JP2023001804A - Prediction method and prediction system for coating film quality of coating film of automotive body and/or automobile component, prediction method and prediction system for coating condition, and multi-layer coating film formation method for and automotive body and/or automobile component - Google Patents

Abstract

To provide a prediction method and a prediction system for coating film quality of a coating film of an automotive body and/or automobile component that can speedily predict an evaluation result of coating film quality of the coating film of the automobile body and/or automobile component, a prediction method and a prediction system for coating condition that can speedily predict optimum coating conditions, and a forming method that can form a coating film having desired coating film quality.SOLUTION: Through machine learning by a computer, a predetermined artificial intelligence model is created which inputs data on coating conditions of a coating process and outputs coating quality of a coating film of an automobile body and/or automobile component after the coating process. The computer inputs the coating conditions of the coating process with the artificial intelligence model so as to calculate and predict the coating film quality of the coating film of the automobile body and/or automobile component after the coating process.SELECTED DRAWING: Figure 1

Description

The present invention relates to a method and system for predicting coating film quality of a coating film on an automobile body and/or automobile parts, a method and system for predicting painting conditions, and a method for forming a multi-layer coating film on an automobile body and/or automobile parts.

BACKGROUND ART Conventionally, various techniques for coating automobile bodies and automobile parts have been proposed (for example, Patent Document 1). The coating process includes, for example, any one or more of a chipping primer coating process, an intermediate coating process, a base coating process, and a clear coating process, and often consists of a plurality of processes.

In such painting, the evaluation of the paint film quality of the paint film of the automobile body and / or automobile parts is performed in the inspection process within the painting process and / or after the painting process (after baking and drying the paint film). Afterwards, the coating film was evaluated for quality such as color, gloss, film thickness, sharpness, and smoothness using a measuring instrument or the like.

JP 2021-041376 A

However, in the production method that evaluates the quality as described above, after the coating process for applying the paint, it is necessary to harden the applied wet paint film so that a measuring instrument etc. can be used. There is a problem that it takes time to obtain the evaluation result of the coating film quality, and if the quality defect is found, the productivity decreases due to repair, repainting, disposal, etc. In addition, even if it is necessary to adjust the coating conditions before painting, it is not possible to instantly judge it or reflect the judgment results in the coating conditions when painting, which may lead to a decrease in productivity. was there.

Therefore, the present invention provides a method for predicting the coating film quality of a coating film on an automobile body and/or an automobile component, which can quickly predict the evaluation result of the coating film quality of the coating film on the automobile body and/or the automobile component, and Prediction system, method and system for predicting coating conditions capable of rapidly obtaining optimum coating conditions, and multi-layers of automobile bodies and/or automobile parts capable of forming coatings with desired coating film quality An object of the present invention is to provide a coating film forming method.

The gist and configuration of the present invention are as follows.
(1) A method for predicting the coating film quality of a coating film obtained by coating an automobile body and/or automobile parts,
Using computer machine learning to create a predetermined artificial intelligence model that takes as input paint condition data in the paint process and outputs the paint film quality of the paint film on the automobile body and/or automobile parts after the paint process. , an artificial intelligence model creation process;
The computer calculates and predicts the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process by inputting the painting conditions in the painting process in the artificial intelligence model. and a film quality prediction step.

(2) A method of predicting the optimum painting conditions for the painting process to achieve a goal when painting an automobile body and/or automobile parts through a painting process,
Machine learning by computer creates a predetermined artificial intelligence model that takes as input paint condition data in the paint process and outputs the paint film quality of the paint film on the automobile body and/or automobile parts after the paint process. an artificial intelligence model creation process;
The computer calculates and predicts the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process by inputting the painting conditions in the painting process in the artificial intelligence model; and a coating film quality prediction step,
In the coating film quality prediction step, the coating conditions in a plurality of the coating processes are input, and the coating film of the auto body and/or the auto parts after the plurality of coating processes corresponding to each of the coating conditions. anticipate quality,
Based on the predicted paint film quality of the paint film of the automobile body and/or automobile parts after the plurality of the painting steps, the optimum painting condition is determined as the optimum painting condition for achieving the target. , a painting condition prediction method, further comprising a painting condition prediction step.

(3) a relationship data input step of inputting into a computer relationship data indicating the relationship between the coating conditions in the coating step and the coating film quality of the coating film of the automobile body and/or automobile parts after the painting step; further includes
The method according to (1) or (2) above, wherein the artificial intelligence model creation step is performed by machine learning with the computer using the relational data input in the relational data input step as teacher data.

(4) Any one of (1) to (3) above, wherein the coating film quality includes any one or more of color, gloss, film thickness, sharpness, and smoothness of the coating film. the method of.

(5) The sharpness and / or smoothness uses a wave scan value as an index,
The method according to (4) above, wherein the wave scan value is any one or more of du, Wa, Wb, Wc, Wd, We, Lw, and Sw.

(6) The method according to any one of (1) to (5) above, wherein the machine learning uses a prediction algorithm based on an ensemble tree composed of a plurality of decision trees.

(7) The method according to any one of (1) to (6) above, wherein the machine learning uses deep learning, partial least squares regression, random forest, or gradient boosting.

(8) The method according to (3) above, further comprising the step of performing data cleansing on the relational data after the relational data input step.

(9) In the painting condition prediction step, the difference from the target paint film quality is the largest among the paint film qualities of the paint film of the automobile body and / or automobile part after the plurality of the painting processes predicted. (2) above, wherein the coating conditions corresponding to the coating film quality of the coating film of the automobile body and/or the automobile parts after the coating process, which becomes smaller, are determined as the optimum coating conditions for achieving the target; The method described in .

(10) The method according to (2) or (9) above, wherein Bayesian optimization is used in the coating condition prediction step.

(11) The painting conditions include the temperature of the painting booth where the painting process is performed, the dry-bulb temperature of the psychrometer of the painting booth, the wet-bulb temperature of the psychrometer, the relative humidity of the painting booth, and the absolute humidity of the painting booth. , paint viscosity, paint temperature, paint type, paint solid content, paint discharge amount in the painting process, linear speed of the painting machine, and the temperature of the object to be coated of the automobile body and / or automobile parts. The method according to any one of (1) to (10) above, comprising

(12) Any one of (1) to (11) above, wherein the coating step includes any one or more of a chipping primer coating step, an intermediate coating step, a base coating step, and a clear coating step. the method of.

(13) A coating condition adjustment step of adjusting the coating conditions predicted by the method described in (2) above;
and a post-adjustment coating step of performing coating under the coating conditions adjusted in the coating condition adjustment step.

(14) a measuring step of measuring the coating film quality of the coating film of the automobile body and/or automobile parts after the post-adjustment painting step;
Updating relational data, inputting relational data indicating a relation between the coating conditions adjusted in the coating condition adjusting step and the coating film quality measured in the measuring step into the computer to update the relational data. The method for forming a multilayer coating film for an automobile body and/or automobile part according to (13) above, further comprising steps.

(15) A system for predicting coating film quality of a coating film on an automobile body and/or automobile parts after a painting process,
A predetermined artificial intelligence model having a function of machine learning, which inputs data of painting conditions in the painting process and outputs the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process. an artificial intelligence model creation unit to create;
In the artificial intelligence model, a paint film quality prediction unit that calculates and predicts the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process by inputting the painting conditions in the painting process. A system for predicting coating film quality of a coating film on an automobile body and/or automobile parts, characterized by comprising a computer comprising:

(16) A system for predicting the optimum painting conditions for the painting process to achieve a goal when painting an automobile body and/or automobile parts through a painting process,
A predetermined artificial intelligence model having a function of machine learning, which inputs data of coating conditions in the coating process and outputs the coating film quality of the coating film of the car body and/or car parts after the coating process. an artificial intelligence modeling unit that creates
In the artificial intelligence model, by inputting the painting conditions in the painting process, the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process is calculated and predicted, predicting the paint film quality. and
The coating film quality prediction unit inputs the coating conditions in a plurality of the coating processes, and the coating film of the auto body and/or the auto parts after a plurality of the coating processes corresponding to each of the coating conditions. anticipate quality,
Based on the predicted paint film quality of the paint film of the automobile body and/or automobile parts after the plurality of the painting steps, the optimum painting condition is determined as the optimum painting condition for achieving the target. , a painting condition prediction system, further comprising a painting condition prediction unit.

INDUSTRIAL APPLICABILITY According to the present invention, a method for predicting the coating film quality of a coating film on an automobile body and/or an automobile component, which can quickly predict the evaluation result of the coating film quality of the coating film on the automobile body and/or the automobile component, and Prediction system, method and system for predicting coating conditions capable of rapidly obtaining optimum coating conditions, and multi-layers of automobile bodies and/or automobile parts capable of forming coatings with desired coating film quality A coating film forming method can be provided.

1 is a flow chart of a method for predicting coating film quality of a coating film on an automobile body and/or automobile parts according to an embodiment of the present invention; BRIEF DESCRIPTION OF THE DRAWINGS It is the schematic of an example of the coating process to which the method of this embodiment targets. 4 is a flowchart of a method for predicting coating conditions in a coating process according to one embodiment of the present invention; 1 is a flow chart of a method for forming a multilayer coating film for an automobile body and/or automobile parts according to one embodiment of the present invention. 1 is a block diagram of a paint film quality prediction system for a paint film on an automobile body and/or automobile part according to an embodiment of the present invention; FIG. 1 is a block diagram of a painting condition prediction system according to an embodiment of the present invention; FIG.

Embodiments of the present invention will be exemplified in detail below with reference to the drawings.

<Method for Predicting Coating Film Quality of Automobile Body and/or Automobile Parts Coating Film>
FIG. 1 is a flowchart of a method for predicting coating film quality of a coating film on an automobile body and/or automobile parts according to one embodiment of the present invention. Here, in this specification, "automobile body" is the body of an automobile and is not included in "automotive parts." Further, "automobile parts" are, for example, automobile bumpers and the like. Hereinafter, one embodiment of a method for predicting the coating film quality of a coating film obtained by coating an automobile body and/or automobile parts will be described with reference to FIG. The method of the present embodiment can be executed, for example, using a system for predicting coating film quality of a coating film on an automobile body and/or automobile parts according to an embodiment of the present invention, which will be described later. Here, the predicted coating film quality preferably includes at least one of color, gloss, film thickness, sharpness, and smoothness of the coating film. For 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. As an example, 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. Alternatively, 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. Other indicators of color space, such as L ^* C ^* h ^* color space (JIS Z8781-4 (2013)), Hunter Lab (J "Photoelectric Color-Difference Meter". Journal of the Optical Society of America 38 (7): 661 (Proceedings of the Winter Meeting of the Optical Society of America)), the XYZ color system (CIE 1931 XYZ color space), and other arbitrary indices 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 a gloss value can be used as an index. The gloss value can be measured using a known gloss measurement method. As an example, the glossiness of the test piece is measured using a specular gloss meter (manufactured by BYK-Gardner, trade name Micro Trigloss), using JIS K A value measured according to -5600-4-7 can be taken as a gloss value. The measurement angle can be appropriately adjusted to 20°, 45°, 60°, 75°, 85°, etc., depending on the purpose and the device to be used. The film thickness can be measured using a known film thickness measurement method, and for example, it can be measured using an electromagnetic film thickness gauge (manufactured by Kett Scientific Laboratory Co., Ltd., trade name LE-200J). . Depending on the film thickness of the sample, an electromagnetic type, eddy current type, eddy current phase type, magnetic type, or electric resistance type film thickness gauge is appropriately selected and used according to the measurement accuracy, etc. to measure the film thickness. can do. Wave scan values are preferably used as indices for the sharpness and/or 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 wave scan value can be measured, for example, by a wave scan ("Wave-Scan Dual" manufactured by BYK-Gardner).

As for the paint used for painting, the one usually used for the painting can be prepared. It can be prepared by adding a resin, a solvent, and an additive to a primary color paint prepared by preparing many types of pigments, dispersing these, and then repeating the adjustment. As will be described later, the coating process includes any one or more of a chipping primer coating process, an intermediate coating process, a base coating process, and a clear coating process. Intermediate coating, base coating, and clear coating) can be used, and the curing system, water system, and solvent system are not particularly limited.

As shown in FIG. 1, in the present embodiment, first, relational data indicating the relationship between the coating conditions in the coating process and the coating film quality of the coating film of the auto body and/or auto parts after the coating process is prepared, Input to the computer (step S101: relational data input step).

The coating process includes any one or more of a chipping primer coating process, an intermediate coating process, a base coating process, and a clear coating process. FIG. 2 is a schematic diagram of an example of a coating process targeted by the method of the present embodiment. As shown in FIG. 2, this example includes an electrodeposition coating process and a top coating process, and the top coating process includes, in order, an intermediate coating process, a base coating process, and a clear coating process. These can be performed in the same painting booth, or can be performed in different booths. The process of FIG. 2 is only an example, and there are other examples, such as 4-wet coating consisting of (in order) a chipping primer process, an intermediate coating process, a base coating process, and a clear coating process, a base coating process (in order), and a clear coating process. A 2-wet coating in which only the coating process is performed, a 3-wet coating including (in order) a first base coating process, a second base coating process, and a clear coating process can be exemplified, and various other variations are possible. . Moreover, a preheating process may be performed before each coating process.

The painting conditions are (according to the variation of the above painting process) the temperature of the painting booth where the painting process is performed, the dry-bulb temperature of the psychrometer in the painting booth, the wet-bulb temperature of the psychrometer, the relative humidity of the painting booth, and the temperature of the painting booth. Any one or more of absolute humidity, paint viscosity, paint temperature, paint type, paint solid content, paint discharge amount in the painting process, linear speed of the paint machine, and temperature of the object to be coated of the automobile body and/or automobile parts is preferably included.
The coating conditions preferably include the properties of the coating film obtained up to (any of the steps) up to the previous step (when the coating step has a plurality of steps). Specifically, if the electrodeposition coating is included in the previous process, it is the surface roughness of the electrodeposition coating film, and if the baking type intermediate coating is included in the previous process, the baking intermediate coating Wavescan values and/or surface roughness of the film surface. This can improve the accuracy of prediction. However, unlike the other coating conditions exemplified above, the properties of the coating film obtained up to the previous step are not subject to subsequent adjustment. Here, the surface roughness can be measured by any known measuring method, and as an example, it can be measured using a surface roughness measuring instrument (manufactured by Mitutoyo Co., Ltd., trade name Surftest SJ-210). can. A contact type or non-contact type measuring instrument can be appropriately selected according to the measurement location and the state of the sample.

It should be noted that the relational data is preferably updated constantly, timely, or periodically in order to enable more accurate prediction.
Moreover, it is preferable to further include a step of performing data cleansing on the relational data after the relational data input step (step S101). Here, 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 σ, the data that exceeds the average value ± 2σ 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. However, in the present disclosure, 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.

Next, as shown in FIG. 1, by machine learning by computer, the data of the painting conditions in the painting process are input, and the paint film quality of the paint film of the automobile body and / or automobile parts after the painting process is output. (step S102: artificial intelligence model creation step). In this example, 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. As relational data, measured values and set values can be used for coating conditions, and measured values for the essence of the coating film after coating under the relevant coating conditions can be used for the coating film quality. , can be data in which they are associated with each other.

Machine learning can be using random forests, gradient boosting methods, deep learning, partial least squares regression, or the like. More preferably, the machine learning uses a prediction algorithm based on an ensemble tree composed of multiple decision trees such as random forest or gradient boosting method.

Next, in the present embodiment, a computer calculates the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process by inputting the painting conditions in the painting process in the above artificial intelligence model. Predict (step S103: coating film quality prediction step).

In the coating film quality prediction step (step S103), it is preferable to narrow down the range of the coating conditions to be input in advance, for example, by referring to past data.

In this way, by inputting the painting conditions in the painting process in the artificial intelligence model, the computer can calculate the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process. . The predicted value of the paint film quality to be calculated may be a single value, or may be composed of a plurality of candidate groups. When the predicted value of the paint film quality to be calculated consists of a plurality of candidate groups, it is preferable to further include the step of selecting a single value from among them appropriately using a predetermined criterion. The predetermined criterion can be various, for example, in the case of color, it can be selected such that the E value (ΔE=(ΔL ² +Δa ² +Δb ² ) ^0.5 ) is minimized.

According to the method for predicting the paint film quality of the paint film of the automobile body and / or automobile parts of the present embodiment, the paint film can be predicted from the painting conditions of the painting process, so the completion of the painting process (all processes) There is no need to wait for the paint to bake and dry. Since the paint film quality of the paint film is predicted by the computer, it is possible to quickly predict the evaluation result of the paint film quality of the paint film. Thereby, for example, when the predicted value is out of the quality standard range, the coating conditions can be quickly adjusted.
In addition, in this embodiment, the case where the top coating process includes the intermediate coating process, the base coating process, and the clear coating process in order was shown, but such a method becomes teacher data for machine learning. If the relevant data can be prepared, various variations of the painting process can be handled.
Especially in recent years, from the viewpoint of reducing carbon dioxide emissions, instead of the production method that goes through the baking process and inspection process after the intermediate coating, the intermediate coating and the top coating are continuously performed without going through the baking process. There are many things. In such a case, there was a problem that the quality of the coating film after the intermediate coating process could not be confirmed until all the series of processes of the top coating process were completed. According to the method, it is possible to quickly obtain a predicted value of the paint film quality without waiting for the completion of the topcoat painting process or the subsequent bake-drying of the paint.

<Prediction method for coating conditions in the coating process>
FIG. 3 is a flowchart of a method for predicting coating conditions in a coating process according to one embodiment of the present invention. Hereinafter, with reference to FIG. 3, one embodiment of a method for predicting the optimum painting conditions for a painting process to achieve a goal when painting an automobile body and/or automobile parts through a painting process is illustrated. explain. The method of predicting the coating conditions of the coating process of the present embodiment can be executed, for example, using a coating condition prediction system according to an embodiment of the present invention, which will be described later. Here, as in the embodiment of FIG. 1, the coating film quality preferably includes at least one of color, gloss, film thickness, sharpness, and smoothness of the coating film. Wave scan values are preferably used as indices for the sharpness and/or 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. Also, the painting process may have various variations (the process shown in FIG. 2 as an example), as in the embodiment of FIG. 1, the temperature of the painting booth where the painting process is performed, the dry bulb temperature of the psychrometer in the painting booth, the humidity Bulb temperature, paint booth relative humidity, paint booth absolute humidity, paint viscosity, paint temperature, paint type, paint solids content, paint discharge rate in the painting process, paint machine line speed, and car body and/or car It preferably includes any one or more of the substrate temperature of the part.
Further details of the above items are the same as those of the embodiment of FIG. 1, and therefore will not be described again.

As shown in FIG. 3, in this embodiment, first, relationship data indicating the relationship between the coating conditions in the coating process and the quality of the coating film of the auto body and/or auto parts after the coating process is prepared, Input into the computer (step S201: relational data input step). Then, by machine learning using a computer, a predetermined artificial intelligence model is created that takes as input the data of the painting conditions in the painting process and outputs the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process. (Step S202: artificial intelligence model creation step). These steps S201 and S202 are the same as in the embodiment of FIG. 1, so detailed description thereof will be omitted.

As shown in FIG. 3, in the present embodiment, a computer then inputs the painting conditions in the painting process in the artificial intelligence model described above, thereby determining the paint film of the automobile body and/or automobile parts after the painting process. The coating film quality is calculated and predicted (step S203: coating film quality prediction step). Here, in the coating film quality prediction step (step S203) in the embodiment of FIG. 3, a plurality of coating conditions selected in advance as candidates are input, and a plurality of coating film properties corresponding to each coating condition are predicted. In selecting a plurality of coating conditions in advance, for example, past data can be referred to.
Also, instead of preparing a large number of candidate data in advance, a sequential search algorithm such as Bayesian optimization can be used to obtain a plurality of candidate data and coating quality prediction values.

In this embodiment, next, based on the plurality of predicted coating film qualities, the optimum painting conditions are determined as the optimum painting conditions for achieving the target (step S204: painting condition prediction step).
Specifically, as an example, in the painting condition prediction step (step S204), the target paint film quality is selected from among the paint film qualities of the paint film of the automobile body and/or automobile parts after a plurality of painting processes predicted. The painting condition corresponding to the paint film quality of the paint film of the automobile body and / or automobile part after the painting process that has the smallest difference (e.g., E value) from the is determined as the optimum painting condition to achieve the goal do.

According to the painting condition prediction method of the present embodiment, since it is possible to predict the optimum painting conditions from the current painting conditions, there is no need to wait for the completion of the painting process (all processes) or for the paint to bake and dry. As soon as the setting values of the obtained coating conditions (e.g., discharge amount, etc.) and the coating conditions that can be measured quickly (e.g., various temperatures and humidity, etc.) are obtained, the optimum coating conditions are predicted by the computer. , the optimum coating conditions can be obtained quickly. Then, when the current coating conditions deviate from the predicted values, the coating conditions can be quickly adjusted. Also, the adjustment range can be easily calculated based on the predicted value (just by taking the difference between the current value and the predicted value).
As in the case of the embodiment of FIG. 1, also in the case of the embodiment of FIG. 3, if it is possible to prepare relational data that serves as teacher data for machine learning, various variations of the painting process can be handled. .

<Method for Forming Multilayer Coating Film for Automobile Body and/or Automobile Parts>
FIG. 4 is a flow chart of a method for forming a multi-layer coating film on an automobile body and/or automobile parts according to one embodiment of the present invention. Hereinafter, one embodiment of a method for forming a multilayer coating film on an automobile body and/or automobile parts will be described with reference to FIG. Here, as in the embodiments of FIGS. 1 and 3, the coating film quality preferably includes at least one of color, gloss, film thickness, sharpness, and smoothness of the coating film. Wave scan values are preferably used as indices for the sharpness and/or 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. Also, the coating process may have various variations (the process shown in FIG. 2 as an example), as in the embodiment of FIGS. 1 and 3 . 1 and 3, the temperature of the coating booth where the coating process is performed, the dry-bulb temperature of the psychrometer in the coating booth, the dry-humidity Wet-bulb temperature of the gauge, relative humidity of the paint booth, absolute humidity of the paint booth, paint viscosity, paint temperature, paint type, paint solid content, paint discharge amount in the painting process, paint machine linear speed, automobile body and / Or it preferably includes any one or more of the temperature of the article to be coated of the automobile part.
Further details of the above items are the same as those of the embodiment of FIGS. 1 and 3, and therefore will not be described again.

As shown in FIG. 4, in the present embodiment, first, relationship data indicating the relationship between the coating conditions in the coating process and the coating film quality of the coating film of the automobile body and/or automobile parts after the coating process is prepared, Input into the computer (step S301: relational data input step). Then, by machine learning using a computer, a predetermined artificial intelligence model is created that takes as input the data of the painting conditions in the painting process and outputs the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process. (Step S302: artificial intelligence model creation step). Next, the computer calculates and predicts the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process by inputting the painting conditions in the painting process in the artificial intelligence model (step S303). : Paint film quality prediction process). Next, based on the plurality of predicted coating film qualities, the optimum painting conditions are determined as the optimum painting conditions for achieving the target (step S304: painting condition prediction step). These steps S301 to S304 are the same as in the embodiment of FIG. 3 (steps S301 and S302 are the same as in the embodiment of FIG. 1), so detailed description thereof will be omitted.

As shown in FIG. 4, in this embodiment, the predicted coating conditions are then adjusted in step S304 (step S305: coating condition adjustment step). As described above, the adjustment range of the coating conditions can be determined based on the difference between the predicted value and the current value. Next, the coating is performed under the coating conditions adjusted in the coating condition adjusting step (step S306: post-adjustment coating step).

Although not shown in FIG. 4, after the post-adjustment painting step (step S306), a measurement step (step S307) for measuring the paint film quality of the paint film of the automobile body and/or automobile parts, and painting condition adjustment. A relationship data update step of inputting relationship data indicating the relationship between the coating conditions adjusted in the step (step S305) and the coating film quality measured in the measurement step (step S307) into a computer and updating the relationship data. (Step S308) is preferably further included. As a result, it is possible to update the relevant data and predict the coating film quality and coating conditions more accurately in future predictions.

According to the method for forming a multilayer coating film for an automobile body and/or automobile parts of the present embodiment, the optimum painting conditions can be predicted from the current painting conditions. There is no need to wait for baking and drying, and as soon as the set values of the coating conditions that have already been obtained (e.g., discharge amount, etc.) and the coating conditions that can be measured quickly (e.g., various temperatures and humidity, etc.) are obtained, the computer Since the optimum painting conditions are predicted, the optimum painting conditions can be quickly obtained. Then, when the current coating conditions deviate from the predicted values, the coating conditions can be quickly adjusted. Also, the adjustment range can be easily calculated based on the predicted value (just by taking the difference between the current value and the predicted value). By performing coating under such adjusted coating conditions, it is possible to form a coating film having desired coating film quality.
As with the embodiments of FIGS. 1 and 3, the embodiment of FIG. 4 can be adapted to various painting process variations if it is possible to prepare relational data that serves as teacher data for machine learning. be able to.

<Prediction system for paint film quality of paint film on automobile body and/or automobile parts>
FIG. 5 is a block diagram of a paint film quality prediction system for a paint film on an automobile body and/or automobile parts according to an embodiment of the present invention. This system is a system for predicting the coating film quality of a coating film on an automobile body and/or automobile parts after a painting process. As shown in FIG. 5 , a system 10 for predicting coating film quality of a coating film on an automobile body and/or automobile parts according to this 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 coating film quality prediction unit 13 . The artificial intelligence model creation unit 12 creates a predetermined artificial intelligence model that takes as input paint condition data in the paint process and outputs the paint film quality of the paint film of the automobile body and/or automobile parts after the paint process. It is. The paint film quality prediction unit 13 calculates and predicts the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process by inputting the painting conditions in the painting process in the artificial intelligence model. be. The artificial intelligence model creation unit 12 and the coating film quality prediction unit 13 can be processors.

The computer 11 receives the above-described relational data. In this embodiment, 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. Moreover, it is preferable that the computer 11 further includes a functional unit that performs data cleansing on relational data. Moreover, it is preferable that the prediction system 10 include a display section (display) for displaying the prediction result.
The details of the coating film quality, coating process, coating conditions, etc. are the same as in the embodiment of the method shown in FIG.

According to this prediction system 10, since the coating film can be predicted from the coating conditions of the coating process, there is no need to wait for the completion of the coating process (all processes) or the baking and drying of the paint, and the already obtained coating conditions As soon as the set values (e.g., discharge amount, etc.) and the coating conditions that can be measured quickly (e.g., various temperatures and humidity, etc.) are obtained, the computer predicts the coating film quality. The evaluation result of coating film quality can be predicted quickly. Thereby, for example, when the predicted value is out of the quality standard range, the coating conditions can be quickly adjusted.
Also, as in the case of FIG. 1, if it is possible to prepare relational data that serves as teaching data for machine learning, it is possible to deal with various variations of the painting process.

<Prediction system for coating conditions>
FIG. 6 is a block diagram of a painting condition prediction system according to an embodiment of the present invention. This system is a system that predicts the optimum painting conditions for a painting process to achieve a goal when painting an automobile body and/or automobile parts through a painting process. As shown in FIG. 6 , the coating condition prediction system 20 of this 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 coating film quality prediction unit 23 , a storage unit 24 and a communication unit 25 . Since these are the same as those described in the artificial intelligence model creation unit 12, the coating film quality prediction unit 13, the storage unit 14, and the communication unit 15 in the embodiment shown in FIG. 5, detailed description is omitted. do.

The coating film quality prediction unit 23 inputs coating conditions in a plurality of coating processes, and predicts the coating film quality of the coating film of the automobile body and/or the automobile parts after the plurality of coating processes corresponding to the respective coating conditions. is configured to The computer 21 determines the optimum painting condition as the optimum painting condition to achieve the target based on the predicted paint film quality of the paint film of the automobile body and / or automobile parts after a plurality of painting processes. A coating condition prediction unit 26 is further provided. The painting condition prediction unit 26 can be a processor. The coating condition prediction unit 26 selects the coating condition corresponding to the coating film quality that minimizes the difference from the target coating film quality from among the plurality of predicted coating film qualities, as the optimum coating condition for achieving the target. It is configured so that it can be determined as a coating condition. Moreover, it is preferable that the computer 21 further has a functional unit that performs data cleansing on the relational data. In addition, the prediction system 20 preferably has a display section (display) for displaying the prediction result.
The details of the coating film quality, coating process, coating conditions, etc. are the same as in the embodiment of the method shown in FIG.

According to the painting condition prediction system of the present embodiment, since it is possible to predict the optimum painting conditions from the current painting conditions, there is no need to wait for the completion of the painting process (all processes) or for the paint to bake and dry. As soon as the setting values of the obtained coating conditions (e.g., discharge amount, etc.) and the coating conditions that can be measured quickly (e.g., various temperatures and humidity, etc.) are obtained, the optimum coating conditions are predicted by the computer. , the optimum coating conditions can be obtained quickly. Then, when the current coating conditions deviate from the predicted values, the coating conditions can be quickly adjusted. Also, the adjustment range can be easily calculated based on the predicted value (just by taking the difference between the current value and the predicted value).
As in the case of the embodiment of FIG. 3, also in the case of the embodiment of FIG. 6, if it is possible to prepare relational data that serves as teacher data for machine learning, various variations of the painting process can be handled. .

Examples of the present invention will be described below, but the present invention is not limited to the following examples.

(Example 1)
In order to confirm the effects of the present invention, a test was conducted in which the coating film quality of the coating film after the coating process on the automobile body was predicted and evaluated in comparison with the measured values, which will be described below.
As shown in Fig. 2, the coating process is such that the top coating process is performed after the electrodeposition coating process, and the top coating process consists of an intermediate coating process, a base coating process, and a clear coating process in order. bottom.

The following steps were performed according to the flow shown in FIG.
First, relational data indicating the relationship between the coating conditions in the coating process and the coating film quality of the coating film on the automobile body after the coating process was prepared and entered into a computer. Here, the coating conditions are the dry-bulb temperature of the psychrometer in the coating booth, the wet-bulb temperature of the above-mentioned psychrometer, the relative humidity in the coating booth, the viscosity of the coating, the type of coating, the amount of coating discharged during the intermediate coating process, and the base coating. The amount of paint discharged in the process and the line speed of the coating machine in the base coating process. In addition, the Lw value of the wave scan value, which is a general index, was used as an index for the coating film quality. Relational data that associates these was prepared by extracting data from past work data.
Then, by machine learning using a computer, a predetermined artificial intelligence model was created in which the data of the painting conditions in the painting process are input and the paint film quality of the paint film of the automobile body after the painting process is output. Two artificial intelligence model algorithms were created: partial least squares regression (PLS) and random forest.
Then, by inputting the painting conditions in the painting process into each artificial intelligence model, the computer calculates and predicts the paint film quality of the paint film on the car body after the painting process, and compares the predicted values with the actual values. evaluated. Actual values were measured by a wave scan ("Wave-Scan Dual" manufactured by BYK-Gardner). The range of coating conditions to be input was narrowed down in advance by referring to past data.
Calculate the MAE (Mean Absolute Error), which is the average absolute error from the actual measurement, for the numerical value of Lw of the prediction result for unknown data, and adopt the criterion that it is possible to predict accurately if the MAE is 1.0 or less. bottom. Since the MAE of Example 1 was 0.8 for PLS and 0.1 for random forest, it was found that the prediction was performed with good accuracy in both cases. In particular, it can be seen that the accuracy of prediction using random forest is higher.
The coefficient of determination ^R2 between the measured value and the predicted value of the Lw value was 0.84 for PLS and 0.99 for random forest.
Therefore, it can be seen that the evaluation result of the paint film quality of the paint film of the automobile body could be rapidly predicted.

(Example 2)
Same as Example 1 except that the Sw value of the wave scan value, which is another general index, was used as an index for the coating film quality, and that only the random forest was used for the algorithm of the artificial intelligence model. Under the conditions of , evaluation was performed by comparing the predicted value and the actual value of the coating film quality. The criteria for determination were the same as in Example 1.
Since the MAE of Example 2 was 0.2, it was found that the prediction could be made with good accuracy.
The coefficient of determination ^R2 between the measured value and predicted value of the Sw value was 0.99.
Therefore, it can be seen that the evaluation result of the paint film quality of the paint film of the automobile body could be rapidly predicted.

10: prediction system,
11: computer,
12: Artificial intelligence model creation department,
13: Paint film quality prediction unit,
14: storage unit,
15: communication unit,
20: prediction system,
21: computer,
22: Artificial intelligence model creation department,
23: Paint film quality prediction unit,
24: storage unit,
25: communication unit,
26: Painting condition prediction unit

Claims (16)

A method for predicting the coating film quality of a coating film obtained by coating an automobile body and / or automobile parts,
Using computer machine learning to create a predetermined artificial intelligence model that takes as input paint condition data in the paint process and outputs the paint film quality of the paint film on the automobile body and/or automobile parts after the paint process. , an artificial intelligence model creation process;
The computer calculates and predicts the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process by inputting the painting conditions in the painting process in the artificial intelligence model. and a film quality prediction step. A method for predicting the optimum painting conditions for a painting process to achieve a goal when painting an automobile body and/or automobile parts through a painting process,
Machine learning by computer creates a predetermined artificial intelligence model that takes as input paint condition data in the paint process and outputs the paint film quality of the paint film on the automobile body and/or automobile parts after the paint process. an artificial intelligence model creation process;
The computer calculates and predicts the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process by inputting the painting conditions in the painting process in the artificial intelligence model; and a coating film quality prediction step,
In the coating film quality prediction step, the coating conditions in a plurality of the coating processes are input, and the coating film of the auto body and/or the auto parts after the plurality of coating processes corresponding to each of the coating conditions. anticipate quality,
Based on the predicted paint film quality of the paint film of the automobile body and/or automobile parts after the plurality of the painting steps, the optimum painting condition is determined as the optimum painting condition for achieving the target. , a painting condition prediction method, further comprising a painting condition prediction step. further comprising a relationship data input step of inputting into a computer relationship data indicating the relationship between the coating conditions in the coating step and the coating film quality of the coating film of the automobile body and/or automobile component after the painting step;
3. The method according to claim 1, wherein said artificial intelligence model creation step is performed by machine learning with said computer using said relational data input in said relational data input step as teacher data. The method according to any one of claims 1 to 3, wherein the coating film quality includes any one or more of color, gloss, film thickness, sharpness, and smoothness of the coating film. The sharpness and / or smoothness uses a wave scan value as an index,
5. The method of claim 4, wherein the wavescan values are any one or more of du, Wa, Wb, Wc, Wd, We, Lw, and Sw. The method according to any one of claims 1 to 5, wherein said machine learning uses a prediction algorithm based on an ensemble tree composed of multiple decision trees. The method according to any one of claims 1 to 6, wherein said machine learning uses deep learning, partial least squares regression, random forest or gradient boosting methods. 4. The method of claim 3, further comprising performing data cleansing on the relational data after the relational data input step. In the painting condition prediction step, the difference from the target paint film quality is the smallest among the paint film qualities of the paint films of the automobile body and/or automobile parts after the plurality of the painting processes predicted. The method according to claim 2, wherein the painting conditions corresponding to the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process are determined as the optimum painting conditions for achieving the target. . 10. The method of claim 2 or 9, wherein the painting condition prediction step uses Bayesian optimization. The painting conditions include the temperature of the painting booth where the painting process is performed, the dry-bulb temperature of the psychrometer of the painting booth, the wet-bulb temperature of the psychrometer, the relative humidity of the painting booth, the absolute humidity of the painting booth, and the viscosity of the paint. , paint temperature, paint type, paint solid content, paint discharge amount in the painting process, linear speed of the painting machine, and the temperature of the object to be coated of the automobile body and / or automobile parts. The method according to any one of claims 1-10. The method according to any one of claims 1 to 11, wherein the coating step includes any one or more of a chipping primer coating step, an intermediate coating step, a base coating step, and a clear coating step. A coating condition adjustment step of adjusting the coating conditions predicted by the method according to claim 2;
and a post-adjustment coating step of performing coating under the coating conditions adjusted in the coating condition adjustment step. a measuring step of measuring the coating film quality of the coating film of the automobile body and/or automobile component after the post-adjustment painting step;
Updating relational data, inputting relational data indicating a relation between the coating conditions adjusted in the coating condition adjusting step and the coating film quality measured in the measuring step into the computer to update the relational data. 14. The method of forming a multi-layer coating film for an automobile body and/or automobile parts according to claim 13, further comprising the steps of: A paint film quality prediction system for a paint film on an automobile body and/or automobile parts after a painting process,
A predetermined artificial intelligence model having a function of machine learning, which inputs data of painting conditions in the painting process and outputs the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process. an artificial intelligence model creation unit to create;
In the artificial intelligence model, a paint film quality prediction unit that calculates and predicts the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process by inputting the painting conditions in the painting process. A system for predicting coating film quality of a coating film on an automobile body and/or automobile parts, characterized by comprising a computer comprising: A system for predicting the optimum painting conditions for a painting process to achieve a goal when painting an automobile body and/or automobile parts through a painting process,
A predetermined artificial intelligence model having a function of machine learning, which inputs data of coating conditions in the coating process and outputs the coating film quality of the coating film of the car body and/or car parts after the coating process. an artificial intelligence modeling unit that creates
In the artificial intelligence model, by inputting the painting conditions in the painting process, the paint film quality of the paint film of the automobile body and/or automobile parts after the painting process is calculated and predicted, predicting the paint film quality. and
The coating film quality prediction unit inputs the coating conditions in a plurality of the coating processes, and the coating film of the auto body and/or the auto parts after a plurality of the coating processes corresponding to each of the coating conditions. anticipate quality,
Based on the predicted paint film quality of the paint film of the automobile body and/or automobile parts after the plurality of the painting steps, the optimum painting condition is determined as the optimum painting condition for achieving the target. , a painting condition prediction system, further comprising a painting condition prediction unit.

Images