Rust determination device

Abstract

To provide a rust determination device which can efficiently determine how easily a joint surface where a plurality of members joint to each other becomes rusty.SOLUTION: A rust determination device 1 includes a determination unit 12 for determining how easily a joint surface BP becomes rusty on the basis of input information Iin including information indicating the complexity of the shape of the joint surface BP. The determination unit 12 may determine how easily the joint surface BP becomes rusty by providing the input information Iin to a trained machine learning model M. Information showing the complexity of the shape of the joint surface BP may be represented by the ratio of the area of the joint surface BP to its peripheral length.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a rust determination device.

Background Art

[0002] As a conventional technique related to rust determination, for example, Patent Document 1 describes a method for evaluating the properties of rust whose formation is predicted on the surface of a steel bridge, based on the influencing factors that affect the formation of the rust, for a steel bridge where rust formation is predicted on the surface. This method divides the surface of the steel bridge into a plurality of parts, and sets the influencing factors as a plurality of influencing factors that differ from each other in the influence on the formation of the rust. Then, for each of the plurality of parts, the plurality of influencing factors are each coefficientized as numerical values reflecting the degree of influence on the formation of the rust, and the coefficientized plurality of influencing factors are each multiplied for each of the plurality of parts. Thereby, the properties of the rust predicted to be formed on the surface of the steel bridge are evaluated for each of the plurality of parts.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Patent Document 1 above describes the rust whose formation is predicted on the surface of a steel bridge, but there is no particular mention of the rust on the joint surface at the joint of the steel bridge. On the other hand, for example, in a structure such as a vehicle, rust may occur on the joint surface where a plurality of members are joined. If it is possible to know in advance the ease of rusting of such a joint surface, it is advantageous for the design of structures such as vehicles.

[0005] The present invention has been made paying attention to the above points, and an object thereof is to provide a rust determination device capable of efficiently determining the susceptibility to rust of a joint surface where a plurality of members are joined.

Means for Solving the Problems

[0006] In order to achieve the above object, the rust determination device according to the present invention is configured to determine the susceptibility to rust of the joint surface based on input information including the complexity of the shape of the joint surface where a plurality of members are joined.

Effects of the Invention

[0007] According to the rust determination device of the present invention, it is possible to efficiently determine the susceptibility to rust of the joint surface where a plurality of members are joined.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Mode for Carrying Out the Invention

[0009] Hereinafter, the present invention will be described based on the illustrated embodiments. However, the present invention is not limited by the embodiments described below.

[0010] In the design of structures such as vehicles as described above, design requirements such as strength, rigidity, NVH (Noise, Vibration, Harshness), and formability have become such that it is possible to automatically determine pass or fail from CAD (computer-aided design) data with the maturity of CAE (computer-aided engineering) analysis technology.

[0011] However, regarding rust on the vehicle body, various factors such as materials, structures, the state of coatings and pretreatment, and the use environment are intertwined, and the rust generation mechanism is complex, so it is difficult to automatically determine from CAD data. Therefore, currently, at the data stage, a process is taken in which a knowledgeable person visually checks the CAD data of the vehicle, estimates rust-concerning locations based on past rules of thumb and knowledge, and incorporates rust prevention measures into the CAD data and drawings for the estimated locations of concern. In addition, rust evaluation is performed by repeatedly conducting a plurality of actual vehicle tests such as immersing the developed actual vehicle (prototype vehicle) in water.

[0012] However, visually checking data requires time and there is also a possibility of overlooking something. In addition, in actual vehicle tests, the test cannot be conducted until an actual vehicle is prepared, and the actual vehicle test itself also takes time. As the development schedule becomes tighter, it is required to establish a method that can accurately and comprehensively determine the areas of concern regarding rust at the data stage, and to obtain results with a small number of actual vehicle tests.

[0013] Generally, in order to ensure the rust prevention property of the vehicle body, for example, the white body of an automobile is subjected to electrodeposition coating such as cationic electrodeposition coating. The white body is in a state where the welding of the members constituting the vehicle body is completed and assembled in the manufacturing process of the automobile. The white body is unpainted and engines, seats, etc. are not installed. However, although there is a slight gap between the joint surfaces of one member and the joint surfaces of the other member at the joint of the white body by welding, it is difficult for the paint to cover the entire joint surfaces and it is difficult to ensure rust prevention. The embodiment of the present invention described below relates to a rust determination device for determining the susceptibility to rust of the joint surfaces of the vehicle body as described above.

[0014] FIG. 1 is a block diagram showing a functional configuration example of a rust determination device 1 according to an embodiment of the present invention. FIG. 2 is a perspective view schematically showing an example of a joint surface BP to be determined in the present embodiment. In FIGS. 1 and 2, the rust determination device 1 of the present embodiment is, for example, a device for determining the susceptibility to rust of a joint surface BP where a plurality of members constituting the vehicle body are joined, with a vehicle such as an automobile as an object.

[0015] The "joint surface BP" to be the object of rust determination means a surface where at least two members overlap for joining. In the example of FIG. 2, a rectangular joint surface BP is formed by joining the lower surface of another member 22 (two-dot chain line) to the upper surface of a member 21 (solid line) that constitutes a part of the vehicle body. Note that the number of members to be joined is not limited to two, and may be, for example, three-member joining, four-member joining, etc.

[0016] As shown in FIG. 1 for example, the rust determination device 1 includes an input unit 11, a determination unit 12, and a display unit 13. The input unit 11 receives input information Iin used to determine the rusting tendency of the joint surface BP to be determined. The input information Iin includes information indicating the complexity of the shape of the joint surface BP. In the present embodiment, an example will be described in which 3D CAD data and CAE analysis data of a vehicle (object) are used as the input information Iin. The details of the input information Iin will be described later.

[0017] The determination unit 12 determines the rusting tendency of the joint surface BP by processing the input information Iin received by the input unit 11 according to a predetermined algorithm, and outputs output information Iout indicating the determination result to the display unit 1. As an algorithm for determining the rusting tendency of the joint surface BP according to the input information Iin, for example, it is possible to use a learned model M in which machine learning using a learning data set has been performed in advance as described later. Here, the "rusting tendency" indicates the degree of rust generation on the joint surface BP when the joint surface BP is placed under predetermined conditions that affect rusting of the joint surface BP. For example, the degree of color change, the degree of corrosion (corrosion depth, etc.), and the degree of change in the surface shape on the joint surface BP are included in the rusting tendency. In the present embodiment, an example will be described in which the determination unit 12 uses the learned model M to determine the rusting tendency of the joint surface BP, but the present invention is not limited to this.

[0018] The display unit 13 generates an image visually showing the output information Iout from the determination unit 12, and displays the image on an image display device (monitor, display) or an image projection device (projector). Thereby, the user of the rust determination device 1 can visually grasp the rusting tendency of the joint surface BP.

[0019] FIG. 3 is a block diagram showing an example of the computer hardware configuration of the rust determination device 1. In FIG. 3, the rust determination device 1 includes a CPU 101, an interface device 102, a display device 103, an input device 104, a drive device 105, an auxiliary storage device 106, and a memory device 107, and these are interconnected by a bus 108.

[0020] The program that realizes the functions of the rust determination device 1 is provided by a recording medium 109 such as a CD-ROM. When the recording medium 109 storing the program is set in the drive device 105, the program is installed from the recording medium 109 via the drive device 105 into the auxiliary storage device 106. Alternatively, the installation of the program does not necessarily have to be performed by the recording medium 109 and can also be performed via a network. The auxiliary storage device 106 stores the installed program and also stores necessary files, data, etc.

[0021] When an instruction to start the program is given, the memory device 107 reads out and stores the program from the auxiliary storage device 106. The CPU 101 realizes the functions of the rust determination device 1 according to the program stored in the memory device 107. The interface device 102 is used as an interface for connecting to other computers through a network. The display device 103 displays a GUI (Graphical User Interface) etc. according to the program. The input device 104 is a keyboard, a mouse, etc.

[0022] Here, the determination process of the rust susceptibility of the joint surface BP executed by the determination unit 12 (FIG. 1) of the rust determination device 1 will be described in detail with reference to FIG. 4. As described above, in the present embodiment, the determination unit 12 has a learned model M obtained by machine learning as an algorithm for determining the rust susceptibility of the joint surface BP. In the determination unit 12, as shown in FIG. 4, the input information Iin received by the input unit 11 is given to the learned model M. As a result, information regarding the rust susceptibility of the joint surface BP corresponding to the input information Iin is output from the determination unit 12 as the output information Iout of the learned model M.

[0023] The input information Iin includes rust determination information necessary for the rust susceptibility determination by the determination unit 12. The rust determination information has at least one input data regarding a factor (hereinafter referred to as "rust influencing factor") that affects the rust susceptibility of the joint surface BP to be determined. The input data is extracted or calculated from the 3D CAD data and CAE analysis data of the vehicle (object), etc. The input data can be numerical data that quantifies the rust influencing factor, but is not limited thereto. For example, it is also possible to use, as the input information Iin, photographic data showing the periphery of the joint surface BP, imaging data output from an imaging unit (not shown) directly or indirectly connected to the determination unit 12, and the like.

[0024] The output information Iout includes information regarding the rust susceptibility of the joint surface BP predicted by the learned model M. The output information Iout may be, for example, output data obtained by quantifying the rust susceptibility of the joint surface BP using an evaluation index, or image data that is displayed on the above-described image display device or the like so that the user can visually recognize the rust susceptibility of the joint surface BP.

[0025] FIG. 5 is a diagram showing a specific example of the input information Iin and the output information Iout in the present embodiment. The upper part of FIG. 5 shows an example of a data table used for machine learning, and the lower part of FIG. 5 shows an example of predicting the rust susceptibility of an unknown joint surface using the learned model M obtained by performing machine learning using the data table.

[0026] In the example of Fig. 5, as input data regarding the rust influence factor of the joint surface BP, data such as the ratio of the area to the perimeter ((area) / (perimeter)) on the joint surface BP, the position of the joint surface BP (part and coordinate values), the plate thickness of the member forming the joint surface BP, the presence or absence of plating and sealing, and the susceptibility to water exposure (water exposure frequency) in the running state of the vehicle are set. Here, a plurality of items of input data are set as the rust influence factor of the joint surface BP. In particular, the value of (area) / (perimeter) of the joint surface BP is important input data for determining the rust susceptibility of the joint surface BP as information indicating the complexity of the shape of the joint surface BP, as will be described in detail later. Further, as output data obtained by quantifying the rust susceptibility of the joint surface BP by an evaluation index, for example, the corrosion depth of the joint surface BP is set.

[0027] In the machine learning of the model for predicting the rust susceptibility of the joint surface BP (upper part of Fig. 5), a plurality of input data corresponding to each of the N different joint surfaces BP1 to BP N are used as explanatory variables, and machine learning (supervised learning) is performed using teacher data (known sample data) indicating the corrosion depth of each joint surface BP1 to BP N as the target variable. In such machine learning, the target variable means the data to be predicted, and the explanatory variable means the data used to predict the target variable. The learned model M obtained by machine learning can be configured as, for example, a random forest or the like.

[0028] In the prediction of the target variable using the learned model M as described above (lower part of Fig. 5), for the joint surface BP N+1 with unknown rust susceptibility, by giving the input information Iin including the above plurality of input data to the learned model M, output data obtained by quantifying the rust susceptibility of the joint surface BP N+1 by the evaluation index (corrosion depth) is output from the determination unit 12 as the output information Iout.

[0029] Here, the ratio of the area to the perimeter ((area) / (perimeter)) on the joint surface BP, which is important input data for determining the rust susceptibility of the joint surface BP described above, will be described in detail.

[0030] The inventor of the present invention has found that, as one of the input data, the value of (area) / (perimeter) at the joint surface BP is effective for determining the susceptibility to rust. Briefly explaining the reason, the shape and area of the joint surface BP affect the susceptibility to rust and the degree of rust of the joint surface BP. The parameter of (area) / (perimeter) of the joint surface BP includes information on both the shape and area of the joint surface BP. Therefore, by using the value of (area) / (perimeter) of the joint surface BP as one of the input data, it becomes possible to effectively determine the susceptibility to rust of the joint surface BP.

[0031] Specifically, first, the influence of the "shape" of the joint surface BP on the susceptibility to rust will be described in detail. When the area of the joint surface BP is constant, the more intricate the shape (the farther the shape is from a perfect circle and the more complex it is), the longer the perimeter becomes, and (area) / (perimeter) becomes smaller. Also, the longer the perimeter, the easier it is for the paint to penetrate into the joint surface BP. Therefore, when the above-described electrodeposition coating is applied to the joint surface BP, the rust prevention performance is improved.

[0032] On the left side of FIG. 6, as a first example regarding the shape of the joint surface BP, a flat and circular joint surface BP1 in a certain member is shown. This joint surface BP1 has an area of S (unit: mm 2 ) and a perimeter of L (unit: mm). After another member is joined to this joint surface BP1 and then electrodeposition coating is performed, as shown on the right side of FIG. 6, in the joint surface BP1 after electrodeposition coating, the paint (hatched portion) penetrates to a position spaced by a distance h (unit: mm) from the outer periphery of the joint surface BP1 to the inside. Note that the black portion near the center of the joint surface BP1 on the right side of FIG. 6 illustrates the state of rust that may occur over time.

[0033] On the left side of FIG. 7, a second example regarding the shape of the joint surface BP is shown. The joint surface BP2 of the second example is flat and generally cross-shaped, and has an area approximately equal to that of the joint surface BP1. The joint surface BP2 has a more complex shape than the circular joint surface BP1.

[0034] Referring to FIGS. 8 and 9, the relationship between the shapes of the joint surfaces BP1 and BP2 will be described. FIG. 8 shows a group of circles consisting of eight circles CR having the same diameter. The center of this group coincides with the center of the joint surface BP1. The eight circles CR are all centered on the circumference of the joint surface BP1, are equally spaced along the circumferential direction of the joint surface BP1, and two adjacent circles CR in the circumferential direction of the joint surface BP1 are in contact with each other.

[0035] In each of the eight circles CR, there are a contact point PT1 with an adjacent circle CR on one side in the circumferential direction of the joint surface BP1 and a contact point PT2 with an adjacent circle CR on the other side in the circumferential direction of the joint surface BP1. Of the two arcs connecting the contact point PT1 and the contact point PT2 on the circumference of the circle CR, the arc on the radially outer side of the joint surface BP1 is indicated by the symbol AC1, and the arc on the radially inner side of the joint surface BP1 is indicated by the symbol AC2. Then, by connecting the arc AC1 of one circle CR and the arc AC2 of the other adjacent circle CR, the joint surface BP2 is formed.

[0036] As shown in FIG. 9, let the area of the overlapping region RG1 between the circle CR and the joint surface BP1 be S1 (unit: mm 2 ). Also, in the circle CR, let the area of the region RG2 that does not overlap with the joint surface BP1 be S2 (unit: mm 2 ). At this time, S1≈S2 holds. Therefore, the joint surface BP1 and the joint surface BP2 have approximately equal areas.

[0037] In Fig. 9, among the two tangent circles CR, the circle located on the counterclockwise side on the paper surface is defined as the first circle CR1, and the circle located on the clockwise side on the paper surface is defined as the second circle CR2. Among the two intersection points of the first circle CR1 and the outer periphery of the joint surface BP1, the intersection point on the counterclockwise side on the paper surface is indicated by the symbol PT3. Further, among the two intersection points of the second circle CR2 and the outer periphery of the joint surface BP1, the intersection point on the clockwise side on the paper surface is indicated by the symbol PT4. The arc formed by the intersection points PT3 and PT4 on the outer periphery of the joint surface BP1 is indicated by the symbol CL1. Let the length of the arc CL1 be L1 (unit: mm). Further, on the outer periphery of the joint surface BP2, let the length of the curve CL2 in the shape of an inverted S connecting the intersection points PT3 and PT4 be L2 (unit: mm). At this time, L1 < L2 holds. Therefore, the peripheral length of the joint surface BP2 is longer than the peripheral length of the joint surface BP1.

[0038] As described above, when comparing the joint surface BP1 and the joint surface BP2, the areas are approximately equal, and the peripheral length of the joint surface BP2 is longer than that of the joint surface BP1.

[0039] Similar to the case of the joint surface BP1 described above, when electrodeposition coating is performed after another member is joined to the joint surface BP2 shown on the left side of Fig. 7, as shown on the right side of Fig. 7, also on the joint surface BP2 after electrodeposition coating, the paint (hatched portion) has entered up to a position spaced by a distance h (unit: mm) from the outer periphery of the joint surface BP2 inward. Note that the black portion on the right side of Fig. 7 also exemplifies the state of rust that may occur over time, similar to the right side of Fig. 6 described above.

[0040] As shown in Figs. 6 and 7, in both the joint surfaces BP1 and BP2, the paint has entered up to a position spaced by a distance h (unit: mm) from the outer periphery inward. As described above, the area of the region where the paint penetrates at the joint surfaces BP1 and BP2 is approximately proportional to the perimeter of each joint surface BP1, BP2. That is, at the joint surfaces BP1 and BP2, since the perimeter of BP2 is larger than that of BP1, the area of the region where the paint penetrates is also larger for BP2. Considering that the areas of the joint surfaces BP1 and BP2 are approximately equal, the ratio of the area of the region where the paint penetrates to the total area of the joint surface is larger for the joint surface BP2. Therefore, the rust prevention effect by electrodeposition painting is greater for the joint surface BP2 than for the joint surface BP1.

[0041] When comparing the joint surface BP1 and the joint surface BP2, as the value of (area) / (perimeter), it is smaller for the joint surface BP2 than for the joint surface BP1. That is, when electrodeposition painting is applied to a joint surface BP of an arbitrary shape, as the value of (area) / (perimeter) becomes smaller, the rust prevention effect becomes greater.

[0042] Next, the influence of the "area" of the joint surface BP on the susceptibility to rust will be explained in detail. Since the area is a two-dimensional quantity and the perimeter is a one-dimensional quantity, as the area increases, the value of (area) / (perimeter) also increases. When the shapes of the joint surfaces BP are similar, when the scale becomes n times, the area becomes n 2 times and the perimeter becomes n times, so the value of (area) / (perimeter) also becomes n times. On the other hand, since the area of the region where the paint penetrates is approximately proportional to the perimeter as described above, it becomes n times. Therefore, the larger the area of the joint surface BP, the greater the proportion of the area where the paint does not penetrate. That is, the larger the area of the joint surface BP, the larger the value of (area) / (perimeter), and the smaller the rust prevention effect by electrodeposition painting.

[0043] Summarizing the influence of the "shape" and "area" of the joint surface BP described above on the susceptibility to rust, the relationships shown in the following (1) and (2) hold. (1) The closer the shape of the joint surface BP is to a perfect circle, the larger the value of (area) / (perimeter), and the smaller the rust prevention effect by electrodeposition painting. (2) The larger the area of the joint surface BP, the larger the value of (area) / (perimeter), and the smaller the rust prevention effect by electrodeposition painting.

[0044] In the above relationship, the value of (area) / (perimeter) can be said to be data that quantitatively numerically represents the complexity of the shape of the bonding surface BP. Here, "complexity" means the degree indicating the complexity of the shape of the bonding surface BP, and it is not only represented by the value of (area) / (perimeter), but may also be represented by the "circularity" described later, information on the shape of the bonding surface BP obtained from drawing data or photos, etc., information set according to the shape of the bonding surface BP, and the like. Generally, the more complex the shape (intricate shape) of the bonding surface BP is, the more boundaries in contact with the electrodeposition coating liquid increase, and it becomes easier for the coating film to adhere within the bonding surface BP. This can be exemplified by the fact that the area of the coast increases as the coastline becomes more intricate like a rias-type coast. Or, it can also be exemplified by the fact that in the small intestine, the area for absorbing nutrients is increased by increasing protrusions and folds.

[0045] If the area of the bonding surface BP is constant, the more complex the shape is, the longer the perimeter becomes, it becomes easier for the coating film to adhere within the bonding surface BP, and it is difficult to rust. This means that when the value of (area) / (perimeter) representing the complexity of the shape of the bonding surface BP becomes smaller, the rust prevention effect by electrodeposition coating becomes larger. Also, if the area of the bonding surface BP is constant, the simpler the shape is, the shorter the perimeter becomes, it becomes difficult for the coating film to adhere within the bonding surface, and it is easy to rust. This means that when the value of (area) / (perimeter) representing the complexity of the shape of the bonding surface BP becomes larger, the rust prevention effect by electrodeposition coating becomes smaller.

[0046] As described above, it can be said that the parameter of (area) / (perimeter) of the bonding surface BP is effective input data for determining (predicting) the rusting tendency of the bonding surface BP. Note that instead of (area) / (perimeter) of the bonding surface BP, for example, the diameter of the inscribed circle 23 in contact with the bonding surface BP as shown in FIG. 10 may be used as input data. Or, instead of (area) / (perimeter) of the bonding surface BP, two parameters in total, (area) / (perimeter) 2 and the area, may be used.

[0047] (area) / (perimeter) 2is a dimensionless value representing the shape of the joint surface BP, and the area represents the size of the joint surface BP. By using these two parameters, the contribution degree of the shape of the joint surface BP and the contribution degree of the size of the joint surface BP can be considered separately. A parameter replacing (area) / (perimeter) as described above also indicates the complexity of the shape of the joint surface BP.

[0048] Furthermore, it is also possible to use a combination of (area) / (perimeter) of the joint surface BP and the roundness of the joint surface BP. The roundness is a numerical value representing the complexity of a figure. With the maximum value being 1, the more complex the figure is, the smaller the numerical value becomes. The roundness can be obtained by the following calculation formula. (Roundness) = 4π×(Area)÷(Perimeter) 2 For example, the roundness of a perfect circle with radius r is as follows, 1 (the maximum value). (Roundness) = 4π×(πr 2 )÷(2πr) 2 = 1 Note that the roundness of a square is 0.785, and the roundness of an equilateral triangle is approximately 0.604. From this, it can be said that an equilateral triangle is a more complex figure than a square.

[0049] Next, the input information Iin used for determining the rust susceptibility of the joint surface BP will be specifically described. As exemplified in FIG. 5 described above, the input information Iin used for determining the rust susceptibility of the joint surface BP includes various input data, and these input data can be roughly classified into the following (A) to (E).

[0050] (A) Dimension and position information of the vehicle body member For example, the 3D CAD data of a vehicle (object) includes dimensional information and position information of each member constituting the vehicle body. Based on the dimensional information of each member, data such as the range, area, perimeter, and shape of the joint surface BP can be obtained, and from these data, a numerical value representing the complexity of the shape of the joint surface BP, such as (area) / (perimeter) of the joint surface BP described above, can be calculated. That is, the dimensional information and position information of each member constituting the vehicle body include information regarding the area of the joint surface BP and information regarding the perimeter of the joint surface BP. Also, it is possible to identify which part of the vehicle body (e.g., main floor, side sill, side member, cowl, front door, etc.) the joint surface BP corresponds to. Furthermore, based on the position information of each member, the three-dimensional coordinate values of the joint surface BP can be obtained. By considering the fact that the degree of influence of the rust influence factor on the joint surface BP changes according to the information regarding the shape and arrangement of the joint surface BP as described above, it becomes possible to improve the accuracy of the rust susceptibility determination. In addition, data on the plate thickness of each member forming the joint surface BP can also be obtained. When the joint surface BP rusts (corrodes), the shape of the joint member is distorted, and by considering the fact that this distortion method changes depending on the plate thickness of the joint member, the accuracy of the rust susceptibility determination is further enhanced. Note that the plate thickness data may be the plate thickness of one member among the plurality of members forming the joint surface BP, or the plate thickness of all members.

[0051] (B) Structural information of the joint surface BP The above 3D CAD data includes structural information such as the shape, material, joining method (welding, press joining, brazing, etc.), number of joined sheets (two-sheet joining, three-sheet joining, etc.), and joining strength of the joint surface BP, and these are useful as input data for determining (predicting) the rust susceptibility of the joint surface BP. As described above, when the joint surface BP rusts (corrodes), the shape of the joint surface BP collapses, and by considering the fact that this collapse method changes depending on the structure of the joint surface BP, it becomes possible to improve the accuracy of the rust susceptibility determination.

[0052] (C) Anti-rust treatment information of the joint surface BP The above 3D CAD data and CAE analysis data include information such as the presence or absence of rust prevention treatment applied around the joint surface BP, the type of rust prevention treatment, and the treatment amount (rust prevention film amount, coating thickness, etc.). These are also useful as input data for determining (predicting) the rust susceptibility of the joint surface BP. Here, "rust prevention treatment" includes, for example, plating (metal film), painting such as cationic electrodeposition coating or anionic electrodeposition coating (non-metal film), application of rust preventives (rust preventive oils and waxes), and sealing (sealing or covering parts prone to being wetted by water). By considering the fact that the rust susceptibility of the joint surface BP changes according to such information on rust prevention treatment, it becomes possible to improve the accuracy of determining rust susceptibility. In the case where no rust prevention treatment is applied to the joint surface BP, information on the surface state (surface properties) of the joint surface BP may also be used as input data.

[0053] (D) Information on rust occurrence on the joint surface BP The above CAE analysis data, etc. include information on the factors (salt spray, water immersion, etc.), frequency (salt spray frequency, water immersion frequency, etc.), and degree of influence (amount of salt spray, amount of water immersion, etc.) of rust occurrence on the joint surface BP in the driving state of the vehicle. These are also useful as input data for determining (predicting) the rust susceptibility of the joint surface BP. By considering the fact that the rust susceptibility of the joint surface BP changes according to such rust occurrence information, it becomes possible to improve the accuracy of determining rust susceptibility.

[0054] (E) Other information In addition to the above (A) to (D), for example, environmental information such as the vehicle's usage environment (country, region, weather, temperature, humidity) and years of use, and the angle of the joint surface BP (one end of the joint surface if the joint surface is a curved surface) with respect to the horizontal plane during vehicle use are also useful as input data for determining (predicting) the rust susceptibility of the joint surface BP. By considering the fact that the way rust influencing factors (water, salt, mud, etc.) adhere to the joint surface BP changes depending on the vehicle's usage environment and the angle of the joint surface BP, it becomes possible to improve the accuracy of determining rust susceptibility.

[0055] As described above, the input information Iin in this embodiment includes various information (input data) regarding the rust influence factors of the joint surface BP to be determined. These information can basically be classified into information (necessary information) indicating the complexity of the shape of the joint surface BP and additional information other than that. In short, the susceptibility to rust of the joint surface BP can basically be determined based on the information (necessary information) indicating the complexity of the shape of the joint surface BP. In this embodiment, by determining the susceptibility to rust of the joint surface BP based on the input information Iin that includes additional information other than the necessary information, the determination accuracy is improved.

[0056] Next, the basic concept of an algorithm (program) for determining the susceptibility to rust of the joint surface BP using the input information Iin as described above will be explained. In the rust determination device 1 of this embodiment, three-dimensional CAD data, CAE analysis data, etc. of a vehicle (object) are input into the determination unit 12 via the input unit 11, and the susceptibility to rust of the joint surface BP is determined while extracting or calculating the data necessary for the determination of the susceptibility to rust of the joint surface BP using the input information Iin.

[0057] Specifically, first, the determination unit 12 determines the presence or absence of the intrusion region of the rust preventive material on the joint surface BP to be determined based on the dimensional and positional information of the vehicle body member, the structural information of the joint surface BP, and the rust preventive treatment information. When there is an intrusion region, information (such as area, range, and position) regarding the intrusion region of the rust preventive material is determined. For example, based on the interval h that intrudes from the outer periphery to the inside of the joint surface BP1 as shown in FIG. 6 recorded in advance or input from the outside, the dimensional and positional information of the vehicle body member, and the structural information of the joint surface BP, the intrusion area, intrusion range, or intrusion position of the rust preventive material with respect to the connection surface BP1 may be determined. Then, by comparing the area, range, and position of the entire joint surface BP with the area, range, and position of the intrusion region of the rust preventive material, the presence or absence of the non-intrusion region of the rust preventive material on the joint surface BP is determined. When there is a non-intrusion region, information (such as area, range, and position) regarding the non-intrusion region of the rust preventive material is determined.

[0058] Next, the determination unit 12 determines the rust susceptibility of the intrusion area based on the rust prevention treatment information of the joint surface BP and the information regarding the intrusion area of the rust preventive material, and determines the rust susceptibility of the non-intrusion area based on the rust prevention treatment information of the joint surface BP and the information regarding the non-intrusion area of the rust preventive material. Then, according to the area ratio of the intrusion area and the non-intrusion area of the rust preventive material, the rust susceptibility of the entire joint surface BP is determined. In determining the rust susceptibility of the entire joint surface BP by a series of processes as described above, information indicating the complexity of the shape of the joint surface BP, such as the value of (area) / (perimeter) of the joint surface BP described above, is considered, thereby improving the accuracy of the determination.

[0059] Furthermore, the determination unit 12 determines the degree of rust (corrosion depth, etc.) of each of the intrusion area and the non-intrusion area of the rust preventive material according to the number of years of use, based on the rust occurrence information of the joint surface BP, the information on the usage environment of the vehicle, and the information on the number of years of use. Then, the determination unit 12 generates output information Iout regarding the rust susceptibility of the entire joint surface BP and the degree of rust of each of the intrusion area and the non-intrusion area of the rust preventive material, and outputs it to the display unit 13.

[0060] In the present embodiment, machine learning of a model for predicting the rust susceptibility of the joint surface BP is performed based on the above-described algorithm. This machine learning is advanced while correcting the algorithm based on the comparison result between the execution result of the algorithm and the teacher data after inputting a considerable amount of learning data into a well-known learning program.

[0061] Specifically, input data (explanatory variables) regarding N joint surfaces BP1 to BP as shown in the upper part of FIG. 5 described above, and teacher data (objective variables) obtained by quantifying the visually confirmed results of the rust susceptibility (corrosion depth) corresponding to each of the joint surfaces BP1 to BP N are input into a learning program corresponding to random forest or the like. Then, according to the algorithm as described above, for each of the joint surfaces BP1 to BP N N ​The degree of rusting in each of the intrusion area and non-intrusion area of the rust preventive material corresponding to the input data for [object] is determined, and based on the result of comparison with the teacher data (actual degree of rusting), correction of the algorithm is repeatedly performed. At this time, the relationship between the value of (area) / (perimeter) etc. for each of the joint surfaces BP1 to BP N is recorded with the teacher data, and by repeating this, a learned model M capable of determining the rust susceptibility of the joint surface BP with high accuracy is generated.

[0062] When the learned model M obtained by machine learning as described above is generated, for the joint surface BP with an unknown degree of rusting as shown in the lower part of FIG. 5 described above, N+1 input information Iin including the value of (area) / (perimeter) etc. is given to the learned model M for the joint surface BP N+1 and output information Iout indicating the degree of rusting (corrosion depth) of the joint surface BP numerically is output from the determination unit 12 to the display unit 13.

[0063] In the display unit 13, for example, as shown in FIG. 11, an image visually showing the output information Iout from the determination unit 12 is generated, and the image is displayed on an image display device (monitor, display), an image projection device (projector), etc. In the example of FIG. 11, for the digital mock-up (DMU) data created using the 3D CAD data of the joint member, image processing (for example, applying gradation or hatching, changing color, etc.) that allows the rust susceptibility to be visually understood is performed on the image areas corresponding to the plurality of joint surfaces BP to be determined.

[0064] Next, the effects of the rust determination device 1 according to the present embodiment will be described. In the rust determination device 1 as described above, based on the input information Iin including information indicating the complexity of the shape of the joint surface BP where a plurality of members 21 and 22 are joined, the susceptibility to rust of the joint surface BP is determined. Since the information indicating the complexity of the shape of the joint surface BP is effective data for determining the susceptibility to rust of the joint surface BP as described above, according to the rust determination device 1, the susceptibility to rust of the joint surface BP can be efficiently determined. By performing vehicle development using such a rust determination device 1, it becomes possible to determine pass / fail regarding rust at the data stage in the early stage of development, and it also becomes possible to automate the visual confirmation of the rust prevention check. As a result, the man-hours and costs of vehicle development can be significantly reduced.

[0065] Further, in the rust determination device 1 of the present embodiment, the susceptibility to rust of the joint surface BP is determined by providing the input information Iin to the learned model M that has been machine-learned. By using the learned model M in this way, the susceptibility to rust can be easily determined without using a complicated calculation formula for determining the susceptibility to rust.

[0066] Further, in the rust determination device 1 of the present embodiment, the information indicating the complexity of the shape of the joint surface BP is represented by the ratio of the area to the perimeter ((area) / (perimeter)) on the joint surface. By using the numerical value indicating the relationship between the area and the perimeter that can be easily grasped from the dimensional information of the joint surface BP as the information indicating the complexity of the shape of the joint surface BP, the susceptibility to rust of the joint surface BP can be determined with simple input information Iin.

[0067] In addition, in the rust determination device 1 of the present embodiment, when the joint surface BP is subjected to rust prevention treatment, it is determined that the joint surface BP is less likely to rust as the degree of complexity of the shape of the joint surface BP is higher. When the joint surface BP to be determined is subjected to rust prevention treatment, the more complex the shape of the joint surface BP is during the rust prevention treatment, the easier it is for the rust preventive agent to penetrate into the joint surface BP and the less likely it is to rust. On the other hand, the simpler the shape of the joint surface BP is, the more difficult it is for the rust preventive agent to penetrate into the joint surface BP and the easier it is to rust. By determining the ease of rusting in consideration of such a tendency, the determination accuracy of the ease of rusting can be improved. When the joint surface BP to be determined is not subjected to rust prevention treatment, the above tendency is reversed. That is, the more complex the shape of the joint surface BP is, the easier it is for rust influencing factors (such as water) to penetrate into the joint surface BP, and thus the easier it is to rust.

[0068] In addition, in the rust determination device 1 of the present embodiment, the input information Iin includes the plate thickness of the member forming the joint surface BP. When the joint surface BP rusts (corrodes), the shape of the member forming the joint surface BP collapses, and this collapse pattern changes depending on the plate thickness of the member. Therefore, by including the plate thickness as the input information Iin for determining the ease of rusting, the determination accuracy can be further improved.

[0069] In addition, in the rust determination device 1 of the present embodiment, the input information Iin includes information regarding the rust prevention treatment of the joint surface BP. The ease of rusting changes according to the presence or absence of rust prevention treatment, the type of rust prevention treatment, and the treatment amount (such as the amount of rust preventive film and the thickness of the coating film). Therefore, by including information regarding the rust prevention treatment as the input information Iin for determining the ease of rusting, the determination accuracy can be further enhanced.

[0070] In addition, in the rust determination device 1 of the present embodiment, the input information Iin includes the angle of the joint surface BP with respect to the reference surface during the use or test of a vehicle (object) equipped with a plurality of members 21, 22. The way in which rust influencing factors (such as water, salt, and mud) adhere to the joint surface BP changes depending on the angle of the joint surface BP. Therefore, by including the angle of the joint surface BP as the input information Iin for determining the ease of rusting, the determination accuracy can be further improved.

[0071] In addition, in the rust determination device 1 of the present embodiment, the input information Iin includes at least one of information regarding the position of the joint surface BP in a vehicle (object) including a plurality of members 21 and 22, and information regarding the type of each of the plurality of members 21 and 22. Since the degree of influence of the rust influence factor on the joint surface BP changes according to the position of the joint surface BP and the type of the joint member, by including such information in the input information Iin for determining the ease of rusting, it is possible to improve the determination accuracy.

[0072] In addition, in the rust determination device 1 of the present embodiment, the input information Iin includes information regarding the frequency or amount of occurrence of an event that affects the occurrence of rust on the joint surface BP. Thereby, the determination result of the ease of rusting can be made closer to the test result of the vehicle member, and further improvement in the rust determination accuracy can be achieved.

[0073] As described above, the embodiments of the present invention have been described. However, the present invention is not limited to the above-described embodiments, and various modifications and changes are possible based on the technical idea of the present invention. For example, in the above-described embodiment, an example in which the rust determination device 1 includes the input unit 11 and the display unit 13, and the input / output of information regarding the determination of the ease of rusting of the joint surface BP is directly performed with respect to the rust determination device 1 has been described. However, as in the modification example shown below, the input / output of information to and from the rust determination device 1 may be performed indirectly by communication or the like.

[0074] FIG. 12 is a block diagram showing the functional configuration of a rust determination system 3 according to a modification example related to the above-described embodiment. In the modification example shown in FIG. 12, a rust determination device 1' provided on a cloud server and an information processing unit 4 provided on the client side are connected to each other via a communication line 5 such as the Internet or an intranet, thereby constructing a rust determination system 3.

[0075] The rust determination device 1' receives the input information Iin sent from the information processing unit 4 on the client side via the communication line 5, determines the susceptibility to rust of the joint surface BP based on the input information Iin, and returns the output information Iout indicating the determination result to the information processing unit 4 via the communication line 5. That is, the rust determination device 1' in the modified example has a function corresponding to the determination unit 12 in the rust determination device 1 of the above-described embodiment, and functions corresponding to the input unit 11 and the display unit 13 are provided on the client side.

[0076] The information processing unit 4 on the client side corresponds to, for example, a computer installed with an application such as 3D CAD. In the example of FIG. 12, the information processing unit 4 includes an input device 41, a control device 42, and an image generation device 43. In the input device 41, various types of information (numerical values, images, etc.) are input by a keyboard, a mouse, a camera, etc., and the input information is transmitted to the control device 42.

[0077] The control device 42 generates the input information Iin used for determining the rust of the joint surface BP based on the 3D CAD data, CAE analysis data, etc. operated according to the input information from the input device 41. Then, the control device 42 transmits the generated input information Iin to the rust determination device 1' via the communication line 5 according to the required communication protocol. Further, the control device 42 receives the output information Iout indicating the susceptibility to rust of the joint surface BP determined by the rust determination device 1', and transmits the output information Iout to the image generation device 43 together with the 3D CAD data, etc.

[0078] The image generation device 43 generates and displays an image (FIG. 11) obtained by performing image processing on the DMU data of the joining member so that the susceptibility to rust of the joint surface BP can be visually understood based on the output information Iout and the 3D CAD data from the control device 42. Thereby, the user operating the information processing unit 4 on the client side can visually grasp the susceptibility to rust of the joint surface BP.

[0079] Note that the rust determination devices 1 and 1' according to the embodiments and modification examples disclosed in this specification have not only the aspect as a device, but also the aspect as a method and the aspect as a computer program.

Explanation of Reference Numerals

[0080] 1, 1'... Rust determination device 11... Input unit 12... Determination unit 13... Display unit 21, 22... Members 23... Inscribed circle 3... Rust determination system 4... Information processing unit 41... Input device 42... Control device 43... Image generation device 5... Communication line BP, BP1, BP2, BP1~BP N+1 ... Joint surface Iin... Input information Iout... Output information

Claims

1. A rust determination device configured to determine the susceptibility of a joint surface to rust based on information indicating the complexity of the shape of the joint surface where a plurality of members are joined.

2. The rust determination device according to claim 1, wherein the information indicating the complexity of the shape of the joint surface is represented by the ratio of the area to the perimeter of the joint surface.

3. The rust determination device according to claim 1, wherein the information indicating the complexity of the shape of the joint surface is determined based on information regarding the area of the joint surface and information regarding the perimeter of the joint surface.

4. The rust determination device according to claim 1, configured to determine that the joint surface is less likely to rust as the degree of complexity of the shape of the joint surface is higher when the joint surface is rust-proof treated.

5. The rust determination device according to claim 1, configured to determine the susceptibility of the joint surface to rust based on input information including information indicating the complexity of the shape of the joint surface.

6. The rust determination device according to claim 5, configured to determine the susceptibility of the joint surface to rust by providing the input information to a trained model that has undergone machine learning.

7. The rust determination device according to claim 5, wherein the input information includes the plate thickness of the member forming the joint surface.

8. The rust determination device according to claim 5, wherein the input information includes information regarding the rust-proof treatment of the joint surface.

9. The rust determination device according to claim 5, wherein the input information includes the angle of the joint surface with respect to a reference surface during use or testing of the object provided with the plurality of members.

10. The rust determination device according to claim 5, wherein the input information includes at least one of information regarding the position of the joint surface in the object provided with the plurality of members and information regarding the type of each of the plurality of members.

11. The plurality of members are members used in a vehicle, The rust determination device according to claim 5, wherein the input information includes information regarding the frequency or amount of occurrence of an event that affects the occurrence of rust on the joint surface.

12. An image generation device that generates an image visually showing information regarding the susceptibility to rust determined by the rust determination device according to claim 1.

13. An information processing unit that receives and processes information regarding the susceptibility to rust determined by the rust determination device according to claim 1.

14. A program for causing a computer to function as the rust determination device according to claim 1.

Images