Covered conductive wire and multi-core cable

Abstract

A covered conductive wire according to one embodiment of the present disclosure comprises a conductor part and a covering part. The covering part covers the conductor part. The conductor part is configured by twisting together a plurality of element wires. The conductor part includes a conductor. The conductor is adjacent to the covering part. The direction in which the conductor part extends is defined as the Z direction. In the conductor, the half-value width is less than or equal to 0.6 times the median value. The half-value width and the median value are calculated from a histogram of the twist pitch of the conductor in the Z direction. Using two cross-sections that are perpendicular to the Z direction and separated from each other in the Z direction, the twist pitch is calculated on the basis of the displacement of the relative positions of the element wires between the two cross-sections in a relative coordinate system in which the center of gravity of the conductor in the cross-sections is treated as the origin. The distance between the two cross-sections is 1 mm. The histogram is created for a section of the conductor, said section having a length of 50 mm in the Z direction.

Description

Insulated conductors and multi-core cables

[0001] This disclosure relates to insulated conductors and multi-core cables. This application claims priority under the international application (PCT / JP2024 / 043807) filed on 11 December 2024. All provisions contained herein are incorporated herein by reference.

[0002] Conventionally, multi-core cables are known that comprise a conductor made of multiple strands twisted together and a covering portion that covers the conductor (see, for example, International Publication No. 2017 / 056278). Conductors made of multiple strands twisted together are manufactured, for example, using a manufacturing apparatus that uses a bobbin fixing method or a manufacturing apparatus that uses a bobbin rotation method (see, for example, Japanese Patent Publication No. 2006-328603 and Japanese Patent Publication No. Hei 4-075733).

[0003] International Publication No. 2017 / 056278, Japanese Patent Publication No. 2006-328603, Japanese Patent Publication No. Hei 4-075733

[0004] The insulated conductor of this disclosure comprises a conductor portion and an insulated portion. The insulated portion covers the conductor portion. The conductor portion is composed of multiple strands twisted together. The conductor portion includes a conductor. The conductor is adjacent to the insulated portion. The direction in which the conductor portion extends is defined as the Z direction. In the conductor, the width at half maximum is 0.6 times or less the median. The width at half maximum and the median are calculated from a histogram of the twist pitch in the Z direction of the conductor. When considering two cross-sections perpendicular to the Z direction and separated from each other in the Z direction, the twist pitch is calculated based on the displacement of the relative positions of the strands between the two cross-sections in a relative coordinate system with the centroid of the conductor in the cross-sections as the origin. The distance between the two cross-sections is 1 mm. The histogram is created for a portion of the conductor with a length of 50 mm in the Z direction.

[0005] Figure 1 is a cross-sectional view of a multi-core cable according to this embodiment. Figure 2 is a cross-sectional view of the insulated conductor in the multi-core cable according to this embodiment. Figure 3 is a cross-sectional view of the insulated conductor in the multi-core cable according to this embodiment. Figure 4 is a cross-sectional view of the conductor portion in the multi-core cable according to this embodiment. Figure 5 is a schematic cross-sectional view of the conductor portion illustrating the method for calculating the stranding pitch of the strands. Figure 6 is a schematic distribution diagram illustrating the histogram of the stranding pitch. Figure 7A is a schematic diagram showing an apparatus for manufacturing a conductor formed by twisting together multiple strands of a multi-core cable according to this embodiment. Figure 7B is a schematic diagram showing an apparatus for manufacturing a multi-core cable according to this embodiment. Figure 8 is a diagram showing the configuration of the multi-core cable quality evaluation system 40. Figure 9 is a diagram showing the configuration of the learning device 60. Figure 10 is a diagram showing an example of a convolutional neural network according to this embodiment. Figure 11 is a diagram showing the configuration of the processed image generation device 50. Figure 12A is a diagram showing a tomographic image input to the neural network. Figure 12B is an enlarged view of the area represented by the rectangle in Figure 12A. Figure 13A shows the processed image when the filter size is {f1=5, f2=1, f3=5}. Figure 13B is a magnified view of the area represented by the rectangle in Figure 13A. Figure 14A shows the processed image when the filter size is {f1=10, f2=1, f3=5}. Figure 14B is a magnified view of the area represented by the rectangle in Figure 14A. Figure 15A shows the processed image when the filter size is {f1=15, f2=1, f3=5}. Figure 15B is a magnified view of the area represented by the rectangle in Figure 15A. Figure 16A shows the processed image when the filter size is {f1=25, f2=1, f3=5}. Figure 16B is a magnified view of the area represented by the rectangle in Figure 16A. Figure 17A shows the processed image when the filter size is {f1=35, f2=1, f3=5}. Figure 17B is an enlarged view of the area represented by the rectangle in Figure 17A. Figure 18A shows the processed image when the filter size is {f1=45, f2=1, f3=5}. Figure 18B is an enlarged view of the area represented by the rectangle in Figure 18A. Figure 19A shows the processed image when the filter size is {f1=55, f2=1, f3=5}.Figure 19B is an enlarged view of the area represented by the rectangle in Figure 19A. Figure 20A shows the processed image when the filter size is {f1=35, f2=1, f3=15}. Figure 20B is an enlarged view of the area represented by the rectangle in Figure 20A. Figure 21A shows the processed image when the filter size is {f1=35, f2=5, f3=15}. Figure 21B is an enlarged view of the area represented by the rectangle in Figure 21A. Figure 22A shows the processed image when the filter size is {f1=35, f2=15, f3=15}. Figure 22B is an enlarged view of the area represented by the rectangle in Figure 22A. Figure 23A shows the processed image when the filter size is {f1=35, f2=1, f3=5}. Figure 23B is an enlarged view of the area represented by the rectangle in Figure 23A. Figure 24A shows the processed image when the filter size is {f1=35, f2=1, f3=10}. Figure 24B is a magnified view of the area represented by the rectangle in Figure 24A. Figure 25A shows the processed image when the filter size is {f1=35, f2=1, f3=15}. Figure 25B is a magnified view of the area represented by the rectangle in Figure 25A. Figure 26A shows the processed image when the filter size is {f1=35, f2=1, f3=25}. Figure 26B is a magnified view of the area represented by the rectangle in Figure 26A. Figure 27A shows the processed image when the filter size is {f1=35, f2=1, f3=45}. Figure 27B is a magnified view of the area represented by the rectangle in Figure 27A. Figure 28A shows a processed image when the filter sizes are {f1=35, f2=1, f3=55}. Figure 28B is an enlarged view of the area represented by the rectangle in Figure 28A. Figure 29 shows appropriate and optimal values ​​for the number and size of filters in the convolutional neural network of Figure 10. Figure 30 shows the configuration of the center position detection device 70. Figure 31 shows the configuration of the index calculation device 80. Figure 32 shows an example of connecting the center positions of the cross-sections of the strands. Figure 33 shows an example of trajectory data for each strand. Figure 34 is a diagram to explain the first twist pitch. Figure 35 is a diagram to explain the second twist pitch. Figure 36 is a diagram to explain the third twist pitch.Figure 37 is a diagram illustrating the fourth twist pitch. Figure 38 is a flowchart showing the operation procedure during learning. Figure 39 is a flowchart showing the quality evaluation procedure for multi-core cables. Figure 40 is a flowchart showing the center position detection procedure for step S202. Figure 41 is a diagram showing the configuration when the functions of the multi-core cable quality evaluation system 40 are realized using software. Figure 42 is a schematic diagram showing an apparatus for manufacturing a conductor formed by twisting together multiple strands of multi-core cables related to samples 9 and 10.

[0006] Conventionally, in multi-core cables connected to sensors used in anti-lock braking systems (ABS), electric parking brakes (EPB), or electromechanical brakes (EMB), there has been a need to suppress the frequency of sparks.

[0007] This disclosure was made to solve the problems described above. More specifically, it provides insulated conductors and multi-core cables with reduced sparking frequency.

[0008] According to this disclosure, it is possible to obtain insulated conductors and multi-core cables in which the frequency of spark generation is suppressed.

[0009] First, the embodiments of this disclosure will be listed and described.

[0010] (1) An image generation method according to one aspect of the present disclosure comprises the steps of: acquiring a tomographic image of a cross-section perpendicular to the longitudinal direction of an object containing a plurality of round rod-shaped materials or a plurality of round rod-shaped hollows; and generating an image from the acquired tomographic image using a trained model that generates an image in which the pixels at the center of the cross-section of the material or hollow in the tomographic image are bright, and the pixels gradually darken toward the radial direction of the cross-section of the material or hollow, thereby generating an image in which the pixels at the center of the cross-section of the material or hollow in the acquired tomographic image are bright, and the pixels gradually darken toward the radial direction of the cross-section of the material or hollow.

[0011] (2) In the image generation method described in (1) above, the trained model is composed of a convolutional neural network having three convolutional layers. A trained model can be generated by a convolutional neural network having three convolutional layers that have high learning ability and are efficient.

[0012] (3) In the image generation method described in (2) above, the size of one side of the filter of the first convolutional layer (kernel size) is the number of pixels that corresponds to a size of 0.88 times or more and 2.64 times or less the diameter of the cross-section of the material or hollow surface. This allows the processing capacity of the first convolutional layer to be appropriately set.

[0013] (4) In the image generation method described in (2) above, the size of one side of the filter of the third convolutional layer is the number of pixels corresponding to a size of 0.59 times or more and 1.47 times or less the diameter of the cross-section of the material or hollow surface. This allows the processing capacity of the third convolutional layer to be appropriately set.

[0014] (5) A method for evaluating the quality of a multicore cable according to one aspect of the present disclosure comprises the steps of generating a tomographic image and generating an image as described in any one of (1) to (4) above, wherein the step of acquiring a tomographic image includes acquiring a tomographic image of a cross-section perpendicular to the longitudinal direction of a multicore cable including a plurality of strands, and the step of generating an image includes, for each of the tomographic images of the plurality of cross-sections of the multicore cable, generating an image such that the pixel at the center of the cross-section of the strand in the tomographic image is bright, and the pixels gradually become darker toward the radial direction of the cross-section of the strand. The method for evaluating the quality of a multicore cable further comprises the steps of detecting the center position of the cross-section of a strand from the generated image, determining the trajectory of a strand by connecting the center positions of the cross-sections of strands in a plurality of cross-sections, and calculating an index representing the twist pitch of the strands based on the trajectory of the strands, or calculating an index representing the twist pitch of an assembly of a plurality of strands twisted together based on the trajectories of a plurality of strands. This allows for the generation of images that easily separate multiple strands into a single strand when they are in contact, and enables the accurate calculation of the twist pitch using the center position of the strands.

[0015] (6) In the multi-core cable quality evaluation method described in (5) above, the step of detecting the center position includes the step of binarizing the obtained image. This makes it possible to appropriately separate an image that is prone to being divided into individual strands into individual strands.

[0016] (7) The multi-core cable quality evaluation method described in (5) above further comprises the step of calculating a statistical quantity of an index representing the twist pitch. The quality of the multi-core cable can be evaluated using the statistical quantity representing the twist pitch.

[0017] (8) In the multi-core cable quality evaluation method described in (7) above, the statistical quantity of the index representing the twist pitch is the half-width of the frequency calculated from the histogram of the index representing the twist pitch. The quality of the multi-core cable can be evaluated according to whether there is a large variation in the twist pitch, using the statistical quantity of the index.

[0018] (9) A covered conductor according to one aspect of the present disclosure comprises a conductor portion and a covering portion. The covering portion covers the conductor portion. The conductor portion is composed of a plurality of strands twisted together. The conductor portion includes a conductor. The conductor is adjacent to the covering portion. The direction in which the conductor portion extends is defined as the Z direction. In the conductor, the width at half maximum is 0.6 times or less the median. The width at half maximum and the median are calculated from a histogram of the twist pitch in the Z direction of the conductor. When two cross-sections perpendicular to the Z direction and separated from each other in the Z direction are considered, the twist pitch is calculated based on the displacement of the relative positions of the strands between the two cross-sections in a relative coordinate system with the centroid of the conductor in the cross-sections as the origin. The distance between the two cross-sections is 1 mm. The histogram is created for a portion of the conductor with a length of 50 mm in the Z direction.

[0019] (10) In the case of the insulated conductor described in (9) above, if the median is 26.0 mm or more, the half-width may be 15.5 mm or less.

[0020] (11) In the covered conductor described in (9) or (10) above, the histogram may be created from the amount of relative positional displacement of each of the multiple strands.

[0021] (12) In the insulated conductors of (9) to (11) above, the conductor may include a first outer conductor and a second outer conductor. The histogram may be created from the twist pitch of the first outer conductor and the second outer conductor, respectively.

[0022] (13) In the insulated wires described in (9) to (12) above, the conductor portion may include an inner conductor. The inner conductor may be arranged separately from the insulated portion.

[0023] (14) A multi-core cable according to one aspect of the present disclosure may include the insulated conductor described in any of (9) to (13) above.

[0024] Next, the details of the embodiments of this disclosure will be described with reference to the drawings. In the following drawings, identical or corresponding parts will be given the same reference numerals, and their descriptions will not be repeated.

[0025] <Configuration of a Multi-Core Cable> Figure 1 is a cross-sectional view of a multi-core cable 1 according to this embodiment. As shown in Figure 1, the multi-core cable 1 according to the embodiment of this disclosure is a multi-core cable 1 connected to, for example, a sensor used in the ABS of a vehicle, an electric parking brake, or an electromechanical brake (EMB), and mainly comprises a plurality of insulated conductors 2 and a sheath layer 30. The insulated conductors 2 are surrounded by the sheath layer 30. The plurality of insulated conductors 2 are composed of two insulated conductors 2a and two insulated conductors 2b. The number of insulated conductors 2 is not particularly limited. The multi-core cable 1 only needs to have at least one of the insulated conductors 2a and the insulated conductors 2b. The number of insulated conductors 2a may be three or more, or it may be one. The number of insulated conductors 2b may be three or more, or it may be one. The number of insulated conductors 2b may be zero.

[0026] Here, the direction in which the multi-core cable 1 extends is defined as the Z direction (the Z direction shown in Figure 1). The directions perpendicular to the Z direction are defined as the X and Y directions. The X direction is perpendicular to the Y direction. The cross-section shown in Figure 1 is a plane perpendicular to the Z direction.

[0027] <Sheath Layer> The sheath layer 30 may have a two-layer structure. Specifically, the sheath layer 30 includes an outer sheath layer 31 and an inner sheath layer 32. The two insulated conductors 2a and the two insulated conductors 2b are surrounded by the outer sheath layer 31. The inner sheath layer 32 is located inside the outer sheath layer 31. Specifically, the inner sheath layer 32 covers the outer circumference of each of the two insulated conductors 2a and the two insulated conductors 2b so as to fill the space between the outer sheath layer 31 and the insulated conductors 2.

[0028] The main component of the inner sheath layer 32 can be any flexible synthetic resin. The material constituting the inner sheath layer 32 may be, for example, polyethylene or polyolefin such as EVA (Ethylene-Vinyl Acetate), polyurethane elastomer, or polyester elastomer.

[0029] The main component of the outer sheath layer 31 may be a synthetic resin with excellent flame retardancy and abrasion resistance. The material constituting the outer sheath layer 31 may be, for example, polyurethane.

[0030] As shown in Figure 1, in the cross-section of the multi-core cable 1 in the Z direction, the shape of the outer sheath layer 31 is annular. The inner sheath layer 32 is connected to the inner circumferential surface of the outer sheath layer 31. In other words, the shape of the outer circumference 32a of the inner sheath layer 32 is circular.

[0031] The lower limit of the shortest distance from the outer circumference 32a of the inner sheath layer 32 to the insulated conductor 2 may be 0.3 mm or 0.4 mm. On the other hand, the upper limit of the shortest distance from the outer circumference 32a of the inner sheath layer 32 to the insulated conductor 2 may be 0.9 mm or 0.8 mm. Also, the lower limit of the outer diameter of the inner sheath layer 32 may be 6.0 mm or 7.3 mm. On the other hand, the upper limit of the outer diameter of the inner sheath layer 32 may be 10 mm or 9.3 mm.

[0032] The thickness of the outer sheath layer 31 may be 0.3 mm or more and 0.7 mm or less. Furthermore, a tape material such as paper may be wrapped around the sheath layer 30 and the insulated conductor 2a as a winding restraint.

[0033] <Insulated Conductors> Figure 2 is a cross-sectional view of the insulated conductor 2a in the multi-core cable 1 according to this embodiment. Figure 3 is a cross-sectional view of the insulated conductor 2b in the multi-core cable 1 according to this embodiment.

[0034] As shown in Figure 2, the insulated conductor 2a includes a conductor portion 4 and an insulated portion 22. The insulated portion 22 covers the conductor portion 4. The conductor portion 4 is composed of multiple strands 5 twisted together. The conductor portion 4 only needs to include a conductor 21 (outer conductor 21a) adjacent to the insulated portion 22, and any configuration can be adopted.

[0035] Specifically, as shown in Figure 2, the conductor portion 4 includes an outer conductor 21a and an inner conductor 21b. The outer conductor 21a is adjacent to the covering portion 22. The inner conductor 21b is not adjacent to the covering portion 22. In other words, the inner conductor 21b is positioned away from the covering portion 22.

[0036] The outer conductor 21a includes a first outer conductor 21a1, a second outer conductor 21a2, a third outer conductor 21a3, a fourth outer conductor 21a4, a fifth outer conductor 21a5, and a sixth outer conductor 21a6. The outer conductors 21a are arranged on the outer periphery of the inner conductor 21b, with the inner conductor 21b at the center. In other words, the first outer conductor 21a1, the second outer conductor 21a2, the third outer conductor 21a3, the fourth outer conductor 21a4, the fifth outer conductor 21a5, and the sixth outer conductor 21a6 are arranged so as to surround the inner conductor 21b, with the arrangement being circular. Thus, the insulated conductor 2a may include multiple outer conductors 21a and inner conductors 21b. On the other hand, as shown in Figure 3, in the insulated conductor 2b, the conductor portion 4 may consist of only one conductor 21.

[0037] The covering portion 22 covers the conductor portion 4 by being laminated on the outer circumference of the conductor portion 4. As shown in Figures 2 and 3, the cross-sectional shape of the covering portion 22 is annular. The thickness of the covering portion 22 is not particularly limited, but for example, it is 0.1 mm or more and 5 mm or less. The thickness of the covering portion 22 may be 1.0 mm or less, or 0.4 mm or less.

[0038] The coating portion 22 only needs to be made of an insulating material. Therefore, the coating portion 22 may be formed from a composition mainly composed of synthetic resin. The main component of the coating portion 22 is a polyethylene resin. Examples of polyethylene resins include high-density polyethylene, low-density polyethylene, linear low-density polyethylene, and ethylene-α-olefin copolymer. Examples of polyethylene resins such as ethylene-α-olefin copolymer include ethylene-vinyl acetate copolymer (EVA), ethylene-ethyl acrylate copolymer (EEA), ethylene-methyl acrylate copolymer (EMA), and ethylene-butyl acrylate copolymer (EBA). Among these, low-density polyethylene and linear low-density polyethylene may be used as polyethylene resins. One or more of these polyethylene resins can be used. When two or more polyethylene resins are mixed and used, it is sufficient that the entirety of the two or more polyethylene resins constitutes the main component of the coating portion 22. The main component of the coating portion 22 is not particularly limited as long as it has insulating properties, but from the viewpoint of flexibility, it may be a copolymer with ethylene-α-olefin. The coating portion 22 may contain additives such as flame retardants, flame retardant enhancers, antioxidants, lubricants, colorants, reflective agents, opacities, processing stabilizers, and plasticizers. Furthermore, the coating portion 22 may contain resins other than the main component resin mentioned above.

[0039] The lower limit of the product C × E (unit: MPa / K) of the coefficient of linear expansion C of the coating portion 22 from 25°C to -35°C and the modulus of elasticity E at -35°C is 0.01 MPa / K. On the other hand, the upper limit of the product C × E is 0.9 MPa / K. The upper limit of the product C × E may also be 0.7 MPa / K or 0.6 MPa / K. If the above product C × E is smaller than the above lower limit, the mechanical properties such as the strength of the coating portion 22 may be insufficient. On the other hand, if the above product C × E exceeds the above upper limit, the coating portion 22 will be less likely to deform at low temperatures. Therefore, the bending resistance of the coated conductor 2 at low temperatures may decrease. Note that the product C × E can be adjusted by the content of α-olefin, the content ratio of the main component resin, etc.

[0040] The lower limit of the coefficient of linear expansion C (unit: / K) of the coated portion 22 from 25°C to -35°C is 1 × 10 －５ It may also be / K, 1 × 10 －４ It may also be / K. On the other hand, the upper limit of the coefficient of linear expansion C of the covering portion 22 is 2.5 × 10 －４ It may also be / K, 2 × 10 －４ It may also be / K. If the coefficient of linear expansion C of the coating portion 22 is smaller than the above lower limit, the mechanical properties such as the strength of the coating portion 22 may be insufficient. On the other hand, if the coefficient of linear expansion C of the coating portion 22 exceeds the above upper limit, the coating portion 22 will be less likely to deform at low temperatures. Therefore, the bending resistance of the coated conductor 2a at low temperatures may decrease.

[0041] The lower limit of the elastic modulus E (unit: MPa) of the coating portion 22 at -35°C may be 1000 MPa or 2000 MPa. On the other hand, the upper limit of the elastic modulus E of the coating portion 22 may be 3500 MPa or 3000 MPa. If the elastic modulus E of the coating portion 22 is smaller than the lower limit, the mechanical properties such as the strength of the coating portion 22 may be insufficient. On the other hand, if the elastic modulus E of the coating portion 22 exceeds the upper limit, the coating portion 22 will be less likely to deform at low temperatures, which may reduce the bending resistance of the coated conductor 2a at low temperatures.

[0042] The coefficient of linear expansion C is the linear expansion coefficient measured in accordance with the test method for dynamic mechanical properties described in JIS-K7244-4 (1999). The coefficient of linear expansion C is a value calculated from the dimensional change of a thin plate in response to temperature changes using a viscoelasticity measuring device (for example, "DVA-220" manufactured by IT Measurement Control Co., Ltd.). The modulus of elasticity E is a value measured in accordance with the test method for dynamic mechanical properties described in JIS-K7244-4 (1999). The modulus of elasticity E is the value of the storage modulus of elasticity measured using a viscoelasticity measuring device (for example, "DVA-220" manufactured by IT Measurement Control Co., Ltd.).

[0043] Figure 4 is a cross-sectional view of the conductor portion 4 in the multi-core cable 1 according to this embodiment. As shown in Figure 4, the conductor portion 4 is constructed by twisting together a plurality of strands 5 at a constant pitch (twist pitch P). The strands 5 only need to be electrically conductive. Therefore, the material constituting the strands 5 is not particularly limited, but may be, for example, copper wire, copper alloy wire, tin-plated soft copper wire, aluminum wire, or aluminum alloy wire.

[0044] In this embodiment, one conductor section 4 contains 72 strands 5. The insulated conductor 2a includes six outer conductors 21a and one inner conductor 21b. Therefore, the insulated conductor 2a contains 504 strands 5. The number of strands 5 may be appropriately changed according to the application of the multi-core cable 1 and the diameter of the strands 5. The number of strands 5 included in one insulated conductor 2 may be 196 or 294 as a lower limit. On the other hand, the number of strands 5 included in one insulated conductor 2 may be 2450 or 2000 as an upper limit.

[0045] The diameter of the wire strand 5 may be 40 μm, 50 μm, or 60 μm as a lower limit. On the other hand, the diameter of the wire strand 5 may be 200 μm or 160 μm as an upper limit.

[0046] In the cross-section of the conductor portion 4 in a direction perpendicular to the Z direction, the lower limit of the area ratio of the gap region between the multiple strands 5 is 5%, but may also be 6% or 8%. On the other hand, the upper limit of the area ratio of the gap region is 20%, but may also be 19% or 18%. If the area ratio of the gap region is smaller than the lower limit, large bending stresses are more likely to be locally applied to the strands 5 when the multi-core cable 1 is bent. As a result, the bending resistance of the multi-core cable 1 may decrease. On the other hand, if the area ratio of the gap region exceeds the upper limit, the extrusion moldability of the covering portion 22 decreases. As a result, the roundness of the covered conductor 2 and the adhesion between the covering portion 22 and the conductor portion 4 may decrease. In other words, when the conductor portion 4 is exposed at the end of the covered conductor 2, the conductor portion 4 becomes more likely to move relative to the covering portion 22. As a result, the processability of the terminal may decrease. In addition, the covered conductor 2 may become more easily deformed, and water may easily penetrate it.

[0047] The area occupied by the gaps between the strands 5 is calculated by using a photograph of the cross-section of the insulated conductor 2, which includes the conductor portion 4 and the covering portion 22 covering the conductor portion 4. The result is the area enclosed by the covering portion 22 (the cross-sectional area of ​​the conductor portion 4, including the gap between the covering portion 22 and the conductor portion 4, and the gaps between the strands 5), minus the sum of the cross-sectional areas of the strands 5. Specifically, the area occupied by the gaps can be determined by image processing, for example, by binarizing the grayscale of the cross-sectional photograph into strand portions and gap portions, and then calculating the area of ​​the gap portions. This image processing can be performed, for example, by using software such as "Paint Shop Pro" to create a two-tone image, visually confirming that the boundaries of the strands 5 are correctly distinguishable and setting a threshold, and then calculating the area ratio of each binarized region using a histogram.

[0048] Here, we will explain how to calculate the twist pitch P. Figure 5 is a schematic cross-sectional view of the conductor section illustrating the method for calculating the twist pitch P of the strands 5. Here, we will explain how to calculate the twist pitch P of any strand 5 included in the conductor section 4. First, the quality evaluation system 40 of the multi-core cable 1, which will be described later, identifies the center positions O01A to O72A for each of the multiple strands 5 in cross-section A at an arbitrary position in the Z direction. The coordinates of the center positions O01A to O72A of the strands 5 in cross-section A are expressed as relative positions in the relative coordinate system VA, with the center position OA of the conductor section 4 as the origin in cross-section A. The center position OA of the conductor section 4 is determined from the center positions O01A to O72A of each of the multiple strands 5 included in the conductor section 4. Specifically, the center positions O01A to O72A of each of the multiple strands 5 are considered to be the centroid positions of each of the multiple strands 5 in cross-section A. Assuming that each of the multiple strands 5 has the same cross-sectional area, the centroid position of the entire set of strands 5 is determined based on the centroid positions of the multiple strands 5, and this is taken as the centroid position of the conductor section 4. This centroid position of the conductor section 4 in cross-section A is considered to be the center position OA of the conductor section 4 in cross-section A. Using a relative coordinate system VA with the center position OA of the conductor section 4 thus identified as the origin, the coordinates of the center position of each of the multiple strands 5 are determined. In this way, the coordinates of the center position of each of the multiple strands 5 in a cross-section at an arbitrary position in the Z direction are determined.

[0049] Next, the center positions O01B to O72B of the multiple strands 5 in cross-section B are determined in the same manner as the method for determining the center position of the strand 5 in cross-section A. Cross-section B is a cross-section located adjacent to (i.e., different from) cross-section A in the Z direction. The coordinates of the center positions O01B to O72B of the strands 5 in cross-section B are expressed as relative positions in a relative coordinate system VB in cross-section B, with the center position OB of the conductor section 4 as the origin. The center position OB of the conductor section 4 is determined from the respective center positions O01B to O72B of the multiple strands 5 included in the conductor section 4, as in the case of cross-section A.

[0050] In Figure 5, assuming that the relative coordinate system VB has been translated so that its origin (center position OB) coincides with the origin (center position OA) of the relative coordinate system VA, the center position O01A of one strand 5 in section A and the center position O01B of one strand 5 in section B are shown. In other words, in Figure 5, the center position O01A of the strand 5 in section A and the center position O01B of the strand 5 in section B are virtually shown on the same plane.

[0051] As shown in Figure 5, the circumferential displacement dθ1 of the strand 5 is determined with respect to the center position OA (OB) of the conductor portion 4. The twist pitch P of the strand 5 is calculated as P = l × 360 / dθ, where l is the distance between cross-sections A and B in the Z direction. In this way, when considering two cross-sections that are separated from each other in the Z direction, the twist pitch P for the portion of the strand 5 located between the two cross-sections is calculated based on the displacement dθ1 of the relative position of the strand 5 between the two cross-sections in a relative coordinate system with the centroid of the conductor portion 4 in the cross-section as the origin. In other words, by setting multiple cross-sections at short intervals in the Z direction for a single strand 5, the twist pitch P (local twist pitch P) for each portion of the strand 5 located between two adjacent cross-sections can be determined.

[0052] Next, we will explain the histogram of the twist pitch P in the Z direction. Figure 6 is a schematic distribution diagram illustrating the histogram of the twist pitch P of the conductor section 4. In Figure 6, the horizontal axis represents the twist pitch P (unit: mm) as described above, and the vertical axis represents the frequency (unit: pieces). The order range of the twist pitch P may be set, for example, at 1 mm intervals. Alternatively, the lower limit of the order range of the twist pitch P may be set as the class value of the twist pitch P. Specifically, in the order range of the twist pitch P between 0 mm and less than 1 mm, the class value of the twist pitch P may be set to 0 mm. In the order range of the twist pitch P between 1 mm and less than 2 mm, the class value of the twist pitch P may be set to 1 mm. In the order range of the twist pitch P between 2 mm and less than 3 mm, the class value of the twist pitch P may be set to 2 mm. The frequency depends on the number of data points of the twist pitch P calculated later. As shown in Figure 6, F1 represents the frequency at the mode of the twist pitch P. F2 is the frequency that corresponds to half the value of frequency F1. The twist pitch P corresponding to frequency F2 may also be calculated by interpolating the frequencies at the class values ​​of two adjacent twist pitches P. For example, if frequency F2 corresponds to 50, the frequency at a class value of 1 mm is 40, and the frequency at a class value of 2 mm is 70, then the twist pitch P corresponding to frequency F2 of 50 is calculated to be 1.3 mm (rounded to the second decimal place). There are two twist pitches P corresponding to frequency F2 calculated in this way. The half-width W is the difference between these two twist pitches P. The half-width W is calculated in this way.

[0053] The number of data points for creating a histogram of the twist pitch P is determined based on the distance between the two cross-sections and the length of the conductor 21 to be measured in the Z direction.

[0054] When the distance l between the two cross-sections is 1 mm, the histogram may be calculated for, for example, a portion of the conductor 4 with a length of 50 mm in the Z direction. In this case, the number of data points for the twist pitch P in one strand 5 is 50. To expand the measurement area of ​​the strand 5 in the Z direction, the length of the conductor 4 in the Z direction may be increased. The histogram may be calculated for a portion of the conductor 4 with a length of 100 mm in the Z direction. In this case, the number of data points for the twist pitch P in one strand 5 is 100. In this way, by increasing the number of data points, it is possible to create a histogram that more accurately reflects the variation in the twist pitch P in the Z direction.

[0055] The distance l between the two cross-sections may be reduced. For example, when the length of the conductor portion 4 in the Z direction is 50 mm, the distance l between the two cross-sections may be 0.1 mm. In this case, the number of data points for the twist pitch P in one strand 5 is 500. By increasing the number of data points in this way, it is possible to create a histogram that reflects the local variation of the twist pitch P in the Z direction with higher accuracy. In other words, the twist pitch P of the finer parts of the conductor portion 4 in the Z direction is reflected in the histogram. In this way, the distance l between the two cross-sections and the length of the conductor portion 4 in the Z direction may be adjusted to set an arbitrary number of data points.

[0056] The histogram may reflect not only the variation in the Z direction, but also the variation in the twist pitch P between multiple strands 5. Specifically, the histogram may be calculated from the relative positional displacement amounts dθ1 to dθ72 for each of the multiple strands 5. One conductor section 4 contains multiple strands 5. Therefore, the local twist pitch P may be calculated for each of the multiple strands 5. If the number of data points includes the twist pitch P of each of the multiple strands 5, the histogram can reflect the variation in the twist pitch P between the multiple strands 5 contained in one conductor section 4. Basically, the histogram of the twist pitch P is created for all the strands contained in the outer conductor 21a.

[0057] When the insulated conductor 2a contains multiple conductor sections 4, the histogram may reflect the variation in the twist pitch P of the strands 5 contained in each of the multiple conductor sections 4. Specifically, each of the first outer conductor 21a1, second outer conductor 21a2, third outer conductor 21a3, fourth outer conductor 21a4, fifth outer conductor 21a5, and sixth outer conductor 21a6 contains multiple strands 5. Therefore, the twist pitch P of the strands 5 contained in each conductor section 4 may be calculated individually. By including the twist pitch P of the strands 5 contained in each conductor section 4 in the number of data points, the histogram can reflect the variation in the twist pitch P of the strands 5 contained in multiple conductor sections 4. In this way, a histogram can be created that also reflects the variation in the twist pitch P of the strands 5 among the multiple conductor sections 4.

[0058] Here, a characteristic of the insulated conductor 2 according to this embodiment is that the half-width W calculated from the histogram of the twist pitch P of the strands 5 in the conductor 21 is 0.6 times or less the median value calculated from the histogram of the twist pitch P. In the insulated conductor 2, the less frequently regions with locally large twist pitch P occur, the more the frequency of sparks in the insulated conductor 2 is suppressed. In particular, a small half-width W of the twist pitch P means that the variation in the twist pitch P is small. In other words, if the half-width W of the twist pitch P of the strands 5 is small, the frequency of regions where the twist pitch P of the strands 5 is locally large will decrease.

[0059] Furthermore, when the multi-core cable 1 is bent, greater bending stress is applied to the outer conductor 21a than to the inner conductor 21b. Therefore, the half-width W of the twist pitch P of the strands 5 in the outer conductor 21a has a greater impact on the bending resistance and sparking frequency of the multi-core cable 1 than the half-width W of the twist pitch P of the strands 5 in the inner conductor 21b.

[0060] (Method for Manufacturing Multicore Cables) Figure 7A is a schematic diagram showing an apparatus for manufacturing a conductor 21 formed by twisting together multiple strands 5 of a multicore cable 1 according to this embodiment. The conductor 21 formed by twisting together multiple strands 5 can be manufactured using the bobbin-fixed manufacturing apparatus shown in Figure 7A. As shown in Figure 7A, the apparatus for manufacturing the conductor 21 mainly comprises multiple rollers 100 to 105, multiple supply reels 200, a sieve plate 300, a die 400, an arm 600, and a winding unit 700. The multiple supply reels 200 are installed so as to be fixed to the floor.

[0061] The supply reel 200 is a bobbin around which strands of wire 5 are wound. The supply reel 200, which is fixed to the floor, rotates, supplying the strands of wire 5. As shown in Figure 7A, multiple strands of wire 5 wound around the supply reel 200 are supplied to the sieve plate 300 via rollers 100 to 105. The median value of the twist pitch P is determined by the rotation speed of the supply reel 200 (feed speed of the strands of wire 5) and the rotation speed of the twisting roller 500 and the arm 600. The rotation speed of the twisting roller 500 and the arm 600 and the feed speed of the strands of wire 5 may be adjusted so that the median value of the twist pitch P is between 10 mm and 60 mm, or the rotation speed of the twisting roller 500 and the arm 600 and the feed speed of the strands of wire 5 may be adjusted so that the median value of the twist pitch P is between 12 mm and 35 mm. When manufacturing insulated wires for electric parking brakes, the rotational speed of the twisting roller 500 and the arm 600 and the feed speed of the strands 5 are adjusted so that the median value of the twist pitch P is, for example, 10 mm or more and 30 mm or less. When manufacturing insulated wires for anti-lock brake systems, the rotational speed of the twisting roller 500 and the arm 600 and the feed speed of the strands 5 are adjusted so that the median value of the twist pitch P is, for example, 20 mm or more.

[0062] The median value of the twist pitch P may be adjusted in the process of twisting together multiple strands 5 shown in Figure 7A to manufacture the conductor 21, and may be further adjusted in the process of twisting together multiple conductors 21 to manufacture the insulated conductor 2, which will be described later. In other words, the rotational speed of the twisting roller 500 and the arm 600 and the feed speed of the strands 5 in the process of twisting together multiple strands 5 shown in Figure 7A to manufacture the conductor 21 may be determined by considering the amount of change in the median value of the twist pitch P in the process of twisting together multiple conductors 21 to manufacture the insulated conductor 2. The median value of the twist pitch P does not change in the process of twisting together multiple insulated conductors 2 shown in Figure 7B to manufacture the multi-core cable 1.

[0063] The planar shape of the spindle plate 300 is, for example, circular. The spindle plate 300 has a plurality of guide holes (not shown). The number of guide holes corresponds to the number of strands 5 to be twisted together. The plurality of guide holes are formed to be distributed, for example, at equal intervals around the center of the spindle plate 300. The plurality of strands 5 are then fed towards the die 400 in a distributed state around the center of the spindle plate 300 by passing through each guide hole.

[0064] Subsequently, the strands 5 dispersed by the sieve plate 300 are gathered in the die 400. The multiple strands 5 gathered in the die 400 are twisted together by the twisting roller 500. The twisted strands 5 are supplied to the winding unit 700 along the arm 600. The arm 600 is rotated in the rotational direction R (shown in Figure 7A). Subsequently, the twisted strands 5 are wound and stored in the winding unit 700. In this way, a conductor 21 made of multiple strands 5 twisted together is manufactured.

[0065] The half-width W of the twist pitch P can be changed by the pass line ratio L. Pass line L1 is the distance from the twisting roller 500 to the nearest supply reel 201 among the multiple supply reels 200. Pass line L2 is the distance from the twisting roller 500 to the furthest supply reel 203 among the multiple supply reels 200. The pass line ratio L is calculated by L2 / L1. Pass lines L1 and L2 correspond to the length of the strand 5 along the transport path of the strand 5 from supply reels 201 and 203 to the twisting roller 500. The pass line ratio can be adjusted by changing the position of the twisting roller 500 (twist point). The pass line ratio can also be reduced by adjusting the relative positions of the multiple supply reels 200 with respect to the twisting roller 500 and the arrangement of rollers 101 to 105. As a result, the half-width W of the twist pitch P of the strand 5 can be reduced, as will be described later.

[0066] Next, the conductor section 4 is manufactured. Specifically, multiple conductors 21 manufactured in Figure 7A are prepared. Each of the multiple conductors 21 is supplied from the winding section 700 manufactured in Figure 7A to the twisting section as six outer conductors 21a and one inner conductor 21b. The multiple conductors 21 are twisted together in the twisting section. The twisted conductors 21 are supplied to the winding section and wound and housed in the winding section.

[0067] Next, the insulated conductor 2 is manufactured. Specifically, the twisted conductor 21 wound and housed in the winding section is supplied to the insulated section. The insulated section heats and melts the resin stored in the storage section and extrudes it to the outside of the twisted conductor 21. In this way, an insulated section 22 is formed on the outside of the twisted conductor 21. The conductor 21 with the insulated section 22 is supplied to the cooling section. In the cooling section, the conductor 21 with the insulated section 22 is cooled, causing the insulated section 22 to harden. In this way, an insulated conductor 2a containing multiple conductors 21 is manufactured. The insulated conductor 2a is supplied to the winding section and wound and housed in the winding section.

[0068] Alternatively, one conductor 21 manufactured as shown in Figure 7A may be prepared, and a covered wire 2b may be manufactured in the same manner as the method for forming the covering portion 22 described above, with the covering portion 22 formed on one conductor 21.

[0069] Figure 7B is a schematic diagram showing a manufacturing apparatus for a multi-core cable 1 according to this embodiment. The multi-core cable 1 can be manufactured using the manufacturing apparatus for the multi-core cable 1 shown in Figure 7B. The manufacturing apparatus for the multi-core cable 1 mainly comprises a plurality of conductor supply reels 801, a twisting section 802, an inner sheath layer covering section 803, a storage section 803a, an outer sheath layer covering section 804, a storage section 804a, a cooling section 805, and a winding section 806.

[0070] The conductor supply reel 801 is, for example, a bobbin around which insulated conductors 2 are wound. The conductor supply reel 801 may also be a winding section around which the aforementioned insulated conductors 2 (insulated conductors 2a or insulated conductors 2b) are wound and housed. The insulated conductors 2 wound around each of the multiple conductor supply reels 801 are supplied to the twisting section 802. The multiple insulated conductors 2 (for example, two insulated conductors 2a and two insulated conductors 2b) are twisted together in the twisting section 802.

[0071] Multiple twisted insulated wires 2 are supplied to the inner sheath layer coating section 803. The inner sheath layer coating section 803 pushes the resin stored in the storage section 803a outwards from the insulated wires 2. In this way, the inner sheath layer 32 is formed on the outside of the insulated wires 2.

[0072] The insulated conductor 2, with the inner sheath layer 32 formed on it, is supplied to the outer sheath layer covering portion 804. The outer sheath layer covering portion 804 pushes the resin stored in the storage portion 804a to the outside of the inner sheath layer 32. In this way, the outer sheath layer 31 is formed on the outside of the inner sheath layer 32.

[0073] The insulated conductor 2, with the outer sheath layer 31 formed on it, is supplied to the cooling unit 805. As the supplied insulated conductor 2 is cooled in the cooling unit 805, the sheath layer 30 formed from the inner sheath layer 32 and the outer sheath layer 31 hardens. In this way, the multi-core cable 1 is manufactured. The multi-core cable 1 is supplied to the winding unit 806 and wound and housed in the winding unit 806.

[0074] As mentioned above, the half-width W of the twist pitch P is determined by the pass line ratio in the process of twisting together multiple strands 5 shown in Figure 7A to manufacture the conductor 21. The twisting conditions may be controlled so that the half-width W of the twist pitch P determined in this way does not change in the subsequent process of twisting together multiple conductors 21 to manufacture the insulated conductor 2.

[0075] <Effects> The insulated conductor 2 according to this disclosure comprises a conductor portion 4 and an insulated portion 22. The insulated portion 22 covers the conductor portion 4. The conductor portion 4 is composed of a plurality of strands 5 twisted together. The conductor portion 4 includes a conductor 21. The conductor 21 is adjacent to the insulated portion 22. The direction in which the conductor portion 4 extends is defined as the Z direction. In the conductor 21, the half-width W is 0.6 times or less of the median value. The half-width W and the median value are calculated from the histogram of the twist pitch P in the Z direction of the conductor 21. When considering two cross-sections perpendicular to the Z direction and separated from each other in the Z direction, the twist pitch P is calculated in a relative coordinate system based on the displacement of the relative position of the strands 5 between the two cross-sections. The relative coordinate system has as its origin the centroid of the conductor 21 in the cross-section (center positions OA, OB of the conductor portion 4). The distance l between the two cross-sections is 1 mm. The histogram is created for the portion of the conductor 21 that has a length of 50 mm in the Z direction.

[0076] In this way, the frequency of occurrence of regions in the strand 5 where the twist pitch P is locally large can be reduced. As a result, a coated conductor 2 with a suppressed frequency of sparks can be obtained. Similarly, the frequency of sparks can be suppressed in the multi-core cable 1 equipped with this coated conductor 2.

[0077] In the above-mentioned insulated conductor 2, if the median is 26.0 mm or more, the half-width W may be 15.5 mm or less. In this way, an insulated conductor 2 with improved bending resistance can be obtained. Similarly, the multi-core cable 1 equipped with this insulated conductor 2 may also have improved bending resistance.

[0078] In the above-described insulated conductor 2, the histogram may be created from the relative positional displacement dθ of each of the multiple strands 5. In this way, not only the variation in the twist pitch P in the Z direction of a single strand 5, but also the variation in the twist pitch P among the multiple strands 5 contained in a single conductor 21 can be reflected in the histogram.

[0079] In the above-described insulated conductor 2a, the conductor 21 may include a first outer conductor 21a1 and a second outer conductor 21a2. The histogram may be created from the twist pitch P of the first outer conductor 21a1 and the second outer conductor 21a2, respectively. In this way, a histogram can be created that includes the variation in the twist pitch P of the strands 5 among multiple conductors 21.

[0080] In the above-described insulated conductor 2a, the conductor portion 4 may include an inner conductor 21b. The inner conductor 21b may be positioned away from the insulated portion 22. In this way, the twist pitch P of the strands 5 in the outer conductor 21a and the twist pitch P of the strands 5 in the inner conductor 21b can be reflected in the histogram.

[0081] The multi-core cable 1 according to this disclosure may include the above-mentioned insulated conductor 2. In this way, a multi-core cable 1 can be obtained in which the frequency of spark generation is suppressed.

[0082] To verify the effectiveness of the insulated conductor 2a according to the above embodiment, the following tests were conducted. Specifically, for a sample of a multi-core cable with the configuration shown in Figure 1, the correlation between the half-width W, the ratio of the half-width W to the median, the pull-out force, the bending resistance, and the frequency of spark anomalies was evaluated.

[0083] <Test Subjects> Tables 1 to 3 show the configuration of the multi-core cable 1 including the insulated conductor 2a for samples 1 to 12, and the pull-out force, bending resistance, and spark abnormality frequency for each sample. The specific configuration of the multi-core cable 1 for samples 1 to 12 will be described below. The multi-core cable 1 for samples 1 to 12 consists of two insulated conductors 2a, two insulated conductors 2b, and a sheath layer 30. The two insulated conductors 2a and the two pre-twisted insulated conductors 2b are further twisted together to form the core wire. The sheath layer 30 is applied around the core wire by extrusion. The sheath layer 30 includes an outer sheath layer 31 and an inner sheath layer 32. The main component of the inner sheath layer 32 is a cross-linked polyolefin. The minimum thickness of the inner sheath layer 32 is 0.45 mm. The outer diameter of the inner sheath layer 32 is 7.4 mm. The main component of the outer sheath layer 31 is flame-retardant crosslinked polyurethane. The thickness of the outer sheath layer 31 is 0.5 mm. The outer diameter of the outer sheath layer 31 is 8.4 mm. The crosslinking of the resin component of the sheath layer 30 was performed by electron beam irradiation at 180 kGy. The insulated conductor 2a has a conductor portion 4 and an insulated portion 22. The conductor portion 4 is composed of six outer conductors 21a and one inner conductor 21b twisted together. The outer diameter of the conductor portion 4 is 2.4 mm. The insulated portion 22 is formed by extruding an insulating layer-forming structure onto the outer circumference of the conductor portion 4. As the insulating layer-forming material, 100 parts by mass of EEA (ethylene ethyl acrylate copolymer) (product name: DPDJ-6182) (ethyl acrylate content 15% by mass) manufactured by ENEOS NUC Corporation was mixed with 70 parts by mass of aluminum hydroxide (product name: Hydelilite H-32) manufactured by Toshin Kasei Co., Ltd. as a flame retardant and 2 parts by mass of a plastic additive (product name: Irganox® 1010) manufactured by BASF as an antioxidant, and thoroughly mixed in a resin kneader. The thickness of the coating portion 22 is 0.3 mm, and the outer diameter of the coated conductor wire 2a is 3 mm. The coating portion 22 was irradiated with an electron beam at 60 kGy to crosslink the resin components. The conductor portion 4 consists of six outer conductors 21a and one inner conductor 21b, each composed of 72 strands 5 twisted together. The diameter of the strands 5 is 80 μm.The insulated conductor 2b consists of a conductor 21 and an insulated portion 22. The outer diameter of the conductor 21 is 0.72 mm. The insulated portion 22 is formed by extruding the same insulating layer-forming material as the insulated conductor 2a onto the outer circumference of the conductor 21. The thickness of the insulated portion 22 is 0.36 mm, and the outer diameter of the insulated conductor 2b is 1.45 mm. The conductor 21 is made up of 60 strands 5 twisted together. The diameter of each strand 5 is 80 μm. The material constituting the strands 5 is soft copper. The multi-core cables 1 of samples 1 to 10 are multi-core cables 1 formed from conductors 21 manufactured using the bobbin-fixed manufacturing apparatus shown in Figure 7A. The multi-core cables 1 of samples 11 and 12 are multi-core cables 1 formed from conductors 21 manufactured using the bobbin-rotating manufacturing apparatus shown in Figure 42.

[0084] Figure 42 is a schematic diagram showing an apparatus for manufacturing a conductor 21 formed by twisting together multiple strands 5 of a multi-core cable 1 according to samples 11 and 12. The multi-core cables 1 of samples 11 and 12 can be formed from a conductor 21 manufactured using the bobbin rotation type manufacturing apparatus shown in Figure 42. As shown in Figure 42, the apparatus for manufacturing a conductor 21 formed by twisting together multiple strands 5 of a multi-core cable 1 of samples 11 and 12 mainly comprises a plurality of supply reels 210, a sieve plate 310, a die 410, and a cradle 211. As shown in Figure 42, the plurality of supply reels 210 are fixed to the cradle 211 so as to be arranged on the circumference of the central axis A1.

[0085] The supply reel 210 is a bobbin around which strands of wire 5 are wound. The cradle 211 rotates on its central axis A1. The supply reel 210, which is fixed to the cradle 211, revolves around the central axis A1, supplying the strands of wire 5. As shown in Figure 42, the multiple strands of wire 5 wound on the supply reel 210 are supplied to the die 410 via the sieve plate 310. The sieve plate 310 rotates in sync with the cradle 211. The multiple strands of wire 5 are gathered and twisted together in the die 410.

[0086] As shown in Figure 42, in a bobbin-rotating manufacturing apparatus, the pass line is the distance L10 from the supply reel 210 to the die 410. Multiple supply reels 210 are arranged on the circumference of the central axis A1. Therefore, the distance L10 from each of the multiple supply reels 210 to the die 410 is equal to one another. Thus, the pass line ratio L is 1. Sample 11 is a multi-core cable 1 manufactured in substantially the same manner as the method disclosed in Japanese Patent Application Publication No. 2006-328603. Sample 12 is a multi-core cable 1 manufactured in substantially the same manner as the method disclosed in Japanese Patent Application Publication No. Hei 4-075733.

[0087]

[0088]

[0089]

[0090] As shown in Tables 1 to 3, the test subjects are multi-core cables 1 containing 12 different types of insulated conductors 2a, from Sample 1 to Sample 12. Samples 1 to 5 are examples. The insulated conductors 2a of Samples 1 to 5 are manufactured with a pass-line ratio L of 1.6. Samples 6 to 12 are comparative examples. The insulated conductors 2a of Samples 6 to 10 are manufactured with a pass-line ratio L of 2.3. The insulated conductors 2a of Samples 11 and 12 are manufactured using a bobbin-rotating manufacturing apparatus with a pass-line ratio L of 1.0. The half-width W of the twist pitch P of the outer conductor 21a in the insulated conductors 2a of Samples 1 to 5 is 0.6 times or less the median value of the twist pitch P of the outer conductor 21a in the insulated conductor 2a (i.e., the half-width W of the twist pitch P relative to the median value of the twist pitch P is 0.6 or less). On the other hand, the half-width W of the twist pitch P of the outer conductor 21a in the coated wires 2a of samples 6 to 12 is greater than 0.6 times the median value of the twist pitch P of the outer conductor 21a in the coated wires 2a (i.e., the half-width W of the twist pitch P relative to the median value of the twist pitch P is greater than 0.6). In particular, the half-width W of the twist pitch P of the outer conductor 21a in the coated wires 2a of samples 11 and 12, which were manufactured using a bobbin rotation type manufacturing apparatus, is 18 mm or more, indicating a large half-width W of the twist pitch P.

[0091] When manufacturing a conductor 21 made by twisting multiple strands 5 of a multi-core cable 1 using a bobbin rotation type manufacturing apparatus as shown in Figure 42, even if the pass line ratio L is 1, centrifugal force is applied to the strands 5 as the bobbin revolves, causing wire runout. On the other hand, when manufacturing a conductor 21 made by twisting multiple strands 5 of a multi-core cable 1 using a bobbin fixed type manufacturing apparatus as shown in Figure 7A, the bobbin does not revolve, so no wire runout occurs in the strands 5. Therefore, the half-width W of the twist pitch P of the outer conductor 21a of samples 1 to 5 manufactured using the bobbin fixed type manufacturing apparatus is smaller than the half-width W of the twist pitch P of the outer conductor 21a of samples 11 and 12 manufactured using the bobbin rotation type manufacturing apparatus.

[0092] As shown in Tables 1 to 3, the median twist pitch P of the outer conductor 21a in the insulated wire 2a from Sample 1 to Sample 12 is between 10.0 mm and 30.0 mm. The median twist pitch P represents the value of the twist pitch P that falls exactly in the middle when the data points are arranged in ascending order when calculating the histogram. When the number of data points is even, the median is the average of the two centrally ranked points.

[0093] In the insulated conductor 2a from Sample 1 to Sample 12, the coefficient of linear expansion C of the insulated portion 22 is 2.0 × 10⁻⁶. －４ The product C × E is 0.56 MPa / K. In the coated conductor 2a from sample 1 to sample 12, the elastic modulus E of the coating portion 22 is 2800 MPa. In other words, in the coated conductor 2a from sample 1 to sample 12, the product C × E is 0.56 MPa / K.

[0094] <Test Method> The method for measuring the pull-out force of the insulation portion 22 is described below. For the insulated conductor 2a from Sample 1 to Sample 12, the insulation portion 22 was removed, leaving 50 mm in the Z direction, to expose the conductor portion 4. Next, a metal plate (5 mm thick) with a hole whose inner diameter was larger than the diameter of the conductor portion 4 and smaller than the outer diameter of the insulation portion 22 was prepared. The conductor portion 4 was passed through the hole in the metal plate, and the metal plate was fixed while the conductor 21 was pulled up at a speed of 200 mm / min. At this time, the insulation portion 22 was caught on the metal plate and could not be pulled up, and only the conductor 21 was pulled out from the insulation portion 22. The force required to pull out a 50 mm length of conductor 21 from a 50 mm length of insulation portion 22 was measured, and the maximum value was defined as the pull-out force. The results are shown in Tables 1 to 3.

[0095] Next, the method for measuring the number of bends of the multi-core cable 1 will be described. The bending resistance of the multi-core cable 1 is evaluated by the number of bends. Specifically, two mandrels with a diameter of 60 mm were prepared, positioned horizontally and parallel to each other. Then, the multi-core cables 1 from sample 1 to sample 12 were passed vertically between the two mandrels. The multi-core cable 1 was bent horizontally by 90° so that the upper end of the multi-core cable 1 abutted against the upper side of one mandrel. Then, the multi-core cable 1 was bent in the opposite direction by 90° so that the upper end of the multi-core cable 1 abutted against the upper side of the other mandrel. This operation was repeated. The test conditions were a downward load of 2 kg applied to the lower end of the multi-core cable 1, a temperature of -30°C, and a bending speed of 60 times / minute. In this test, the number of bends until the multi-core cable 1 broke (became unable to conduct electricity) was measured. The results are shown in Tables 1 to 3. Furthermore, when the number of bends is measured using the multi-core cable 1 from Sample 1 to Sample 12 as the measurement target, it is preferable that the number of bends be 27,000 or more.

[0096] Next, a method for measuring the frequency of spark anomalies using a spark test will be described. The frequency of spark occurrence is evaluated by the spark anomaly frequency. The spark test is performed during the extrusion process in which the covering portion 22 is formed on the outer circumference of the conductor portion 4. The spark anomaly frequency is the number of times pinholes and exposed conductors are detected per 1 mm (megameter) of the covered conductor wire 2a. A spark tester (HB15F) manufactured by Clinton was used for the spark test. The spark tester was set to a voltage of 5 kV. When measuring the spark anomaly frequency using the multi-core cable 1 from sample 1 to sample 12 as the measurement target, it is preferable that the spark anomaly frequency is 0.1 times / mm or less.

[0097] <Evaluation> In samples 1 to 12, where the porosity is 4%, the pull-out force of the coating portion 22 is 70 N.

[0098] As shown in Tables 1 to 3, it can be seen that a smaller half-width W corresponds to a larger number of bends. Specifically, the number of bends in Samples 1 to 4, where the half-width W is 15.5 mm or less, is 27,000 or more. In particular, the number of bends in Sample 4, where the median is 26.0 mm or more and the half-width W is 15.5 mm or less, is 29,000, which is more than 27,000. On the other hand, the number of bends in Sample 5, where the median is 26.0 mm or more and the half-width W is greater than 15.5 mm, is 18,700, which is less than 27,000. Therefore, it can be seen that if the median is 26.0 mm or more, a coated conductor 2 with improved bending resistance can be obtained if the half-width W is 15.5 mm or less. Note that the number of bends in Samples 11 and 12, which were manufactured using a bobbin rotation type manufacturing device, is 5,000 or less in both cases. This is because the full width at half maximum (FWHM) W is large in samples 11 and 12, which were manufactured using a bobbin-rotating manufacturing device.

[0099] As shown in Tables 1 to 3, the spark anomaly frequency in samples 1 to 5, where the ratio of the full width at half maximum (FWHM) W to the median is 0.6 or less, is lower than the spark anomaly frequency in samples 6 to 12, where the ratio of the FWHM W to the median is greater than 0.6, and is 0.1 times / Mm or less. From this result, it can be seen that when the ratio of the FWHM W to the median is 0.6 or less, the spark anomaly frequency is low. In particular, the spark anomaly frequencies in samples 11 and 12, which were manufactured using a bobbin rotation type manufacturing device, are both greater than 1.0 times / Mm. This is because the FWHM W is large in samples 11 and 12, which were manufactured using a bobbin rotation type manufacturing device, and the ratio of the FWHM W to the median is greater than 0.6.

[0100] Based on the above results, the ratio of the half-width W to the median value at the twist pitch P of the outer conductor 21a is smaller for the multi-core cable 1 formed from conductors 21 manufactured using a bobbin-fixed manufacturing device than for the multi-core cable 1 formed from conductors 21 manufactured using a bobbin-rotating manufacturing device.

[0101] Furthermore, the smaller the half-width W, the greater the number of bends. In particular, even if the median is 26.0 mm or more, if the half-width W is 15.5 mm or less, the number of bends will exceed 27,000. In other words, a smaller half-width W improves bending resistance.

[0102] Furthermore, when the ratio of the full width at half maximum (FWHM) W to the median is 0.6 or less, the frequency of spark anomalies is small, and the frequency of spark anomalies is 0.1 times / Mm or less. In other words, when the ratio of the FWHM W to the median is 0.6 or less, the occurrence of pinholes in the coating portion 22 of the coated conductor 2a is suppressed. It was also confirmed that even if the material of the coating portion 22 is different from the material (polyolefin) used for the coating portion 22 from sample 1 to sample 12, the frequency of spark anomalies is small when comparing between the same materials when the ratio of the FWHM W to the median is 0.6 or less. In addition, when the thickness of the coating portion 22 is, for example, 0.1 mm or more and 0.4 mm or less, it was confirmed that the frequency of spark anomalies is small when comparing with the same coating thickness when the ratio of the FWHM W to the median is 0.6 or less. In other words, it was found that regardless of the material and thickness of the coating portion 22, when the ratio of the half-width W to the median is 0.6 or less, the frequency of spark anomalies is small and the occurrence of pinholes in the coating portion 22 is suppressed.

[0103] Referring to Figure 8, the configuration of the multi-core cable quality evaluation system 40 will be described. The multi-core cable quality evaluation system 40 comprises a processed image generation device 50, a learning device 60, a learned model storage device 41, a center position detection device 70, and an index calculation device 80.

[0104] The learning device 60 generates a trained model that generates a processed tomographic image from a tomographic image of a cross-section perpendicular to the longitudinal direction of a multi-core cable containing multiple strands. The processed tomographic image is an image in which the pixels at the center of the strands in the tomographic image are bright, and the pixels gradually become darker toward the radial direction of the cross-section of the strands.

[0105] The processed image generation device 50 generates processed images from acquired tomographic images using a trained model.

[0106] The trained model storage device 41 stores the trained model. The center position detection device 70 detects the center position of the strands included in the processed image.

[0107] The index calculation device 80 calculates an index representing the strand twist pitch based on the strand trajectories obtained from the center positions of the strands in multiple processed images.

[0108] Referring to Figure 9, the configuration of the learning device 60 will be described. The learning device 60 comprises a data acquisition unit 61, a learning data generation unit 62, and a model generation unit 63.

[0109] The data acquisition unit 61 acquires a tomographic image of a cross-section perpendicular to the longitudinal direction of a multi-core cable containing multiple strands.

[0110] The training data generation unit 62 generates a processed image from a tomographic image based on user input, such that the pixels at the center of each strand in the tomographic image are bright, and the pixels gradually darken towards the radial direction of the cross-section of each strand. The training data generation unit 62 takes one tomographic image as input data and generates training data using the processed image generated from that tomographic image as training data. The training data generation unit 62 generates multiple training data sets by using multiple tomographic images or parts of tomographic images. Specifically, the user places circles of a certain diameter at the positions of each strand in the tomographic image while viewing the tomographic image. For each placed circle, the training data generation unit 62 generates a processed image such that the center of the circle is the brightest (highest pixel value), gradually darkens along the radial direction of the strand, and darkens most at the circumference (pixel value is 0). The pixel values ​​of the processed image consist only of the pixel values ​​of the circles and do not include the pixel values ​​of the original tomographic image.

[0111] If an image with constant pixel values ​​across the entire circle is used as training data, rather than a processed image like the one in this embodiment, the binarization process described later may fail to cut the two touching strands. On the other hand, if an image with constant pixel values ​​at only one point in the center of the strands is used as training data, learning becomes difficult. This is because the learning process may attempt to generate an image with small errors from the training data, where all pixel values ​​are 0.

[0112] The model generation unit 63 uses multiple training data to generate a trained model that generates a processed image from a tomographic image in which the pixels at the center of the cross-section of the strands in the tomographic image are bright, and the pixels gradually become darker towards the radial direction of the cross-section of the strands. Specifically, the model generation unit 63 inputs input data to the first stage convolutional layer CV1 of the neural network in Figure 10, and learns the coefficients of filters FL1, FL2, and FL3 so that the sum of the squares of the errors between the data output from the third stage convolutional layer CV3 and the training data is small. The model generation unit 63 stores the generated trained model in the trained model storage device 41.

[0113] Referring to Figure 11, the configuration of the processed image generation device 50 will be described. The processed image generation device 50 comprises a data acquisition unit 51 and an image generation unit 52.

[0114] The data acquisition unit 51 acquires a tomographic image of a cross-section perpendicular to the longitudinal direction of a multi-core cable containing multiple strands.

[0115] The image generation unit 52 uses the trained model stored in the trained model storage device 41 to generate a processed image from the acquired tomographic image such that the pixels at the center of the cross-section of each strand in the tomographic image are bright, and the pixels gradually become darker towards the radial direction of the cross-section of each strand. Specifically, the image generation unit 52 inputs the tomographic image to the first stage convolutional layer CV1 of the trained neural network shown in Figure 10, and outputs the data output from the third stage convolutional layer CV3 as the processed image.

[0116] Next, we show the experimental results when the filter size of the neural network is changed. Figure 12A is a diagram showing the tomographic image input to the neural network. Figure 12B is a magnified view of the region represented by the rectangle in Figure 12A.

[0117] Figures 13A to 19A show the processed images when the filter size f1 of the first convolutional layer CV1 is changed.

[0118] Figure 13A shows the processed image output from the neural network when the filter size is {f1=5, f2=1, f3=5}. Figure 13B is an enlarged view of the area represented by the rectangle in Figure 13A. Figure 14A shows the processed image output from the neural network when the filter size is {f1=10, f2=1, f3=5}. Figure 14B is an enlarged view of the area represented by the rectangle in Figure 14A. Figure 15A shows the processed image output from the neural network when the filter size is {f1=15, f2=1, f3=5}. Figure 15B is an enlarged view of the area represented by the rectangle in Figure 15A. Figure 16A shows the processed image output from the neural network when the filter size is {f1=25, f2=1, f3=5}. Figure 16B is an enlarged view of the area represented by the rectangle in Figure 16A. Figure 17A shows the processed image output from the neural network when the filter size is {f1=35, f2=1, f3=5}. Figure 17B is an enlarged view of the area represented by the rectangle in Figure 17A. Figure 18A shows the processed image output from the neural network when the filter size is {f1=45, f2=1, f3=5}. Figure 18B is an enlarged view of the area represented by the rectangle in Figure 18A. Figure 19A shows the processed image output from the neural network when the filter size is {f1=55, f2=1, f3=5}. Figure 19B is an enlarged view of the area represented by the rectangle in Figure 19A.

[0119] Figures 20A to 22A show the processed images when the filter size f2 of the second convolutional layer CV2 is changed.

[0120] Figure 20A shows the processed image output from the neural network when the filter size is {f1=35, f2=1, f3=15}. Figure 20B is an enlarged view of the area represented by the rectangle in Figure 20A. Figure 21A shows the processed image output from the neural network when the filter size is {f1=35, f2=5, f3=15}. Figure 21B is an enlarged view of the area represented by the rectangle in Figure 21A. Figure 22A shows the processed image output from the neural network when the filter size is {f1=35, f2=15, f3=15}. Figure 22B is an enlarged view of the area represented by the rectangle in Figure 22A.

[0121] Figures 23A to 28A show processed images when the filter size f3 of the third convolutional layer CV3 is changed.

[0122] Figure 23A shows the processed image output from the neural network when the filter size is {f1=35, f2=1, f3=5}. Figure 23B is an enlarged view of the area represented by the rectangle in Figure 23A. Figure 24A shows the processed image output from the neural network when the filter size is {f1=35, f2=1, f3=10}. Figure 24B is an enlarged view of the area represented by the rectangle in Figure 24A. Figure 25A shows the processed image output from the neural network when the filter size is {f1=35, f2=1, f3=15}. Figure 25B is an enlarged view of the area represented by the rectangle in Figure 25A. Figure 26A shows the processed image output from the neural network when the filter size is {f1=35, f2=1, f3=25}. Figure 26B is an enlarged view of the area represented by the rectangle in Figure 26A. Figure 27A shows the processed image output from the neural network when the filter sizes are {f1=35, f2=1, f3=45}. Figure 27B is an enlarged view of the area represented by the rectangle in Figure 27A. Figure 28A shows the processed image output from the neural network when the filter sizes are {f1=35, f2=1, f3=55}. Figure 28B is an enlarged view of the area represented by the rectangle in Figure 28A.

[0123] Based on the above results, we will explain the appropriate and optimal values ​​for the number and size of filters in the convolutional neural network of Figure 10 with reference to Figure 29.

[0124] The number of filters FL1 in the first stage convolutional layer CV1 (n1) is set to 128. The side size f1 of the filter FL1 in the first stage convolutional layer CV1 is set to 15 to 45 pixels. Furthermore, the side size f1 may be set to 35 pixels. The number of filters FL2 in the second stage convolutional layer CV2 (n2) is set to 64 pixels. The side size f2 of the filter FL2 in the second stage convolutional layer CV2 is set to 1, 5, or 15 pixels. The side size f3 of the filter FL3 in the third stage convolutional layer CV3 is set to 10 to 25 pixels. Furthermore, the side size f3 may be set to 15 pixels.

[0125] The diameter d of the strands in the tomographic image is represented by 17 pixels in the original image shown in Figure 12A. The appropriate and optimal values ​​based on the diameter d are as follows.

[0126] A side size of 15 to 45 pixels in FL1 corresponds to 0.88 to 2.64 times the diameter d of the strand. A side size of 35 pixels in FL1 corresponds to 2.06 times the diameter d of the strand. The first stage filter is responsible for extracting local features from the original image. Generally, to determine whether a strand exists in a certain area of ​​an image, it is necessary to examine an area equivalent to one to several strands, centered on that area. The fact that a filter equivalent to about two strands is optimal supports this fact. On the other hand, if the filter is made too large, it will capture information from an area that is unnecessarily wide for determining whether a strand exists, and the accuracy of the conversion will not improve, or may even decrease.

[0127] The side sizes of 1, 5, and 15 pixels in FL2 correspond to 0.06, 0.29, and 0.88 times the wire diameter d, respectively. The second-stage filter is responsible for understanding the interrelationships between multiple feature maps extracted by the first-stage filter. If the feature quantities extracted by the first-stage filter are sufficiently good, the accuracy can be high even with a filter size of 1, meaning that information in the in-plane direction is not included. In some cases, when combining the feature quantities extracted by the first-stage filter, bundling features at positions slightly shifted in the in-plane direction may result in higher conversion accuracy.

[0128] A side size of 10 to 25 pixels for FL3 corresponds to 0.59 to 1.47 times the diameter d of the strand. A side size of 15 pixels for FL3 corresponds to 0.88 times the diameter d of the strand. The third stage filter is responsible for generating images where strands do not touch each other, based on the image features compiled by the second stage filter. Ideally, it is necessary to generate many shapes that are about the size of a strand, with a bright center that gradually darkens towards the ends of the strand. Therefore, when referencing features, by comprehensively incorporating information from an area about the size of a strand from that point, it becomes possible to determine with higher accuracy whether that point is the center of the strand or away from the center of the strand. From this, it can be said that it is reasonable for the size of f3 to be about the diameter of the strand, or at least the radius of the strand.

[0129] Referring to Figure 30, the configuration of the center position detection device 70 will be described. The center position detection device 70 comprises a binarization unit 71, a segmentation unit 72, a distance conversion unit 73, a strand region detection unit 74, and a center detection unit 75.

[0130] The binarization unit 71 binarizes the processed image. The segmentation unit 72 detects multiple consecutive regions with the same pixel value from the binarized image.

[0131] The distance conversion unit 73 generates a distance conversion image for each region. The strand region detection unit 74 generates subregions by binarizing the distance conversion image for each region with an adaptive threshold. The strand region detection unit 74 selects an appropriate threshold based on the number of subregions. Specifically, for each region, the strand region detection unit 74 counts the number of subregions in which pixels having a first pixel value included in the binarized distance conversion image are consecutive. For each region, the strand region detection unit 74 obtains multiple count values ​​by sequentially increasing the threshold TH from an initial value until the count value decreases. The strand region detection unit 74 determines the number of strands to be the number of subregions included in the binarized distance conversion image generated by one of the multiple threshold TH values ​​when the count value becomes the maximum count value, if the ratio R of the threshold TH when the multiple count values ​​become the maximum count value is greater than or equal to a reference value.

[0132] The center detection unit 75 detects the position of any pixel in the subregion (for example, the position of the pixel with the highest pixel value) as the center position of the strand.

[0133] Referring to Figure 31, the configuration of the index calculation device 80 will be described. The index calculation device 80 comprises a trajectory data generation unit 81, a twist pitch calculation unit 82, a histogram creation unit 83, and a statistical quantity calculation unit 84.

[0134] Assume that the processed image generation device 50 generates N processed images RI(i) from N tomographic images X(i), where i = 1 to N.

[0135] As shown in Figure 32, the trajectory data generation unit 81 obtains trajectory data for each strand by connecting the center position of the cross-section of each strand included in the processed image RI(i) to the closest center position among the cross-sections of multiple strands included in the processed image RI(i+1), thereby obtaining trajectory data for each strand as shown in Figure 33.

[0136] In this specification, determining the locus of a wire means not determining the actual locus of the wire itself, but rather determining the coordinates of the endpoints of each of a series of line segments that approximate the actual locus of the wire.

[0137] Here, L ｉ(i = 1 to LN) represents the longitudinal position of the multi-core cable. (X, Y) represents the center position of the cross-section of the individual wire.

[0138] The locus of the individual wire is approximated by a line segment connecting the center positions (X, Y) at L ｉ and the center positions (X, Y) at L ｉ＋１ i = 1 to LN - 1. The interval between L ｉ and L ｉ＋１ can be, for example, 10 to 100 μm. LN can be, for example, 3,000 to 5,000.

[0139] The twist pitch calculation unit 82 calculates the first twist pitch, the second twist pitch, the third twist pitch, and the fourth twist pitch as follows. The third twist pitch and the fourth twist pitch are the twist pitches of the individual wires. The first twist pitch and the second twist pitch are the twist pitches of the aggregate formed by twisting a plurality of individual wires.

[0140] The twist pitch calculation unit 82 selects the longitudinal positions L ｉ (i = 1 to LN - 1) every K. Let the selected longitudinal positions be Z ｉ (i = 1 to N - 1). Z ｉ = L Ｋ×ｉ where N = LN / K. The interval between Z ｉ and Z ｉ＋１ can be, for example, 1 mm. N can be, for example, 51.

[0141] FIG. 34 is a diagram for explaining the first twist pitch. In the cross-section of Z ｉ , let the centroid position of the center positions of all the individual wires included in one of the coated conductors 2a be O ｉ (1), and the centroid position of the center positions of all the individual wires included in the other coated conductor 2a be O ｉ (2). Let the centroid position of the center positions of all the individual wires included in the two coated conductors 2a be O ｉ .

[0142] In the cross-section of Z ｉ , the angle formed by the line segment connecting O ｉ and O ｉ (j) (j = 1 to 2) and the X-axis is θｉ Let (j) be the case. Z ｉ＋１ In the cross-section, O ｉ＋１ and O ｉ＋１ The angle between the line segment connecting (j) (j=1 to 2) and the x-axis is θ. ｉ＋１ (j) The twist pitch calculation unit 82 calculates the following values.

[0143] dθ ｉ (j) = θ ｉ＋１ (j) - θ ｉ (j) PT ｉ (j) = 360 / dθ ｉ (j) For i = 1 to N-1 and j = 1 to 2, PT ｉ (j) is the first indicator. The sample size for the first indicator is (N-1) × 2.

[0144] Figure 35 is a diagram illustrating the second twist pitch. Z ｉ In the cross-section, the outer conductor included in one of the insulated wires 2a is defined as follows: The centroid position of all strands included in the outer conductor 21a1 is P ｉ (1,1) The centroid position of all strands included in the outer conductor 21a2 is P ｉ (1, 2) The centroid position of all strands included in the outer conductor 21a3 is P ｉ (1, 3) The centroid position of all strands included in the outer conductor 21a4 is P ｉ (1, 4) The centroid position of all strands included in the outer conductor 21a5 is P ｉ (1, 5) The centroid position of all strands included in the outer conductor 21a6 is P ｉ (1, 6) Let P be the centroid of the center position of all strands included in the outer conductors 21a1, 21a2, 21a3, 21a4, 21a5, and 21a6. ｉ (1)

[0145] Z ｉ In the cross-section, the outer conductor included in the other insulated conductor 2a is defined as follows: The centroid position of all strands included in the outer conductor 21a1 is P ｉ (2,1) The centroid position of all strands included in the outer conductor 21a2 is P ｉ (2,2) The centroid position of all strands included in the outer conductor 21a3 is Pｉ (2,3) The centroid position of all strands included in the outer conductor 21a4 is P ｉ (2,4) The centroid position of all strands included in the outer conductor 21a5 is P ｉ (2, 5) The centroid position of all strands included in the outer conductor 21a6 is P ｉ (2, 6) Let P be the centroid of the center position of all strands included in the outer conductors 21a1, 21a2, 21a3, 21a4, 21a5, and 21a6. ｉ (2)

[0146] Z ｉ In the cross-section, P ｉ (j) and P ｉ Let θ be the angle between the line segment connecting (j, k) (j=1 to 2, k=1 to 6) and the x-axis. ｉ Let (j, k). Z ｉ＋１ In the cross-section, P ｉ＋１ (j) and P ｉ＋１ Let θ be the angle between the line segment connecting (j, k) (j=1 to 2, k=1 to 6) and the x-axis. ｉ＋１ Let (j, k). The twist pitch calculation unit 82 calculates the following values.

[0147] dθ ｉ (j, k) = θ ｉ＋１ (j, k) - θ ｉ (j,k) PT ｉ (j,k)=360 / dθ ｉ (j, k) For i = 1 to N-1, j = 1 to 2, k = 1 to 6, PT ｉ (j, k) is the second indicator. The sample size for the second indicator is (N-1) × 2 × 6.

[0148] Figure 36 is a diagram illustrating the third twist pitch. ｉ In the cross-section, the outer conductor included in one of the insulated wires 2a is defined as follows: The centroid position of the M strands included in the outer conductor 21al (k=1 to 6) is P ｉ (1, k), the center position of each strand included in the outer conductor 21al (k = 1 to 6) is P ｉ Let (1, k, m) (m = 1 to M).

[0149] Z ｉIn the cross-section, the outer conductor included in the other insulated conductor 2a is defined as follows: The centroid position of the M strands included in the outer conductor 21al (k=1 to 6) is P ｉ (2, k), the center position of each strand included in the outer conductor 21al (k = 1 to 6) is P ｉ Let (2, k, m) (m = 1 to M).

[0150] Z ｉ In the cross-section, P ｉ (j, k) and P ｉ Let θ be the angle between the line segment connecting (j, k, m) (j=1 to 2, k=1 to 6, m=1 to M) and the X-axis. ｉ Let (j, k, m). Z ｉ＋１ In the cross-section, P ｉ＋１ (j, l) and P ｉ＋１ Let θ be the angle between the line segment connecting (j, k, m) (j=1 to 2, k=1 to 6, m=1 to M) and the X-axis. ｉ＋１ Let (j, k, m). The twist pitch calculation unit 82 calculates the following values.

[0151] dθ ｉ (j, k, m) = θ ｉ＋１ (j, k, m) - θ ｉ (j, k, m) PT ｉ (j, k, m) = 360 / dθ ｉ (j, k, m) For i = 1 to N-1, j = 1 to 2, k = 1 to 6, m = 1 to M, PT ｉ (j, k, m) is the third indicator. The sample size for the third indicator is (N-1) × 2 × 6 × M.

[0152] Figure 37 is a diagram illustrating the fourth twist pitch. ｉ In the cross-section, the inner conductor 21b included in one of the insulated conductors 2a is defined as follows: The centroid position of the M strands included in the inner conductor 21b is S. ｉ (1) The center position of each strand included in the inner conductor 21b is S ｉ Let (1, m) (m = 1 to M).

[0153] Z ｉIn the cross-section of, for the inner conductor 21b included in the other coated conductor 2a, it is defined as follows. The centroid position of the M strands included in the inner conductor 21b is S ｉ (2), the center position of each strand included in the inner conductor 21b is S ｉ (2, m) (m = 1 to M).

[0154] Z ｉ In the cross-section of, the angle formed by the line segment connecting S ｉ (j) and S ｉ (j, m) (j = 1 to 2, m = 1 to M) and the X-axis is θ ｉ (j, m). Z ｉ＋１ In the cross-section of, the angle formed by the line segment connecting S ｉ＋１ (j) and S ｉ＋１ (j, m) (j = 1 to 2, m = 1 to M) and the X-axis is θ ｉ＋１ (j, m). The twist pitch calculation unit 82 calculates the following values.

[0155] dθ ｉ (j, m) = θ ｉ＋１ (j, m) - θ ｉ (j, m) PT ｉ (j, m) = 360 / dθ ｉ (j, m) for i = 1 to N - 1, j = 1 to 2, m = 1 to M, PT ｉ (j, m) is the fourth index. The number of samples of the fourth index is (N - 1) × 2 × M.

[0156] The statistic calculation unit 84 calculates the statistic of the index using the histogram of the calculated index. The statistic is, for example, the average value, median, variance, standard deviation, or percentile of the index, etc. The statistic may also be the half-value width of the frequency calculated from the histogram of the index.

[0157] Referring to FIG. 38, the operation procedure during learning will be described. In step S10, the data acquisition unit 61 acquires a tomographic image of a cross-section perpendicular to the longitudinal direction of a multi-core cable including a plurality of strands.

[0158] In step S102, the training data generation unit 62 generates a processed image from the tomographic image based on the user's operation, such that the pixels at the center of the strands in the tomographic image are bright, and the pixels gradually darken toward the radial direction of the cross-section of the strands. The training data generation unit 62 takes one tomographic image as input data and generates multiple training data sets using the processed image generated from that tomographic image as training data.

[0159] In step S103, the model generation unit 63 uses multiple training data to generate a trained model that generates a processed image from a tomographic image in which the pixels at the center of the cross-section of the strands in the tomographic image are bright, and the pixels gradually become darker towards the radial direction of the cross-section of the strands. The model generation unit 63 stores the generated trained model in the trained model storage device 41.

[0160] Referring to Figure 39, the procedure for evaluating the quality of a multi-core cable will be explained. In step S201, the data acquisition unit 51 of the processed image generation device 50 acquires a tomographic image of the cross-section of a multi-core cable containing multiple strands.

[0161] In step S202, the image generation unit 52 of the processed image generation device 50 uses the trained model stored in the trained model storage device 41 to generate a processed image from the tomographic image acquired in step S201, such that the pixels at the center of the cross-section of each strand in the tomographic image are bright, and the pixels gradually become darker in the radial direction of the cross-section of each strand.

[0162] In step S203, the center position detection device 70 detects the center position of each strand from the processed image.

[0163] In step S204, the trajectory data generation unit 81 of the index calculation device 80 generates trajectory data for each strand by connecting the center positions of each strand in multiple processed images.

[0164] In step S205, the twist pitch calculation unit 82 of the index calculation device 80 calculates an index representing the twist pitch of each strand, or an index representing the twist pitch of an assembly of multiple strands twisted together, based on the trajectory data of each strand. Specifically, the twist pitch calculation unit 82 calculates at least one of the first index, the second index, the third index, and the fourth index.

[0165] In step S206, the histogram creation unit 83 of the index calculation device 80 creates a histogram of the calculated index.

[0166] In step S207, the statistical calculation unit 84 of the index calculation device 80 calculates the statistical quantity of the calculated index using a histogram.

[0167] Referring to Figure 40, the procedure for detecting the center position in step S202 will be explained. In step S301, the binarization unit 71 binarizes the processed image to generate a first binarized image consisting of a plurality of pixels with a first pixel value (for example, 1) or a second pixel value (for example, 0).

[0168] In step S302, the segmentation unit 72 detects each region in which the pixels of the first pixel value of the first binarized image are consecutive.

[0169] In step S303, the distance conversion unit 73 calculates the shortest distance from the pixel with the first pixel value (corresponding to a candidate region where a strand exists) in each region to the pixel with the second pixel value (corresponding to a candidate region where a strand does not exist).

[0170] In step S304, the distance conversion unit 73 generates a distance conversion image for each region, where the shortest distance is the pixel value.

[0171] In step S305, the wire region detection unit 74 sets the threshold value TH to an initial value for each region.

[0172] In step S306, the strand region detection unit 74 binarizes the distance conversion image for each region based on a threshold TH to generate a binarized distance conversion image consisting of multiple pixels with a first pixel value (1) or a second pixel value (0).

[0173] In step S307, the strand region detection unit 74 counts the number of subregions in each region where pixels having a first pixel value (1) included in the binarized distance-converted image are continuous.

[0174] In step S308, if the count value changes to a decrease, the process proceeds to step S310; if the count value does not change to a decrease, the process proceeds to step S309.

[0175] In step S309, the wire region detection unit 74 increases the threshold TH by the step width ΔdTH. Then, the process returns to step S305.

[0176] In step S310, the wire region detection unit 74 calculates, for each region, the ratio R of the number of threshold values ​​TH that result in the largest count value among multiple count values ​​to the total number of set threshold values ​​TH.

[0177] In step S311, if the ratio R for each region is equal to or greater than the reference value, the process proceeds to step S312.

[0178] In step S312, the strand region detection unit 74 determines the number of strands for each region by the number of subregions generated by one or more thresholds that result in the maximum count value (for example, the maximum threshold, the minimum threshold, or the average of both).

[0179] In step S313, the center detection unit 75 detects the position of any pixel among the pixels in one or more subregions as the center position of the strand. For example, the center detection unit 75 may detect the position of the pixel with the maximum pixel value in the distance-transformed image, or the position of the pixel at the centroid of the subregion, as the center position of the strand.

[0180] In the above embodiments, the multi-core cable quality evaluation system 40 generates a processed image from a tomographic image of a cross-section perpendicular to the longitudinal direction of a multi-core cable containing multiple strands, such that the pixels at the center of the strands in the tomographic image are bright, and the pixels gradually darken toward the radial direction of the cross-section of the strands. The quality of the multi-core cable is then evaluated using the processed image. However, the objects whose quality can be evaluated are not limited to multi-core cables. More precisely, the method described above can be applied to any object whose existence can be recognized by some change in brightness in the original tomographic image of the cross-section. The quality evaluation system 40 may also generate a processed image from a tomographic image of a cross-section perpendicular to the longitudinal direction of an object containing multiple round rod-shaped materials or a tomographic image of a cross-section perpendicular to the longitudinal direction of an object containing multiple round rod-shaped hollows, such that the pixels at the center of the material or hollow cross-section in the tomographic image are bright, and the pixels gradually darken toward the radial direction of the material or hollow cross-section. The quality of the object is then evaluated using the processed image.

[0181] Referring to Figure 41, the configuration of the multi-core cable quality evaluation system 40 when its functions are implemented using software will be described. The multi-core cable quality evaluation system 40 includes a processor 502 and a memory 501 connected to a bus 503. The functions described in the above embodiment are performed when the processor 502 executes a program stored in the memory 501.

[0182] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The basic scope of this disclosure is indicated by the claims rather than the embodiments described above, and all modifications within the scope of the claims are intended to be included in the meaning of equivalents and within the scope.

[0183] 1 Multi-core cable, 2, 2a, 2b Insulated conductor, 4 Conductor section, 5 Strand, 21 Conductor, 21a Outer conductor, 21a1 First outer conductor, 21a2 Second outer conductor, 21a3 Third outer conductor, 21a4 Fourth outer conductor, 21a5 Fifth outer conductor, 21a6 Sixth outer conductor, 21b Inner conductor, 22 Insulation section, 30 Sheath layer, 31 Outer sheath layer, 32 Inner sheath layer, 32a Outer perimeter, 100, 101, 102, 103, 104, 105 Rollers, 200, 201, 202, 203, 210 Supply reel, 211 Cradle, 300, 310 Die plate, 400, 410 Die, 500 Twisting roller, 600 Arm, 700 801 Winding section, Conductor supply reel, 802 Twisting section, 803 Inner sheath layer coating section, 804 Outer sheath layer coating section, 803a, 804a Storage section, 805 Cooling section, 806 Winding section, A, B Cross section, C Coefficient of linear expansion, E Modulus of elasticity, C × E Product, F1, F2 Frequency, L Pass line ratio, L1, L2, L10 Pass line, O01A, O01B, O72A, O72B, OA, OB Center of gravity position, P Twist pitch, A1 Central axis, VA, VB Relative coordinate system, W Width at half maximum, dθ Displacement, l Distance between cross section, 40 Quality evaluation system, 41 Learned model storage device, 50 Processed image generation device, 51, 61 Data acquisition unit, 52 Image generation unit, 60 Learning device, 62 63 Training data generation unit, 70 Model generation unit, 70 Center position detection device, 71 Binarization unit, 72 Segmentation unit, 73 Distance conversion unit, 74 Strand region detection unit, 75 Center detection unit, 80 Index calculation device, 81 Trajectory data generation unit, 82 Twist pitch calculation unit, 83 Histogram creation unit, 84 Statistical calculation unit, 501 Memory, 502 Processor, 503 Bus, CV1, CV2, CV3 Convolutional layers, FL1, FL2, FL3 Filters.

Claims

1. A covered conductor comprising a conductor portion and a covering portion covering the conductor portion, wherein the conductor portion is composed of a plurality of strands twisted together, the conductor portion includes a conductor adjacent to the covering portion, and in the conductor, if the direction in which the conductor portion extends is the Z direction, the half-width calculated from the histogram of the twist pitch in the Z direction of the conductor is 0.6 times or less the median value calculated from the histogram of the twist pitch, and when considering two cross-sections perpendicular to the Z direction and separated from each other in the Z direction, the twist pitch is calculated based on the amount of displacement of the relative positions of the strands between the two cross-sections in a relative coordinate system with the centroid of the conductor in the cross-sections as the origin, the distance between the two cross-sections is 1 mm, and the histogram is created for a portion of the conductor with a length of 50 mm in the Z direction.

2. The insulated conductor according to claim 1, wherein when the median is 26.0 mm or more, the half-width is 15.5 mm or less.

3. The insulated conductor according to claim 1 or 2, wherein the histogram is created from the amount of relative positional displacement in each of the plurality of strands.

4. The insulated conductor according to any one of claims 1 to 3, wherein the conductor comprises a first outer conductor and a second outer conductor, and the histogram is created from the twist pitch of the first outer conductor and the second outer conductor, respectively.

5. The insulated wire according to any one of claims 1 to 4, wherein the conductor portion includes an inner conductor, and the inner conductor is arranged away from the covering portion.

6. A multi-core cable comprising a covered conductor according to any one of claims 1 to 5.

Images