得サブ工程、前記性状変動量予測サブ工程、前記製造条件変動量予測サブ工程、前記条件調整後製造サブ工程、及び前記性状測定サブ工程を繰り返す、請求項10に記載の塗布物の製造方法。 - 前記製造条件の変動量と、前記性状測定工程において測定された前記塗膜の性状の変動量と、の関係を示す関係データを前記コンピュータに入力して前記関係データをアップデートする、関係データアップデート工程をさらに含む、請求項10に記載の塗布物の製造方法。
- 対象物に塗料を塗布して得られた塗膜の性状の変動量を予測するシステムであって、
製造条件の変動量を入力として含み、前記塗膜の性状の変動量を出力とする、所定の人工知能モデルを取得する、人工知能モデル取得部と、
前記人工知能モデルにおいて、前記製造条件の変動量を入力することによって、前記塗膜の性状の変動量を算出して予測する、性状予測部と、を備えたコンピュータを備えていることを特徴とする、システム。 - 対象物に塗料を塗布して塗布物を製造するに際し、製造条件を変動させて調整することにより、塗布物の塗膜の目標となる性状を有する塗布物を製造する場合に、目標を達成するために最適な前記製造条件の変動量を予測するシステムであって、
前記製造条件の変動量を入力として含み、前記塗膜の性状の変動量を出力とする、所定の人工知能モデルを取得する、人工知能モデル取得部と、
前記人工知能モデルにおいて、複数の前記製造条件の変動量を入力することによって、複数の前記塗膜の性状の変動量を算出して予測する、性状予測部と、
予測した前記複数の前記塗膜の性状の変動量に基づいて、最適な前記製造条件の変動量を、前記目標を達成するために最適な前記製造条件の変動量として決定する、製造条件変動量予測部と、を備えたコンピュータを備えていることを特徴とする、システム。 - 前記製造条件は、複数の工程の各々における工程製造条件を有し、
請求項2に記載の方法により予測した製造条件の変動量の分だけ前記製造条件を調整した際の、各前記工程製造条件をモニタリングし、
各前記工程製造条件のうち、前記変動量から所定の値以上ずれて調整された工程製造条件が観測された場合に、
前記人工知能モデルにおいて、モニタリングされた前記工程製造条件を固定した上で、他の複数の前記製造条件の変動量を入力することによって、複数の前記塗膜の性状の変動量を算出して予測する、性状再予測工程と、
予測した前記複数の前記塗膜の性状の変動量に基づいて、最適な前記製造条件の変動量を、前記目標を達成するために最適な前記製造条件の変動量として再決定する、製造条件変動量再予測工程と、をさらに含み、
現工程より後の工程における前記製造条件の変動量を、前記製造条件変動量再予測工程の結果に基づいて調整し直す、請求項11に記載の製造方法。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2023530482A JPWO2022270506A1 (ja) | 2021-06-21 | 2022-06-21 | |
CN202280040372.XA CN117529370A (zh) | 2021-06-21 | 2022-06-21 | 涂膜的性状的变化量的预测方法和预测系统、涂敷物的制造条件的变化量的预测方法和预测系统、涂敷物的制造方法 |
AU2022299794A AU2022299794A1 (en) | 2021-06-21 | 2022-06-21 | Paint film property change amount prediction method and system, coated object production condition change amount prediction method and system, and coated object production method |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021-102753 | 2021-06-21 | ||
JP2021102753 | 2021-06-21 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022270506A1 true WO2022270506A1 (ja) | 2022-12-29 |
Family
ID=84545442
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2022/024739 WO2022270506A1 (ja) | 2021-06-21 | 2022-06-21 | 塗膜の性状の変動量の予測方法及び予測システム、塗布物の製造条件の変動量の予測方法及び予測システム、塗布物の製造方法 |
Country Status (4)
Country | Link |
---|---|
JP (1) | JPWO2022270506A1 (ja) |
CN (1) | CN117529370A (ja) |
AU (1) | AU2022299794A1 (ja) |
WO (1) | WO2022270506A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116641035A (zh) * | 2023-07-26 | 2023-08-25 | 南京诺源医疗器械有限公司 | 一种用于腹腔镜光学件的镀膜方法 |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06142565A (ja) * | 1992-09-16 | 1994-05-24 | Nissan Motor Co Ltd | 自動車の塗装条件管理装置 |
JP2001070846A (ja) * | 1999-07-31 | 2001-03-21 | Abb Res Ltd | 塗料噴霧ユニットに関する噴霧パラメータの決定方法 |
JP2002180295A (ja) * | 2000-12-20 | 2002-06-26 | Fuji Heavy Ind Ltd | 電着塗装解析方法および電着塗装解析装置 |
JP2006218425A (ja) * | 2005-02-14 | 2006-08-24 | Kansai Paint Co Ltd | 塗装方法、塗装制御装置及び塗装設備 |
JP2007029765A (ja) * | 2005-06-20 | 2007-02-08 | Jfe Steel Kk | 塗装膜厚制御方法及びそのシステム |
JP2007508937A (ja) * | 2003-10-24 | 2007-04-12 | イー・アイ・デュポン・ドウ・ヌムール・アンド・カンパニー | 塗装パラメータを予測し適用するための方法およびその使用 |
JP2017200681A (ja) * | 2016-05-06 | 2017-11-09 | ダイハツ工業株式会社 | 塗装条件設定方法及びプログラム |
WO2019171498A1 (ja) * | 2018-03-07 | 2019-09-12 | 日立化成株式会社 | 塗布制御システム、塗布制御方法、および塗布制御プログラム |
-
2022
- 2022-06-21 JP JP2023530482A patent/JPWO2022270506A1/ja active Pending
- 2022-06-21 WO PCT/JP2022/024739 patent/WO2022270506A1/ja active Application Filing
- 2022-06-21 CN CN202280040372.XA patent/CN117529370A/zh active Pending
- 2022-06-21 AU AU2022299794A patent/AU2022299794A1/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06142565A (ja) * | 1992-09-16 | 1994-05-24 | Nissan Motor Co Ltd | 自動車の塗装条件管理装置 |
JP2001070846A (ja) * | 1999-07-31 | 2001-03-21 | Abb Res Ltd | 塗料噴霧ユニットに関する噴霧パラメータの決定方法 |
JP2002180295A (ja) * | 2000-12-20 | 2002-06-26 | Fuji Heavy Ind Ltd | 電着塗装解析方法および電着塗装解析装置 |
JP2007508937A (ja) * | 2003-10-24 | 2007-04-12 | イー・アイ・デュポン・ドウ・ヌムール・アンド・カンパニー | 塗装パラメータを予測し適用するための方法およびその使用 |
JP2006218425A (ja) * | 2005-02-14 | 2006-08-24 | Kansai Paint Co Ltd | 塗装方法、塗装制御装置及び塗装設備 |
JP2007029765A (ja) * | 2005-06-20 | 2007-02-08 | Jfe Steel Kk | 塗装膜厚制御方法及びそのシステム |
JP2017200681A (ja) * | 2016-05-06 | 2017-11-09 | ダイハツ工業株式会社 | 塗装条件設定方法及びプログラム |
WO2019171498A1 (ja) * | 2018-03-07 | 2019-09-12 | 日立化成株式会社 | 塗布制御システム、塗布制御方法、および塗布制御プログラム |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116641035A (zh) * | 2023-07-26 | 2023-08-25 | 南京诺源医疗器械有限公司 | 一种用于腹腔镜光学件的镀膜方法 |
CN116641035B (zh) * | 2023-07-26 | 2023-10-13 | 南京诺源医疗器械有限公司 | 一种用于腹腔镜光学件的镀膜方法 |
Also Published As
Publication number | Publication date |
---|---|
AU2022299794A1 (en) | 2024-01-18 |
CN117529370A (zh) | 2024-02-06 |
JPWO2022270506A1 (ja) | 2022-12-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP7116266B1 (ja) | 塗料性状の予測方法、補正配合組成の予測方法、塗料性状の予測システム、補正配合補正の予測システム、及び塗料の製造方法 | |
WO2022270506A1 (ja) | 塗膜の性状の変動量の予測方法及び予測システム、塗布物の製造条件の変動量の予測方法及び予測システム、塗布物の製造方法 | |
US10401224B2 (en) | Method for ascertaining texture parameters of a paint | |
JP2017508952A (ja) | 色処方を決定するための方法及びシステム | |
JP2017508952A5 (ja) | ||
WO2021140690A1 (ja) | 品質管理システムおよび品質管理プログラム | |
JP2006218426A (ja) | 塗装方法、塗装制御装置及び塗装設備 | |
US10174426B2 (en) | Coated metal sheet, method for manufacturing same, and exterior building material | |
US20230304941A1 (en) | System and method for assessing a coated surface with respect to surface defects | |
CN113362194B (zh) | 涂装质量预测装置以及已学习模型的生成方法 | |
Mao et al. | Coating thickness modeling and prediction for hot-dip galvanized steel strip based on GA-BP neural network | |
US20230173533A1 (en) | Coating production line system | |
US20220107631A1 (en) | Coating Process and Quality Control of Coated Objects | |
CN104136913A (zh) | 利用湿色测量预测由涂料成分得到的涂料金属光泽的方法 | |
JP2006218425A (ja) | 塗装方法、塗装制御装置及び塗装設備 | |
US12085903B2 (en) | Method, device, and system for configuring a coating machine | |
Uribe-Padilla et al. | A viscosity-mediated model for relating gloss and film thickness of coatings | |
WO2022202365A1 (ja) | 検査支援システム、検査支援方法、及びプログラム | |
JP2023001804A (ja) | 自動車ボディ及び/又は自動車部品の塗膜の塗膜品質の予測方法及び予測システム、塗装条件の予測方法及び予測システム、自動車ボディ及び/又は自動車部品の複層塗膜形成方法 | |
JP7429456B2 (ja) | 自動調整方法及び自動調整システム | |
US10378111B2 (en) | Coated metal sheet, method for manufacturing same, and exterior building material | |
WO2024143003A1 (ja) | 塗料の製造方法 | |
Tout et al. | Non-stationary process monitoring for change-point detection with known accuracy: Application to wheels coating inspection | |
CN114016023B (zh) | 钢材激光熔覆工艺参数的优化方法和应用 | |
WO2023013463A1 (ja) | 学習モデル生成方法、プログラム、記憶媒体、学習済みモデル |
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: 22828422 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2023530482 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202280040372.X Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2022299794 Country of ref document: AU Ref document number: AU2022299794 Country of ref document: AU |
|
ENP | Entry into the national phase |
Ref document number: 2022299794 Country of ref document: AU Date of ref document: 20220621 Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 22828422 Country of ref document: EP Kind code of ref document: A1 |