JP3218789B2 - Hollow fiber module inspection equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an inspection apparatus for a hollow fiber module used in a dialyzer for exchanging a substance between blood passing through a hollow fiber and dialysate passing outside the hollow fiber. In particular, a defect in which the resin agent has not penetrated to a required position (hereinafter referred to as an impermeability defect) and a resin agent that solidifies the end of the yarn bundle of the hollow fiber enter the hollow fiber from the end thereof. The present invention relates to an apparatus for inspecting a hollow fiber module suitable for detecting a defect (hereinafter, referred to as an imperfect yarn defect).

[0002]

2. Description of the Related Art As is well known, a hemodialyzer separates blood and dialysate by a semipermeable membrane, and the blood and dialysate are introduced into the dialyzer by a pressure and concentration difference between the blood and the dialysate. By exchanging substances between dialysates, impurities in blood are removed and the blood is purified.

As such a hemodialyzer, as shown in FIG. 25, a plurality of hollow semi-permeable membranes having both ends open in a substantially cylindrical transparent container A having both ends open and symmetrical. There is a type using a hollow fiber module M in which a substantially cylindrical yarn bundle B composed of yarns is disposed. The hollow fiber module M is made of a thermosetting resin material such as urethane resin which is in close contact with the inner peripheral surface of the container A and the outer peripheral surface of the hollow fiber constituting the yarn bundle B at both ends of the container A. It has resin material layers C1, C2 having a thickness.
C2 separates the blood chamber inside the hollow fiber from the dialysate chamber outside the hollow fiber. That is, in the dialyzer using the hollow fiber module M as shown in FIG. 25, a substance exchange is performed between blood passing through the hollow fiber and dialysate passing outside the hollow fiber in the container A.

A hemodialyzer illustrated in FIG. 25 is generally configured as follows.

[0005] First, a cylindrical yarn bundle B in which a large number of hollow fibers are bundled is placed in the container A as shown in FIG. Next, a process for forming resin agent layers C1 and C2 on both end surfaces of the container A is performed, and FIG.
And the end caps D1 and D2 are attached to both ends of the container A. A nozzle d1 previously formed on these end caps D1 and D2
, D2 are a blood introducing part and a blood discharging part, and nozzles d3, d4 formed in the container A in advance are a dialysate discharging part and a dialysate introducing part.

The hollow fiber module shown in FIG. 25 is constructed as described above. In this hollow fiber module, in the step of forming the resin agent layers C1 and C2, the hollow fiber module is placed between the single yarns of the hollow fiber bundle. A defect that the resin agent does not sufficiently penetrate (hereinafter referred to as impervious defect) and a defect in which the hollow hole of the hollow fiber is closed by the resin agent (hereinafter referred to as impervious yarn defect) may occur. Since such impervious defects and impervious yarn defects cause various problems such as reducing the therapeutic effect in artificial dialysis treatment, it is necessary to inspect each hollow fiber module in the manufacturing process.

The above impervious defects are apt to occur at the center of the end surfaces of the resin agent layers C1 and C2 where the hollow fibers are densely packed. The degree of is strong. On the other hand, in the defective portion due to the imperfect yarn defect, the end of the hollow fiber is solidified with the resin material and the open end of the hollow fiber is also closed, so that the degree of gloss of the defective portion on the end surfaces of the resin material layers C1 and C2 is normal. Stronger than part. Conventionally, such impervious defects and impervious yarn defects are inspected by an inspector visually identifying the degree of whiteness of the end surface of the resin material layer and the degree of partial gloss of the yarn bundle. Was.
However, it is expected that the accuracy of such visual inspection does not always satisfy a certain level due to the inspection ability of the inspector and the physical condition at that time.

In the visual inspection, a great deal of attention and a great deal of inspection time are taken to prevent the above-described inspection errors from occurring. This is a hindrance to shortening the time. That is, in order to completely automate the manufacturing process and reduce the cost, it is necessary to automatically detect the impervious defect and the non-permeability defect as described above.

As a general method for automatically detecting the impervious defects and impervious yarn defects in the hollow fiber module without visual observation, the end faces of the resin material layers C1 and C2 irradiated with light are imaged by a television camera. Then, a method of processing and determining the image signal can be considered. Japanese Patent Application Laid-Open No. 61-20 / 1986 discloses a method for evaluating the surface condition by imaging an inspection target surface irradiated with light and processing the image signal.
There is a method disclosed in Japanese Patent Publication No. 1105. According to the method disclosed in this publication, the surface to be inspected is irradiated with light obliquely,
By binarizing the image signal of the inspection target surface with a predetermined threshold value, the area of a portion corresponding to the shadow of the inspection target surface is calculated, and from the magnitude of the calculated area, the inspection target surface is calculated. To evaluate the surface condition.

[0010]

However, the method disclosed in the above publication has the following problems when detecting the above-described impervious defects and impervious yarn defects.

First, in the inspection for impervious defects, it is necessary to uniformly illuminate the entire end surfaces of the resin material layers C1 and C2, which are the inspection target surfaces. The method of irradiating light obliquely has a problem that the luminance distribution on the inspection target end face becomes non-uniform because the distance from the light source to the inspection target end face is not constant. In the case of irradiating the light obliquely as described above, the luminance distribution can be made uniform by moving the light source away from the inspection target surface, but at the same time, a problem arises that a high luminance point is generated by regular reflection. This is a problem that also occurs when a light source is placed at a position orthogonal to the end face of the inspection object in order to make the luminance distribution uniform. On the other hand, when a method of obliquely irradiating the inspection target end face is adopted, it is also conceivable to divide the inspection target end face into a plurality of regions in which the luminance distribution is within a certain range, and to perform inspection by changing conditions for each of the divided regions. However, such a method is not practical because the above-mentioned condition setting is complicated and requires an examination time.

As an irradiating means for uniformly irradiating the surface to be inspected and generating no local high-brightness points, a plurality of light sources 701 as shown in FIG. 27 and a diffusing plate 702 whose shape and characteristics are devised are used. Although a combination of light irradiating means 700 is known, such a light irradiating means 700 has a drawback that the design is delicate, complicated, and expensive.

By the way, in order to uniformly irradiate the end face of a generally circular hollow fiber module, it is easily presumed that the shape of the light source is suitably circular or ring-shaped. However, in the case of circular or ring illumination, it is unavoidable that the illumination effect on the central axis of the ring is usually strongest and weak in the radial direction due to the geometrical effect. Further, when the ring diameter is small or when the ring is arranged far away, a high luminance point due to regular reflection occurs as in the case described above. further,
Irradiation unevenness in the radial direction in the ring illumination is alleviated even if a diffusive cylindrical reflector is provided in front of the ring illumination, and the illumination effect for accurately grasping the impervious defect as described above cannot be obtained.

On the other hand, as described above, the resin agent layers C 1,
The defective portion at the end face of C2 is glossier than the normal portion. As described above, the most common method of detecting a portion having a high degree of gloss on a surface is a method of detecting the degree of specular reflection light of oblique illumination. However, with such a method, it is not possible to accurately detect a thread break defect for the following reasons (1) to (3). (1) In a module having a thick hollow fiber, the difference between the normal portion and the defective portion of the impervious yarn is not remarkable. (2) In the step of forming the end faces of the resin agent layers C1 and C2, there may be a case where it is difficult to make a difference between the normal portion and the defective yarn defect on the end faces. (3) When the non-threading defect occurs in the outer peripheral portion of the yarn bundle portion, it is indistinguishable from the portion of only the resin agent outside the yarn bundle portion.

The inventor of the present invention has observed the defective yarn portions at the end faces of the resin material layers C1 and C2. As a result, in the defective yarn portions, the resin material that closes the open end of the hollow fiber and the defective yarn defects The present inventors have found that a white small lump, which is a group of air bubbles in which microfine air bubbles are dense, is present in the resin agent for solidifying the hollow fibers. The small lumps of the white air bubbles may not only enter the surface of the end face but also enter the inside of the resin agent layer.
That is, in order to detect such white air bubbles, the air bubbles on the end face surface are detected, and the white air bubbles inside the resin material layer, that is, the white bubbles that are in the periphery of the yarn bundle and lie below the end face surface It is necessary to grasp the bubbles. For this reason, it is necessary to consider not only regular reflection but also uniform illumination of the end face as in the case of the above-described impervious defect detection, and also consider illumination inside the resin layer.

[0016]

SUMMARY OF THE INVENTION A first object of the present invention is to provide an apparatus for inspecting the quality of a hollow fiber module having a resin layer in which the ends of a plurality of hollow fibers are solidified with a resin, particularly the quality based on the presence or absence of impervious defects. It is an object of the present invention to provide a hollow fiber module inspection apparatus capable of accurately and easily detecting an impervious defect in a resin layer of a hollow fiber module by uniformly illuminating a surface to be inspected by a simple method. .

A second object of the present invention is an apparatus for inspecting the quality of a hollow fiber module having a resin material layer in which the ends of a plurality of hollow fibers are solidified with a resin material, and in particular, the quality of the hollow fiber module based on the presence or absence of a defective yarn. It is another object of the present invention to provide a hollow fiber module inspection apparatus capable of accurately and easily detecting a non-through yarn defect in a resin layer of a hollow fiber module by devising a lighting means for a surface to be inspected.

[0018]

According to a first aspect of the present invention, there is provided an apparatus for inspecting a hollow fiber module having a resin material layer in which ends of a plurality of hollow fibers are solidified with a resin material. Support means for supporting the hollow fiber module;
A ring-shaped light emitting portion, which is disposed substantially coaxially with the module at a predetermined distance from the end face of the hollow fiber module and whose outer diameter is equal to or larger than the outer diameter of the module, faces the end face of the hollow fiber module; A light metering means for measuring the amount of light in each subdivided area of the end face of the hollow fiber module at a position where regular reflection light of the light emitted from the light emitting portion at the end face does not enter; and the hollow fiber Light intensity measurement at any one selected point on the end face of the module or 2
A threshold value is calculated by adding an arbitrary offset amount to the average light measurement amount at the selected point, and a binarized signal obtained by binarizing the light measurement amount of each area measured by the light measuring means with the threshold value is output. It is provided with a binarization processing means and a determination means for determining whether or not the hollow fiber module is defective based on the processing result of the binarization processing means.

According to a second aspect of the present invention, there is provided an apparatus for inspecting a hollow fiber module having a resin material layer in which ends of a plurality of hollow fibers are solidified with a resin material, the hollow fiber module being supported. A supporting means, a ring-shaped first light emitting portion that is disposed substantially coaxially with the module at a position separated from the end face of the hollow fiber module by a predetermined distance and whose outer diameter is equal to or larger than the outer diameter of the module; A second ring-shaped member surrounding the resin material layer to irradiate the hollow fiber module with light from the side at a position corresponding to the filled resin material.
And a subdivision of the end face of the hollow fiber module at a position facing the end face of the hollow fiber module and not receiving regular reflection light of the light emitted from the first light output section at the end face. A photometric means for measuring the photometric light quantity for each of the formed regions, and adding an arbitrary offset amount to the photometric light quantity at any one selected point or the average photometric light quantity at two or more selected points on the end face of the hollow fiber module. A binarization processing unit that calculates a threshold value and outputs a binarization signal obtained by binarizing the light measurement amount of each area measured by the photometry unit with the threshold value; and a processing result of the binarization processing unit. Determining means for determining whether or not the hollow fiber module is defective based on the first light emitting unit and the second light emitting unit,
The distribution of the measured light amount in the hollow fiber module in which the end of the hollow fiber is not clogged by the resin agent, which is imaged by the imaging camera, irradiates light that becomes flat at least in the hollow fiber region on the end face. .

A third aspect of the present invention provides the inspection apparatus for a hollow fiber module according to the first or second aspect, further comprising a classification unit for classifying the defective hollow fiber module based on a result of the determination by the determination unit.

[0021]

According to the hollow fiber module inspection device of the present invention, the end face of the hollow fiber module supported by the support means is substantially equal to the light quantity of the peripheral part by the ring-shaped light emitting part. Alternatively, irradiation is performed with a light amount distribution that is larger than the light amount in the peripheral portion. The binarization processing means calculates a threshold value individually for each hollow fiber module to be inspected, and binarizes the light amount on the end face of the hollow fiber module to be inspected for each subdivided area.
It is determined whether or not the hollow fiber module is defective based on the binarized signal output from the binarization processing means.

According to the hollow fiber module inspection apparatus of the present invention, the end of the hollow fiber is not clogged with the resin agent by irradiating light from the first and second light emitting portions. The light amount distribution on the end face of the module is made substantially flat. The binarization processing means calculates a threshold value individually for each hollow fiber module to be inspected, and binarizes the light amount on the end face of the hollow fiber module to be inspected for each subdivided area. It is determined whether or not the hollow fiber module is defective based on the value signal.

According to the hollow fiber module inspection apparatus of the third aspect, the defective hollow fiber module can be automatically removed from the production line based on the determination result by the determination means.

[0024]

FIG. 1 shows an embodiment of a hollow fiber module inspection apparatus (hereinafter simply referred to as an inspection apparatus) according to the present invention. The inspection device in FIG. 1 includes a support device 1 for a hollow fiber module M, a first light irradiation device 2A, and a second light irradiation device 2A.
B, third light irradiating means 2D, fourth light irradiating means 2E, first light measuring device 3A, second light measuring device 3D, third light measuring device 3E, abnormality detection signal forming section 4, discriminating device 5 , Display unit 6 and operation command unit 7.

The inspection apparatus constructed as described above is used for the I.P. Impervious defects, II. Y. yarn defect, III. Imperfect yarn defect, IV. Peeling defect, V. Inspect for foam defects. The configuration and operation of each unit will be described below for each of these inspection items.

I. <Inspection of Impervious Defect> Since the impervious defect has already been described, a description thereof will be omitted.

As shown in FIG. 2, the support device 1 has a hollow fiber module M mounted thereon, and the axis C0 of the hollow fiber module M is the optical center of the condenser lens 32A and the television camera 31A to be described later. The inclination of the hollow fiber module M can be adjusted so as to be parallel to the axis 32Aa, and the hollow fiber module M is moved up, down, left, and right in accordance with the outer diameter that is different for each product type. The center C0 can be aligned with the optical center axis 32Aa.

In the inspection for the impervious defect, a first light irradiation device 2A is used as a light irradiation device. As shown in FIG. 2, the first light irradiation device 2A includes a first light source 201A.
And a first ring-shaped emission section 202A.

The first light source 201A is controlled by a ring-shaped light emitting section 2 based on a command given by an operation command section 7 described later.
On / off of emission of light from 02A and adjustment of light intensity are performed.

As shown in FIG.
A is preferably a type having a large light emitting area with diffusibility. In this embodiment, a commercially available annular fluorescent lamp is used. The ring-shaped emission section 202A uniformly irradiates the end face of the hollow fiber module M, and prevents the high-intensity point from being generated in an image of the end face of the hollow fiber module M captured by the CCD camera 31 described later. Is provided at a position that satisfies the following conditions.

[0031]

D> (1 + L / L0) .D

Here, d is the inner diameter of the ring-shaped emitting portion 202A (the diameter of the inscribed circle when the ring-shaped emitting portion 202A is polygonal), and L is from the end face of the hollow fiber module M to the ring-shaped emitting portion 202A. , L0 is the distance between the end face and the CCD camera 31, and D is the diameter of the end face of the hollow fiber module M. When the inner diameter d, the distance L0 and the diameter D are constant, the maximum value L1 of the distance L from the end face of the hollow fiber module M to the ring-shaped emission part 202A that satisfies the above equation (1).
Is

[0033]

L1 = (d-D) .L0 / D

## EQU1 ## In the hollow fiber module M having a circular cross section, the ring-shaped emission section 202A and the hollow fiber module M
Assuming that the distance L between the end faces is the maximum value L1 obtained by Equation 2, FIG.
As shown in (a), the light amount distribution in the diameter direction at the end face has a high mountain shape near the center O of the end face.
When the distance L is smaller than the maximum value L1, FIG.
As shown in (b) and (c), when the peak value of the peak near the center O decreases and the distance L is further reduced, FIG.
As shown in (1), the distribution of the light amount decreases at the center O of the end face.

In the light quantity distribution as shown in FIG. 4A, a relatively large difference in light quantity occurs between the center of the yarn bundle of the hollow fibers at the end face and the periphery of the yarn bundle. There is a problem that the detection accuracy is not uniform between the central portion and the peripheral portion of the bundle. That is, the light quantity distribution shown in FIG. 4A lacks uniformity of illumination over the entire end face, and is not preferably used for detecting impervious defects. FIG.
As shown in (d), the light quantity distribution falling at the center O of the end face is not suitable for detecting impervious defects because the light quantity distribution at the center of the end face at the center is extremely low.

For the detection of impervious defects in a hollow fiber module having an ideal end face perpendicular to the axis C0, a light quantity distribution as shown in FIG. preferable. However, in reality, the end face may be distorted depending on the hollow fiber module,
In order to detect such impervious defects in the hollow fiber module M, the distribution is slightly increased at a position corresponding to the center of the end face in order to reliably avoid a drop in the light amount distribution.
The light quantity distribution shown in FIG.

As described above, the ring-shaped emission section 202A
In order to uniformly illuminate the end face C1, it is necessary to use a type having a light-emitting surface that is more diffusive and has a larger light-emitting area than a light-emitting surface having directivity.
For this reason, a predetermined object can be achieved by using an annular fluorescent lamp as in the embodiment. However, fluorescent lights
It is not suitable for adjusting the intensity, and is not suitable for repeated blinking for switching inspection items. For this reason, instead of the fluorescent lamp, FIG.
As shown in the figure, a ring-shaped main body 202Aa whose inner diameter is larger than the outer diameter of the end of the hollow fiber module M, and the main body 202A
A ring-shaped fiber light guide composed of a plurality of optical fibers 202Ab accommodated in a may be used.
In the case of FIG. 5, a lamp is used as the first light source 201A, and a light beam emitted from the first light source 201A is transmitted by the optical fiber bundle 203A to the ring-shaped emission portion 202A.
Lead to the optical fiber 202Ab of the main body 202A.
Light can be applied to the end face of the hollow fiber module M from the end face a.

The distal end face of each optical fiber 202Ab of the ring-shaped fiber light guide is annularly arranged on one end face of the main body 202Aa along the outer peripheral edge of the main body 202Aa. In this case, in the case where the emission direction, that is, the tip of the optical fiber 202Ab is arranged in the direction perpendicular to the end face of the hollow fiber module M, the ring-shaped emission section is located at a position satisfying the expression 1 in order to avoid regular reflection. When 202A is placed, the amount of light incident on the end face of the hollow fiber module M becomes weak. Therefore, the fiber light guide as shown in FIG.
In this case, the tip of the optical fiber 202Ab is preferably inclined at, for example, about 45 ° toward the center O of the end face of the hollow fiber module M.

It is desirable that the optical fiber 202Ab has a wide exit aperture angle. However, even a wide-angle type optical fiber has an aperture angle of only 120 °, which is insufficient in diffusivity as compared with a fluorescent lamp, as shown in FIG. , Ring-shaped emitting section 202A
It is preferable to dispose the diffusing plate 204A in front of the substrate.

In the inspection for the impervious defect, the first photometric device 3A is used as a photometric device. The first photometric device 3A has, for example, 512 (horizontal) × 480 (vertical) in the field of view.
A known camera 2 having a camera body 31 having a photosensor group divided into minute regions each of which is arranged to correspond to each of these minute regions, and a condenser lens 32 for condensing light on the photosensor group. It is a three-dimensional television camera (see FIG. 2). As described above, by operating the moving mechanism of the supporting device 1, the axis C0 of the hollow fiber module M can be made to coincide with the optical center axis 32Aa of the condenser lens 32A. The apparatus 3 can photograph the end face of the resin layer C1 of the hollow fiber module M. The first photometric device 3 takes an image of the end face of the resin material layer C1 irradiated with the light from the first ring-shaped emission part 202A, and displays a first light quantity representing a light amount of a minute area corresponding to each photosensor. Outputs analog image signals. This first
Is supplied to the abnormality detection signal forming section 4.

The abnormality detection signal forming section 4 includes an A / D converter 401, an image storage section 402, a binarization processing section 403, and a binarized image display section 404.

The A / D converter 401 converts the first analog image signal supplied from the first photometric device 3A into a first digital image signal and sends it to the image storage unit 402.

The image storage section 402 temporarily stores the first digital image signal.

The binarization processing section 403 binarizes the first digital image signal with a predetermined threshold value B1P to form a first binarized image signal. That is, this binarization processing unit 4
03 replaces the image signal (the part shown by the broken line in FIG. 7) having a higher level than the threshold value B1P with a white level,
An image signal having a lower level is replaced with a black level (see FIG. 7).

Further, the binarization processing section 403 sets the threshold value B1P by the following method.

The amount of light at the end face of the hollow fiber module M illuminated by the ring-shaped emission section 202A is a difference between lots of the resin material, the deterioration of sharpness of a molding machine for molding the end face, or the first light source 201A. Due to the deterioration of the lamp and the like, it is difficult to avoid variations among the hollow fiber modules. In order to adapt to this phenomenon, the binarization processing unit 403 according to the present embodiment uses the first digital image signal stored in the image storage unit 402 as shown in FIG. (J-
1) The light amount B1i (i = 1 to 1) of the minute areas p1 to p (j-1)
j-1) is measured. Where R is the hollow fiber module M
Is the inner diameter of the main body case. Further, k is 1>k> 0, and is set arbitrarily regardless of the type.

The measured light quantity B1i of the end face of the hollow fiber module M measured as described above (where i = 0, 1,..., J−
In the case of 1), the value B0 at the center of the end face usually becomes maximum as in the case of FIG. 4B described above. Further, when the impervious defect occurs at the center of the end face, the light intensity at the center becomes higher than that of a normal product. Also, normally, as shown in FIG.
Even in the case of a light amount distribution in which the light amount in the central portion is slightly higher as shown in the figure, if an impervious defect occurs, even if the occurrence location is the peripheral portion of the end face, the light amount of the portion corresponding to the impervious defect is It becomes higher than the central peak of the light quantity distribution. In the binarization processing unit 403, in order to understand that the peak of the light amount distribution on the end face rises high due to the impervious defect, it is possible to separate the high light amount part which is not in the case of a normal product by the equation (3). A binary level, that is, a threshold value B1P is calculated.

[0048]

## EQU3 ## B1P = bar B1 + .DELTA.B1

Here, bar B1 is the average value of (j-2) light quantity data excluding the maximum and minimum values of the measured light quantity at point j, and ΔB1 is the constant of the binarization level depending on the type. Offset. The constant offset amount ΔB1 is basically determined by estimating the light amount distribution on the end face of a normal product in accordance with the chromaticity difference between lots of the resin material and the temporal change of the light source. Here, the slope of the light amount distribution is determined as being substantially constant depending on the product type. The method of determining the bar B1 may be determined by calculating from the range (maximum-minimum) of the light intensity value distribution for each minute area in the field of view photographed by the television camera or the standard deviation value of this distribution. A generally known method may be adopted. As described above, the bar B1 is preferably obtained as an average value or a deviation value from the viewpoint of improving detection accuracy. However, the light amount B10 at the center of the end face of the hollow fiber module M may be directly used as the bar B1. .

The threshold value B1P obtained as described above is calculated as shown in FIG.
As shown by a solid line and an alternate long and short dash line, the hollow fiber module slightly rises and falls according to the overall color difference between the individual hollow fiber modules.

The binarized image display unit 404 has an image obtained from the first binarized image signal replaced with the white level and the black level by the binarization processing unit 403, that is, a high level due to the presence of the impervious part. An image in which only the light amount is displayed in white is displayed.

The discriminating device 5 includes a storage unit 501 and an operation judging unit 502. The discriminating device 5 operates as follows to detect an impervious defect.

The calculation judging section 502 calculates a first high light amount total area based on the first binarized image signal displayed in white on the binarized image display section 404. This total area is one of the photo sensors of the camera body 31.
The output image signal corresponds to the number of photo sensors in which the white-level image signal has been replaced. Also, the operation determination unit 502
Reads out the area reference value for the impervious defect inspection of the part number input from a data input device not described from among the reference values for each part number of the hollow fiber module M stored in the storage unit 501 in advance.

The operation determining unit 502 compares the binarized total area of the first high light intensity portion with the area reference value for the impervious defect inspection to determine the quality of the hollow fiber module. Here, when the comparison result satisfies the reference value> total area, the hollow fiber module M is determined to be non-defective with respect to the impervious defect inspection, and when the reference value ≦ total area, the hollow fiber module M is determined to be defective. .

The display unit 6 includes a CRT display and a printer, and displays the result of the pass / fail judgment on the impervious defect inspection of the hollow fiber module.

The operation instructing unit 7 receives, from the calculation judging unit 502, the judgment results of the yarn imperfection defect inspection, the non-through yarn inspection, the peeling inspection, and the foaming inspection, which will be described later, in addition to the judgment results of the impervious defect inspection described above. Thereafter, the hollow fiber module transfer device 8 shown in FIG. 8 is operated to remove the hollow fiber module M mounted on the support device 1 and mount a new module on the support device 1.

In FIG. 8, the hollow fiber module is attached to and detached from the supply line 81 from the supply line 81 to the support device 1 and then to the next step transfer section 82, and the next step transfer section 82 or the defective module collection section 86 based on the determination result of non-defective / defective products. Sorting into
The automatic hand 80 automatically performs the operation according to a command from the operation command unit 7. Therefore, the auto hand 80 can move along the guide 83 as shown by the two-dot chain line. The operation command unit 7 also determines the axis C0 of the hollow fiber module mounted on the support device 1 based on the information on the product number of the hollow fiber module M input by the input device.
A moving mechanism (not shown) for the support device 1 is controlled so as to coincide with the optical center axis 2.

The operation command section 7 further has a separating function of separating the hollow fiber module M removed from the supporting device 1 into a non-defective product and a non-defective product based on the determination result of the calculation determining portion 502. That is, the operation command unit 7
If the determination result indicates that the hollow fiber module M mounted on the support device 1 is a non-defective product, the hollow fiber module transfer device 8 shown in FIG. A command to move the module M to the next process transfer unit 82 is issued. If the comparison result indicates that the hollow fiber module M is defective, the hollow fiber module transfer device 8 is sent from the support device 1 to the hollow fiber module transfer device 8. A command to move the detached hollow fiber module M to the defective hollow fiber module collection unit 86 is issued.

With the inspection apparatus configured as described above, a hollow fiber module having an impervious defect as described above can be excluded as a defective product.

In the above-described embodiment, the operation determining unit 502
Only whether or not the total area of the first high light intensity portion calculated in step (1) exceeds the reference value read from the storage unit 501 is used as a condition for determining whether the hollow fiber module is a good product or a bad product. However, in still another inspection apparatus for a hollow fiber module according to the present invention, instead of the method using the area as an index, the shape of a high-light amount portion displayed in white on the binarized image display unit 404 is reduced. The feature, that is, the feature such as the length of the high light amount portion may be quantified and the feature amount may be compared with a reference amount.

II. <Inspection of Yarn Defects> Yarn defect is caused by the fact that a part of the hollow fiber constituting the yarn bundle B is in close contact with the resin agent layers C1 and C2, and is gathered around the center of the container A. This is a defect that is scattered around the yarn bundle B, bent, and protrudes from the yarn bundle B. The inspection apparatus of the hollow fiber module of FIG. 1 can also inspect such a yarn disorder defect.

The structure and operation of the support device 1 for inspecting a yarn disorder defect are the same as those of the above-described inspection of the impervious defect, and therefore, detailed description thereof will be omitted.

In the inspection for the yarn disorder defect, the second light irradiation device 2B is used as the light irradiation device. As shown in FIG. 2, the second light irradiation device 2B includes a second light source 201B.
And a ring-shaped emission part 202B.

The second light source 201B turns on / off the emission of the light from the ring-shaped emission section 202B and adjusts the light intensity based on a command given from the operation command section 7.

As shown in FIG. 9, the second ring-shaped emitting portion 202B includes a ring-shaped main body 202Ba having an inner diameter φ larger than the outer diameter of the end portion of the hollow fiber module M, and a main body 202Ba.
This is a ring-shaped fiber light guide including a plurality of optical fibers 202Bb housed in Ba. The tip surface of each optical fiber 202Bb is arranged along the inner peripheral surface of the main body 202Ba, and
It faces the center of a. In FIG. 2, the optical fibers 202Bb are arranged at intervals in the circumferential direction for the sake of simplicity, but it is preferable that the optical fibers 202Bb are arranged as densely as possible in the circumferential direction. In any case, the ring-shaped emission section 202B uniformly illuminates the side surface of the end of the hollow fiber module M in the circumferential direction. The plurality of optical fibers 202Bb extend outside the main body 202Ba via an optical fiber connector or the like, and are each optically connected to the second light source 201B.

The end of the hollow fiber module M can be inserted into the inner peripheral side of the second ring-shaped emission part 202B by moving in the axial direction of the hollow fiber module M.
In this embodiment, the second ring-shaped emission portion 202B
Is an optical fiber 202B as shown by a solid line in FIG.
The emission end face b is fixed at the axially central portion X of the resin layer C1 of the hollow fiber module M so as to face the side surface of the resin layer C1. The inner diameter φ of the main body 202Ba is set according to the emission angle α of the optical fiber 202Bb so that the light emitted from the emission end face of the optical fiber 202Bb can irradiate the entire side surface of the resin material layer C1 in the axial direction. ing. When the side surface of the resin material layer C1 is irradiated in this manner, the resin material layer C1 is almost parallel to the axis C0.
While the hollow fiber reaching the end face of the resin layer C1 hardly generates reflected light on the end face side of the resin agent layer C1, the abnormal hollow fiber protruding from the cylindrical thread bundle B and being curved is Strong reflected light is produced toward the end face of the resin layer C1.

The fixing position of the ring-shaped light emitting portion 202B is not necessarily limited to the position shown by the solid line in FIG. The position may be shifted in the direction. However, as shown by a dashed line in FIG. 10, when the tip of the optical fiber 202Bb is located closer to the tip than the position Y that coincides with the outer end face of the resin material layer C1,
Since the light emitted from the optical fiber 202Bb directly irradiates the outer end face of the resin material layer C1, strong reflected light is generated at this end face, and the reflected light at this end face causes contrast with the reflected light from the abnormal hollow fiber as described above. Deteriorates and causes a detection error. Also, as shown by a two-dot chain line in FIG.
When the tip of the optical fiber 202B is located further inward from the position Z which coincides with the inner end face of the resin material layer C1, the light emitted from the optical fiber 202Bb illuminates the bubbles trapped in the resin material layer C1 and produces strong reflected light. Make
This easily causes a detection error. However, even if the tip position of the optical fiber 202Bb is an intermediate position between Z and X, a detection error similar to the above may occur.
It is preferable to set within a range of about 8 mm inward from the Y position.

It should be noted that a ring-shaped fluorescent lamp may be employed as the ring-shaped light emitting portion 202B in addition to the above-described ring-shaped fiber light guide. However, it is difficult for fluorescent lamps to avoid uneven light intensity at the base. Further, since the fluorescent lamp is omnidirectional and illuminates not only the vicinity of the end face of the resin agent layer C1 but also the trunk portion of the hollow fiber module, the detection error is more likely to occur as compared with the fiber light guide. More preferably, a light guide is used.

The photometric device used in the inspection of the yarn disorder defect is the first photometric device 3A, as in the inspection of the impervious defect. Therefore, a detailed description of the first photometric device 3A is omitted here. First photometric device 3
In the inspection for the yarn disorder defect, the resin material layer C is irradiated with light from the second ring-shaped emission portion 202B.
1 is photographed, and a second analog image signal representing the amount of light in a minute area corresponding to each photosensor is output. This second analog image signal is also provided to the abnormality detection signal forming unit 4.

The operations of the A / D converter 401 and the image storage unit 402 of the abnormality detection signal forming unit 4 when receiving the second analog image signal are the same as those in the case of the above-described inspection for impervious defects. .

In the case of the inspection for the yarn defect, that is, when the abnormality detection signal forming unit 4 inputs the second analog image signal, the binarization processing unit 403 converts the second analog image signal into an A / D signal. The converted second digital image signal is binarized at a predetermined threshold value SP to form a second binarized image signal. That is, the binarization processing unit 403 determines whether the threshold S
Replace the image signal higher than P with white level,
An image signal whose level is lower than the threshold value SP is replaced with a black level.

Further, the binarization processing section 403 sets the threshold value SP by the following method. The ring-shaped emitting section 202
When the resin material layer C1 is illuminated from the side surface by the light emitted from B toward the center, the light is scattered at the periphery of the yarn bundle B, and hardly reaches the center of the yarn bundle B. As a result, in a state where the light is emitted from the ring-shaped emission portion 202B, the light amount of the end surface of the resin agent layer C1 of the hollow fiber module M having no abnormal hollow fiber such as the hollow fiber b1 is measured by the photometric device 3. Then, the light quantity distribution in the radial direction of the hollow fiber module M becomes highest at the peripheral edge of the yarn bundle B as shown by a solid line in FIG. However, if an abnormal hollow fiber is present, as described above, the abnormal hollow fiber produces a strong reflected light directed toward the end face side of the resin material layer C1.
As shown by the dashed line in (b), an abnormal portion E where the amount of light is much higher than the periphery of the yarn bundle B is formed. The threshold S
P is set to isolate the abnormal part E. The threshold value SP is obtained by adding a predetermined offset amount ΔS to the reference light amount S corresponding to the peripheral portion of the yarn bundle B, which is the highest in the light amount distribution indicated by the solid line in FIG.

[0073]

## EQU4 ## SP = S + .DELTA.S

The offset amount ΔS is a constant determined in advance for each type of the hollow fiber module M.

On the other hand, the reference light amount S is also the offset amount Δ
Like S, it can be determined in advance for each type of hollow fiber module M. However, FIG.
It is unavoidable that the overall level of the light quantity distribution indicated by the solid line in (b) differs depending on each hollow fiber module M to be inspected mainly for the following reasons (a) to (c). (a) There is a lot-to-lot difference in the color of the resin itself forming the resin layer C1. (b) The surface state of the end of the resin material layer C1 when forming the resin material layer C1 may be different for each hollow fiber module M from a transparent one to a cloudy one. (c) The amount of emitted light decreases due to deterioration of the lamp used for the light source 201.

Therefore, it is difficult to obtain high inspection accuracy by the method of setting the reference light amount S in advance.
For this reason, in this embodiment, as shown in FIG. 1A, the light amounts Si (i = 1 to n) of n minute regions t1 to tn arranged at equal intervals on the circumference of the diameter mR are determined. Measure,
The average value of these light amounts Si is defined as the reference light amount S. Here, R is the inner diameter of the container A of the hollow fiber module M. M is a constant that can be selected from 1>m> 0 according to the type of the hollow fiber module M. It is preferable that the constant m is set so that the minute regions t1 to tn can be provided at a position corresponding to or closest to the maximum light amount generating portion at the periphery of the normal yarn bundle B. The optimum setting positions of such minute regions t1 to tn differ depending on the type of the hollow fiber module M. Therefore, the constant m needs to be determined for each type of the hollow fiber module M. A method of determining the binarization threshold value SP is to determine the binarization threshold value SP by calculating from the range (maximum-minimum) of the light amount value distribution for each minute area in the field of view photographed by the television camera or the standard deviation value of this distribution. For example, a generally known method may be adopted.
The method of determining using the above formula (1) is more preferable because good accuracy can be obtained while being simple.

The binarized image display unit 404 uses an image obtained from the second binarized image signal replaced with the white level and the black level by the binarization processing unit 403, that is, the reflected light from the abnormal hollow fiber. An image in which only the high light intensity portion is displayed in white is displayed.

In the discriminating device 5, based on the second binarized image signal, the operation judging section 502 calculates the total area of the second high-light portion displayed in white on the binarized image display section 404. Calculate. As in the case of the first high-light portion total area, the second high-light portion total area also corresponds to the photo sensor of the camera body 31 in which the output image signal is replaced with a white-level image signal. It corresponds to the number of sensors. In addition, the operation determination unit 502 determines, from among the reference values for each part number of the hollow fiber module M stored in the storage unit 501 in advance, the yarn disorder defect area reference value of the part number input from a data input device not described. Is read.

Further, the operation judging section 502 compares the binarized total area of the second high light quantity portion with the area reference value for the yarn defect inspection, and judges the quality of the hollow fiber module. Here, when the comparison result satisfies the reference value> total area, the hollow fiber module M is determined to be non-defective with respect to the yarn disorder defect inspection, and when the reference value ≦ total area, the hollow fiber module M is determined to be defective. .

The display unit 6 displays the result of the pass / fail judgment on the inspection of the yarn defect of the hollow fiber module in the same manner as in the case of the impermeability inspection.

The operation of the operation command unit 7 in the yarn defect inspection is the same as that in the above-described impervious defect inspection.
Here, the description is omitted.

With the inspection apparatus configured as described above, a hollow fiber module having an abnormal hollow fiber can be eliminated as a defective product. The center of the yarn bundle B of the hollow fiber module M does not always coincide with the center of the container A, and the shape of the yarn bundle B as viewed from the end face of the resin agent layer C1 is not necessarily a perfect circle. Since illumination is performed such that a high light amount portion is formed on the end surface of the resin agent layer C with the reflected light by the abnormal hollow fiber, and the high light amount portion is determined, the presence of the abnormal hollow fiber is determined. The presence of an abnormal hollow fiber is detected regardless of the discrepancy between the center of the yarn bundle B and the center of the container A and the roundness of the yarn bundle B, and the hollow fiber module having the abnormal hollow fiber can be excluded as a defective product.

In the above embodiment, the operation judging section 502
Only whether or not the second high light amount partial total area obtained in step (1) exceeds the reference value read from the storage unit 501 is used as a condition for determining whether the hollow fiber module is a good or defective product. However, in the other hollow fiber module inspection apparatus, instead of the method using the size of the area as an index, the shape characteristic of the high light amount portion displayed in white on the binarized image display unit 404, that is, this height A feature such as the length of the light amount portion may be quantified and the feature amount may be compared with a reference amount.

III. <Inspection of Non-Threading Defect> The non-threading defect has already been described in detail, and a description thereof will be omitted.

The structure and operation of the support device 1 for inspecting a non-thread defect are the same as those for the above-described inspection of a non-penetrating defect and a yarn turbulence defect, and a detailed description thereof will be omitted.

In the inspection for the defective yarn defect, a first light irradiation device 2A and a second light irradiation device 2B are used as light irradiation devices. These first and second light irradiation devices 2A, 2B
Is as described above, and the detailed description is omitted here.

The first ring-shaped light emitting portion 202A of the first light irradiation device 2A has the light amount distribution on the end face of the hollow fiber module M due to the light irradiation from the first ring-shaped light emitting portion 202A. As shown in FIG. 4 (a) or 4 (b), it is provided at a position where the light amount distribution becomes high at the center of the end face. Thereby, the first light source 201A
Illuminates not only the white deteriorated polymer near the center surface of the end surface of the hollow fiber module M, but also a group of microbubbles slightly buried inside the resin agent layer near the center surface.

On the other hand, the second ring-shaped emitting portion 202B of the second light irradiation device 2B is an annular fiber light guide similar to the case shown in FIG. However, this second
The ring-shaped emitting portion 202B is fixed at a position slightly different from that at the time of the inspection of the yarn disorder described above in the inspection of the non-through yarn defect. That is, the position of the second ring-shaped emission part 202B in the inspection of the yarn disorder defect is a position in front of the end face which irradiates the periphery of the yarn bundle near the end face of the hollow fiber module M as shown by a dashed line in FIG. And a position retracted from the end face as shown by a solid line in FIG. At the position retreated from the end face as shown by the solid line in FIG. 10, the white microbubbles at the defective portion of the impervious yarn are illuminated from behind, so that the vicinity of the end face of the hollow fiber module M where the hollow fibers are densely arranged is effectively formed. It is preferable because a light diffusion action can be obtained. However, as shown by a two-dot chain line in FIG. 10, at a position further receded from the end face than the position indicated by the solid line (a position receded by about 10 mm or more from the end face), an effect like a backlight is strong, and the end face of the hollow fiber module is The image contrast is too low to use.

The light amount distribution by irradiating light from the second ring-shaped emitting portion 202B set as described above is similar to the light amount distribution in the case of a yarn disorder defect shown in FIG. The amount of light is large around the end face of
The light intensity distribution is small at the center. by this,
The light emitted from the second ring-shaped emission section 202B illuminates a group of white microbubbles near the periphery of the yarn bundle on the end face of the hollow fiber module M.

In the inspection for the defective yarn defect,
An annular fluorescent lamp may be used as the second ring-shaped emission part 202B. However, fluorescent lamps are difficult to avoid uneven light intensity at the base, and because they are omnidirectional, they illuminate not only the end near the cut surface but also the body of the dialyzer widely, so the end surface to be inspected is Image contrast is poor. Therefore, the above-mentioned fiber light guide is more preferable as the second ring-shaped emission part 202B.

In the inspection for the defective yarn defect, based on the command given from the operation command unit 7, as shown in FIG. 12, the end face of the hollow fiber module M is irradiated by the first light irradiation means 2A. Light intensity distribution M1 and second light irradiation means 2
The first light quantity distribution M2 in the radial direction of the end face by the irradiation of B gives a substantially flat light quantity distribution M0 over the entire area of the end face.
And intensity adjustment of the second light sources 201A and 201B, and the first and second ring-shaped emission portions 202A and 202B.
Adjust the position of. Also, the first and second light sources 20
ON and OFF of 1A and 201B are performed simultaneously.

The photometric device used in the inspection of the imperfect yarn defect is the first photometric device 3A, as in the inspection of the impermeability defect and the yarn defect defect. Therefore, a detailed description of the first photometric device 3A is omitted here.
The first photometric device 3 performs the first
The end surface of the resin material layer C1 is photographed in a state where the light is irradiated from both the ring-shaped light emitting portion 202A and the second ring-shaped light emitting portion 202B, and the light amount of the minute region corresponding to each photo sensor is indicated. 3 is output. This third analog image signal is also provided to the abnormality detection signal forming unit 4.

The operation of the A / D converter 401 and the image storage unit 402 of the abnormality detection signal forming unit 4 when receiving the third analog image signal is performed in the case of the above-described inspection for the impermeability defect and the yarn disorder defect. Is the same as

In the case of the inspection for the non-threading defect, that is, when the abnormality detection signal forming unit 4 inputs the third analog image signal, the binarization processing unit 403 converts the third analog image signal into an A / D signal. The converted third digital image signal is binarized by a predetermined threshold value B2P to form a second binarized image signal. That is, the binarization processing unit 403 performs the threshold B
Replace the image signal higher than 2P with white level,
An image signal whose level is lower than the threshold value B2P is replaced with a black level.

The binarization processing section 403 sets the threshold value B2P for detecting a non-through thread defect in the same procedure as the threshold value B1P for the same reason as when the threshold value B1P is determined in the impervious defect detection. That is, from the third digital image signal stored in the image storage unit 402, (j-1) pieces of (j-1) pieces which are uniformly arranged on the light amount B20 and the radius kR of the center O of the end face of the hollow fiber module M shown in FIG. Micro area p1
The light quantity B2i (i = 1 to j-1) of .about.p (j-1) is measured.
The constant k may be the same value as that at the time of detection of the impervious defect described above, or may be a different value as needed.

The locations of the defective yarns appearing at various places in the yarn bundle portion, and the white degraded polymer present around the defective yarns has a luminance level slightly higher than that of the normal portion even when illuminated. . The binarization processing unit 403 binarizes such that even if the light amount rises high at any part of the end face of the hollow fiber module M due to the imperfect yarn defect, a portion having a high light amount which is not a normal product can be separated. The level, that is, the threshold value B1P is calculated by the equation (5).

[0097]

## EQU5 ## B2P = bar B2 + .DELTA.B2

Here, bar B2 is the average value of (j-2) light quantity data excluding the maximum value and the minimum value of the light quantity measurement values at point j, and ΔB2 is the constant of the binarization level depending on the type. Offset. Similar to the offset amount ΔB1, the constant offset amount ΔB2 is determined by estimating the light amount distribution on the end face of the normal product by adapting to the chromaticity difference between the lots of the resin agent and the aging of the light source.

As with the bar B1 in the case of the above-described impervious defect inspection, the method of determining the bar B2 is, in addition, to the range of the light amount distribution (maximum- A generally known method, such as determining from the minimum) or the standard deviation value of this distribution, may be employed. As described above, it is preferable to obtain the bar B2 as an average value or a deviation value from the viewpoint of improving the detection accuracy. However, the light amount B20 at the center of the end face of the hollow fiber module M may be directly used as the bar B2. .

The binarized image display unit 404 displays an image obtained from the third binarized image signal that has been replaced with a white level and a black level by the binarization processing unit 403, that is, a high level due to the presence of the impervious yarn. An image in which only the light amount is displayed in white is displayed.

In the discriminating device 5, based on the third binarized image signal obtained by the binarizing processing unit 403, the arithmetic judging unit 50
2 calculates a third high light amount partial total area displayed in white on the binarized image display unit 404. As in the case of the first and second high light quantity partial total areas, the third high light quantity partial total area is also replaced by a white-level image signal in the photo sensor of the camera body 31. The number of photo sensors performed. In addition, the operation determination unit 502 determines, from among the reference values for each part number of the hollow fiber module M stored in the storage unit 501 in advance, the area reference value for the broken yarn defect of the part number input from a data input device not described. Is read.

Further, the operation determining section 502 compares the binarized total area of the third high light quantity portion with the area reference value for non-through yarn defect to determine the quality of the hollow fiber module.
Here, when the comparison result satisfies the reference value> total area, the hollow fiber module M is determined to be non-defective with respect to the inspection of the defective thread, and when the reference value ≦ total area, the hollow fiber module M is determined to be defective. I do.

The display unit 6 displays the result of the pass / fail judgment on the inspection for the imperfect yarn defect of the hollow fiber module in the same manner as in the case of the impermeability defect inspection and the yarn disorder defect.

The operation of the operation command section 7 in the non-thread defect inspection is the same as that in the above-described non-penetration defect inspection and yarn disorder defect inspection, and the description thereof is omitted here.

With the inspection apparatus configured as described above, the hollow fiber module having the defective yarn defect as described above can be excluded as a defective product.

In the above-described embodiment, the operation determining unit 503
Only whether or not the third total area of the high light quantity obtained in Step 3 exceeds the reference value read from the storage unit 502 is used as a condition for determining whether the hollow fiber module is a good product or a bad product. However, in still another inspection apparatus for a hollow fiber module according to the present invention, instead of the method using the size of the area as an index, the third high light intensity portion displayed in white on the binarized image display unit 405 is used. A shape feature, that is, a feature such as a length of the third high light amount portion may be quantified and the feature amount may be compared with a reference amount.

IV. <Examination of Peeling Defect> The peeling defect is a defect in which the container A of the hollow fiber module M and the resin agent layers C1 and C2 in FIG.
The inspection apparatus for the hollow fiber module shown in FIG. 1 inspects this peeling defect by the following configuration and operation.

As shown in FIG. 13, the support device 1 shown in FIG.
1, a drive motor 102, a drive roller 103, module support rollers 104 and 105, a driven roller 106, and a belt 107. The drive motor 102 and each of the rollers 103 to 106 are mounted on a support 101. The driving roller 103 is provided with the driven roller 1.
06 and is rotationally driven by the drive motor 102.

As described above, the supporting apparatus 1 having the above-described configuration can mount the hollow fiber module M with its axis C0 kept horizontal. The hollow fiber module M mounted on the support device 1 is rotated about the axis C0 by operating the drive motor 102. The setting, starting, and stopping of the rotation speed of the drive motor 102 are controlled by a drive control unit 108, and the drive control unit 108 operates based on a command signal given by an operation command unit 7 described later. The rotation angle of the drive roller 103, which is rotationally driven by the drive motor 102, is detected by a rotation angle detector 109 such as a known optical encoder and is given to an operation command unit 7 described later.

As described above, the supporting device 1 is
The mounted hollow fiber module M can be moved up and down by a vertical movement mechanism (not shown) while keeping the axis C0 horizontal. Thus, the detection position M0 of the hollow fiber module M can be moved up and down. In the hollow fiber module M, the detection position M0 is the resin material layer C1.
Alternatively, it is the position of the interface K located at the highest position among the interfaces K between C2 and the container A.

The attachment / detachment of the hollow fiber module M to / from the support device 1 can be automatically performed by the above-described hollow fiber module transfer mechanism 8, which is schematically shown in FIG.

In the inspection for the peeling defect, the third light irradiation device 2D in FIG. 1 is used as the light irradiation device. FIG.
As shown in FIG. 3, the third light irradiation device 2D includes a third light source 201D and a slit light emitting unit 202D.

The third light source 201D turns on / off the light emission and adjusts the light intensity based on a command given from the operation command unit 7.

The slit light emitting section 202D is connected to the third light source 2 by a plurality of optical fibers 203D, for example.
01D and these optical fibers 203
After arranging D in a horizontal line, these optical fibers 203
The light emitted from the terminal of D is
The light is condensed in a slit shape in 4D, and this slit light is emitted. As shown in FIGS. 14 and 15, at the detection position M0, the slit light H is incident at an incidence angle α0 on a minute region P having a slit width ds formed by subdividing the interface K in the circumferential direction. . As described above, the drive motor 1
02 to rotate the hollow fiber module M, the slit light H emitted from the slit light emitting section 202D completely passes through the interface K between the container A and the resin material layer C1 or C2 at the detection position M0. Irradiation can be performed around the circumference, that is, over the entire area.

As shown in FIG. 14, a part of the slit light H emitted from the slit light emitting section 202 is reflected on the surface of the container A to become reflected light R11, and the other part is the reflected light R11. Invade. The slit light H that has entered the container A reaches the detection position M0. This detection position M0
If the container A and the resin agent layer C1 or C2 are separated at the interface K located at the position (2), the slit light H is reflected at this interface at a reflection angle β0 with respect to the perpendicular CL and becomes reflected light R12. The reflection angle β0 is a reflection angle when the slit light H arriving at the interface between the container A and the air layer at the incident angle α0 is reflected at this interface. If the container A and the resin material layer C1 or C2 are normally adhered at the interface K located at the detection position M0 and the two are not separated, the slit light H is mostly transmitted through the resin material through this interface. It penetrates the layer C1 or C2. The slit light H that has entered the resin material layer C1 or C2 is reflected on the surface of the hollow fiber constituting the yarn bundle B of the hollow fiber module M to become reflected light R13.

As can be easily analogized from FIG.
If the slit width ds is large, there is a disadvantage that the reflected lights R11 and R12 overlap. The slit width ds needs to be set so that such inconvenience does not occur. For example, the diameter D (see FIG. 14) at the end of the container A is usually 28 mm or more, but in any case, the slit width d
s is preferably about 1 to 2 mm. Further, it is preferable that the center in the width direction of the minute area P irradiated by the slit light H completely coincides with the detection position M0, and it is preferable that the error be kept within 1 mm. When this error increases, the intensity and angle of the reflected light R12 fluctuate, and it becomes impossible to obtain a highly accurate detection result. Further, the incident angle α0 and the reflection angle β0 are preferably set so that the reflected light R2 is the strongest, or set so that the reflected lights R12 and R13 are sufficiently separated from each other so as not to interfere. The incident angle α0 is preferably set in the range of 20 ° to 50 °.

FIG. 15 shows the minute region P of the interface K located at the detection position M0 divided into six in the axial direction. In the micro area P in FIG. 15, the hollow fiber module M
The three divided regions a, b, and c (the portions indicated by the hatched lines) on the distal end side are the peeled portions, and the other three divided regions d, e, and f are the normal bonded portions. When the slit area H shown in FIG. 15 is irradiated with the slit light H emitted from the slit light emitting section 202, the slit area H illuminating the divided areas a, b, and c is strongly reflected at the interface K and the reflected light R12 is reflected. Becomes On the other hand, the slit light H irradiating the divided areas d, e, and f almost has the interface K
And enters the resin agent layer C1 or C2, and irradiates the yarn bundle B to become the reflected light R13.

In the inspection for the peeling defect, a second photometric device 3D is used as a photometric device. The second photometer 3D is a known two-dimensional CCD camera. The second photometric device 3D receives the reflected lights R11 to R13 by the light emitted from the slit light emitting unit 202D in two dimensions, and converts the image data representing the luminance of the light of each pixel on the two dimensions. Output as a fourth analog image signal. The second photometric device 3D needs to be provided at least at a position capable of receiving the reflected light R12 reflected at the reflection angle β0 with respect to the perpendicular CL in the minute region P of the interface K located at the detection position M0, The reflection angle β0 is preferably provided at a position where the reflected light in the range of 15 ° to 70 ° is received. The fourth analog image signal output from the second photometric device 3D is also supplied to the abnormality detection signal forming unit 4 in the same manner as the first to third analog image signals.

A / D converter 401 of abnormality detection signal forming device 4 when receiving the fourth analog image signal
The operation of the image storage unit 402 is based on the aforementioned impermeability defect,
This is similar to the case of the yarn disorder defect and the non-thread defect.

In the case of the inspection for the peeling defect, that is, when the abnormality detection signal forming unit 4 inputs the fourth analog image signal, the binarization processing unit 403 executes the operation determining unit 502 described later.
Every time a binary command signal is received from the fourth analog image signal, the fourth digital image signal for each pixel obtained by A / D conversion of the fourth analog image signal is binarized by a predetermined threshold value and replaced with a white level and a black level. 4 to form a binary image signal. The binarization command signal is output each time the interface K rotates in the circumferential direction by the slit width ds. The rotation of the interface K by the slit width ds in the circumferential direction is based on the rotation angle detector 1.
The operation command section 7 outputs based on the detection result of the step 09. Further, the threshold value for binarization is determined in consideration of the light intensity of the light source 201D and noise due to stray light or the like such that only the pixel data of the image of the peeled portion in the minute area P is larger than the threshold value. To set.

The reflected light R12 is reflected on the container A and the resin material layer C1.
, C2 are relatively strong as compared to the normal case, so that the binarization threshold can be a fixed value determined empirically, but is calculated from the light intensity distribution of the light intensity meter 3D. A generally known method such as determination may be applied.
The method using the fixed value determined empirically as described above is more preferable because good accuracy can be obtained easily.

In the inspection of the peeling defect, the binarized image display unit 404, as shown in FIG.
The binarized images of the reflected light receiving portions R11 'to R13' and the reflected light non-receiving portion N are reflected by the reflected light receiving portions R11 'to R13'.
Are displayed in white, and the reflected light non-light receiving portion N is displayed in black.
Here, the reflected light receiving portion R11 'is a light receiving portion for the reflected light R11, the reflected light receiving portion R12' is a light receiving portion for the reflected light R12, and the reflected light receiving portion R13 'is a light receiving portion for the reflected light R13. . As shown in FIG. 16, the reflected light receiving portions R11 'to R1
3 'appears on the screen 404a of the binarized image display unit 404 in a state where there is an interval in one direction corresponding to the circumferential direction of the interface K. Further, in the hollow fiber modules M having the same shape, each light receiving portion R1 is provided between the hollow fiber modules M.
The positions of 1 ', R12', and R13 are respectively displayed on the screen 404a.
Appears in almost the same position above. In particular, since the inside diameter of the container A is formed in a substantially perfect circle, the light receiving portion R12 'appears at a predetermined position with high accuracy.

As described above, FIG. 16 shows the case where the microscopic region P shown in FIG.
The binarized image displayed on the binarized image display unit 404 is shown. Therefore, in FIG. 16, the reflected light receiving portion R12 'is composed of regions a', b ', and c' corresponding to the divided regions a, b, and c of the minute region P. Further, the reflected light receiving portion R13 'is provided with divided areas d, e,
It is composed of areas d ', e', f 'corresponding to f.

In the discriminating device 5, the calculation judging section 502 calculates the area of the portion displayed in white on the binarized image display section 4.

The calculation judging section 502 is a binarized image signal larger than the threshold value of the binarizing section 403 in a region m of a predetermined width including the position where the light receiving section R12 'appears on the screen 404a. The area formed by the pixel corresponding to the white level signal, that is, the area of the light receiving section R12 'is calculated. The area of the light receiving portion R12 'calculated here is a peeling area proportional to the peeling area in the minute area P with the slit width ds at the interface K, which is irradiated with the slit light emitted from the slit light emitting section 202D. Is the exponent value of. In the binarized image shown in FIG. 16, the area of the light receiving portion R12 'is proportional to the sum of the areas of the regions a', b ', and c'.

The area of the light receiving portion R13 'is
Is correctly finished in a perfect circular cylindrical shape, it is proportional to the area of the normal bonding portion in the minute area P (the sum of the areas of the divided areas d, e, and f in FIG. 15). However, usually, the yarn bundle B is deformed, and the light receiving portion R13 'of the reflected light R13 transmitted through the interface K has many positional shifts, so that the light receiving portion R13' is merely qualitative information indicating that it is normal.

The calculation judging section 502 sequentially sums up the areas of the light receiving sections R12 'of the respective minute areas P. When the hollow fiber module M is rotated by 360 degrees by the drive motor 102, the calculation determining unit 502 calculates the integrated value corresponding to the area of the light receiving unit R12 'over the entire circumference of the interface K, that is, the peeling area in the entire area of the interface K, This integrated value, that is, the area calculation value, is compared with a peel defect inspection area reference value for each product number previously input by a data input device such as a keyboard not described from the storage unit 501, and based on the magnitude, the hollow fiber is determined. The quality of the module M is determined. That is, when the area calculation value ≧ the area reference value, the hollow fiber module M is determined to be defective with respect to the peeling defect inspection, and when the area calculation value <the area reference value, the hollow fiber module M is determined to be good. The area reference value can be determined empirically.

The display unit 6 displays the comparison result output by the calculation judging unit 502 for each hollow fiber module, that is, the result of judging pass / fail of the hollow fiber module regarding the peeling defect.

The operation command section 7 operates the hollow fiber module transfer mechanism 8 shown in FIG. 8 as described in the above description of the inspection for impervious defects and the like. The operation command unit 7 issues a command to the drive control unit 108 to drive the drive motor 10.
2 is operated to rotate the hollow fiber module M continuously or intermittently by 360 degrees so that the interface K moves in the circumferential direction by the slit width ds. The operation command unit 7 is provided with the hollow fiber module M provided to the storage unit 501.
The vertical movement mechanism (not shown) for the support device 1 is controlled based on the information on the part number so that the highest position of the interface K of the hollow fiber module of the part number coincides with the detection position M0.

With the inspection apparatus constructed as described above, the total area of the peeled portion formed between the container A of the hollow fiber module and the resin agent layer C1 or C2 is determined. In some cases, the hollow fiber module can be rejected as a defective product.

In the above embodiment, the light guided from the light source 201D by the optical fiber is transmitted to the slit light irradiating section 2D.
02D is condensed in a slit shape, and this slit light is irradiated to the uppermost end position M0 of the hollow fiber module M. The slit light irradiation unit 202D uses the light source 201D as a laser light source and uses the laser light as a cylindrical lens. And the like, and the top end position M0 of the hollow fiber module M
May be applied.

In the above embodiment, the second photometric device 3
Although a two-dimensional CCD camera is used as D, other inspection devices according to the invention may employ a line sensor such as a one-dimensional CCD camera as the second photometric device 3D. In this case, the second photometric device 3D is provided at a position where only the reflected light reflected at the detection position M0, that is, the reflected light R12 in FIG. 14, is captured. On the other hand, the binarization processing unit 403 binarizes the brightness of each pixel captured by the second photometric device 3D by the scanning with a predetermined threshold value, and
Counts the number of pixels exceeding the threshold. The total number of pixels counted by the operation determining unit 502 is a peeling index that is proportional to the area of the peeled portion in the minute region P. In this case, the threshold value for binarization considers the scanning time of the line sensor in addition to the noise due to the light intensity of the light source 201D and stray light in the case of the previous embodiment. As described above, even when the line sensor is used as the second photometric device 3D, the configuration and function of the other parts than the above-described parts are as follows.
Except for the binarized image display unit 404, the configuration is the same as that of the previous embodiment. Therefore, description of other parts in this inspection apparatus is omitted here.

In the inspection for the peeling defect, a single light receiving element can be used as the second photometric device 3D. When a single light receiving element is used as the second photometric device 3D as described above, the light receiving element is disposed near the hollow fiber module M so that the light receiving element does not capture the reflected lights R11 and R13 shown in FIG. A window 9 (see FIG. 14) as shown by a dashed line for shielding the reflected lights R11 and R13 is provided.
In this case, as shown in FIG. 17, the second photometric device 3D '
The output signal of the single light receiving element is sent to the abnormality detection signal forming unit 4 'without being subjected to the binarization processing. The abnormality detection signal forming section 4 'subtracts a noise component set in advance in consideration of stray light and slight reflected light generated at the normal bonding portion from the output signal level of the second photometric device 3D'. Then, a peeling index that changes in proportion to the peeling area in the minute region P is obtained, and a peeling signal that is a signal of the peeling index, that is, an abnormality detection signal is output.

As described above, even when a single light receiving element is used as the second photometric device, the second photometric device 3 shown in FIG.
1 except that the second photometric device 3D ′ and the abnormality detection signal forming unit 4 ′ are used in place of D and the abnormality detection signal forming unit 4 and that the window 9 is newly provided. It is almost the same as the inspection apparatus of the example. Therefore,
Here, a detailed description of the operation of the inspection apparatus is omitted.

Next, in the above-described embodiment, the operation determining section 502
Whether the hollow fiber module is a non-defective product or a defective product is determined only by determining whether or not the integrated value corresponding to the area of the peeled portion at the interface K exceeds the threshold value read out by the storage unit 501. And However, in the inspection apparatus of another hollow fiber module, in addition to the peeling signal forming unit 4 in FIG. 1, the divided areas a, b, c, d, e, and
A second peeling signal forming unit for forming a peeling signal indicating the presence or absence of peeling for each f is provided. In the discriminating device 5, each of the divided areas a, b, c or a, b, c of a predetermined number of continuous minute areas P is provided.
When c, d, e, and f all detect a peeling signal, the hollow fiber module can be regarded as a defective product even if the integrated value does not exceed the threshold value. .

In this manner, since the separation of the divided areas a, b, and c extends over a predetermined number of continuous small areas P, the aging of the resin agent layers C1, C2 causes the divided areas a, b, and c to change. The hollow fiber module M in which there is a risk that the peeling of the fiber is diffused can be determined as a defective product. In the case where the second peeling signal forming section as described above is provided, any one of the minute areas P is divided into the divided areas a, b, c, d, and
In the case where the peeling signal is detected in all of e and f, the hollow fiber module may be determined to be defective regardless of other determination conditions. This means that even if one micro area P, the divided areas a, b, c,
This is because if peeling occurs in all of d, e, and f, blood leakage or the like may occur therefrom.

[0137] What kind of peeling state is detected and whether or not the hollow fiber module M is determined to be defective is determined based on the characteristics of the hollow fiber module and the results of the process analysis.

V. <Inspection of Bubble Defect> The foam defect is a defect in which a minute concave portion such as a bubble mark is formed on the end surfaces of the resin material layers C1 and C2 cut along the end surface of the container A. The inspection apparatus for the hollow fiber module shown in FIG. 1 inspects this foaming defect by the following configuration and operation.

The structure of the supporting device 1 for inspecting a foaming defect is the same as that of the inspection for the peeling defect shown in FIG. Therefore, only the reference numerals in FIG. 18 are used here, and the detailed description of the support device 1 is omitted.

In the inspection for the foaming defect, the fourth light irradiation device 2E in FIG. 1 is used as the light irradiation device. FIG.
As shown in FIG. 8, the fourth light irradiation device 2E includes a fourth light source 201E and a slit light irradiation unit 202E.

The fourth light source 201E is formed of a halogen lamp or the like, and performs on / off of light emission and adjustment of light intensity based on a command given from the operation command unit 7. The fourth light source 201E adjusts the light intensity so that the luminance measured by a third photometry device 3E described later becomes a light intensity that is saturated in any of the pixels. The adjustment of the light intensity will be described later.

The slit light irradiating section 202E is connected to the light source 201E by a plurality of optical fibers 203E, for example.
After arranging these optical fibers 203E in a horizontal line, the light emitted from the terminals of these optical fibers 203E is transmitted to a cylindrical lens (not shown).
The light is condensed in a slit shape, and this slit light is emitted. As shown in FIG. 19, the slit-shaped light is generated when the end surface C11 of the resin agent layer C1 of the hollow fiber module M is an ideally formed surface that is perpendicular to the axis C0 over the entire end surface. Is incident on the diameter of the end face C11 at an incident angle α1, and irradiates a slit-shaped irradiation area H having a width h including the diameter of the end face C11. In the following, the irradiation area H of the slit light applied to the end face C11 when the end face is the ideal forming face will be described as the irradiation area H for an arbitrary end face C11.

When the end face C11 is the ideal end face, the slit light L emitted from the slit light emitting portion 202E
As shown by a solid line in FIG. 4, the light is reflected at the reflection angle β21 in the slit-shaped irradiated area H and becomes regular reflection light R21.

If there is a concave portion DP due to a bubble mark or the like on the end face C11, the regular reflection light R22 of the slit light L is generated as shown in FIG.
As shown by an imaginary line, the light is reflected at the concave portion DP at a reflection angle β22 different from the reflection angle β21. It is preferable that the reflected light does not receive irregularly reflected light but receives specularly reflected light.

Also, when the end face C11 is cut obliquely with respect to the ideal end face or is an end face having distortion, the regular reflection light R21 'of the slit light L has a reflection angle different from the reflection angle β21. It reflects at the angle β21 ′. But,
In this case, the inclination of the end surface C11 in the irradiation area H is smaller than that in the case of the concave portion DP, and therefore, the reflection angle β21 ′ of the regular reflection light R21 ′ is larger than the reflection angle β22 of the regular reflection light R22.
Is closer to the reflection angle β21 at the ideal end face.

The width h of the slit-shaped irradiated area H
The setting method of will be described later.

In the inspection for the foaming defect, a third photometer 3E is used as a photometer. The third photometric device 3E is mainly composed of a known two-dimensional television camera. The third photometric device 3E captures an image of the end face C11 two-dimensionally. In addition, the third photometric device 3E
An analog image signal representing the luminance of light of each pixel of the image in the window W corresponding to the diameter of the end face C11 as indicated by a virtual line in FIG. 19 is output.

The width w of the window W is set to, for example, five times or more the pixel width σ. In the case where the two-dimensional television camera is a model using a CCD element, the pixel is a dimension determined on condition of each CCD element of the camera, a focal length of a lens, a distance between the television camera and the end face C11, and the like. Resolution and the end face C11 on the image
Is the minimum structural unit. That is, when the photometric device 3 is a two-dimensional CCD camera, a large number of two-dimensional CCDs are arranged.
Since the end face image is formed on the element and an image is formed from the output of each CCD element, each pixel on the image and the CCD
The minimum units of the elements are in a one-to-one correspondence. The width σ of the pixel is preferably １ ／ or less, more preferably 以下 or less, of the width d of the smallest recess DP among the recesses DP to be detected. If the pixel width σ is increased so that no pixel is completely included in the image of the concave portion DP, the detection accuracy decreases. In FIG.
1 / the width d of the smallest concave portion DP where the width σ of the pixel is to be detected.
It is set to 4. In FIG. 22, the pixels are square, but may be rectangular.

The width w of the window W has a great influence on the processing speed of the inspection apparatus. In order to improve the processing speed, it is preferable to increase the width w of the window W. However, as described later, if the width w of the window W is too large, there is a problem that a part of the window W may not be uniformly irradiated when the end face C11 has a large inclination or distortion. Therefore, window W
Is preferably equal to or less than 幅 of a width h of a region to be irradiated H described later.

The width of the window W may be changed according to the type of the hollow fiber module M.

The width h of the slit-shaped irradiated area H is such that when the reflected light received by the third photometric device 3E is the regular reflected light R21 or R21 'shown in FIG. The width is set such that it can be positioned in the light receiving range of the reflected light R21 or R21 '. In this embodiment, as shown in FIG. 22, the width (2d + 2σ) obtained by supplementing the width σ of the pixel on both sides of the width (2d) that is twice the width d of the minimum recess DP is the width h of the irradiation region H. And As described above, since the width d of the minimum concave portion DP is four times the pixel width σ, the width h of the irradiated region H is set to 10σ, that is, 10 times the pixel width σ. Will be.

The width h of the irradiation area H is preferably
It is preferable to set the width to be 10 times the pixel width σ (in the case of FIG. 22) or more. According to the experiment, when the width h of the irradiation region H is smaller than 10 times the width σ of the pixel,
When the end face C11 is obliquely cut or the end face C11 is distorted, even if such a distortion or the like is within an allowable range, the specularly reflected light R21 or R21 'from a part of the window W.
May shift. In such a case, although there is no concave portion DP to be detected, the luminance of the pixel in the portion where the regular reflection light R1 or R21 'is not irradiated due to the deviation is set to the same level as that of the portion corresponding to the concave portion DP. A down problem occurs.

As described above, it is preferable that the width h of the irradiation area H be at least twice the width w of the window W. Accordingly, it is possible to prevent a pixel portion that is not irradiated with the regular reflection light R21 or R21 'from being generated in the window W due to an inclination of the end face C11 of the hollow fiber module M or an error in mounting the hollow fiber module M to the support device 1.

When the slit-shaped irradiation area H is set as described above, if the end face C11 is an ideal end face,
As shown in FIG. 21A, the window W is set at the center of the light receiving region H1 of the regular reflection light R21 reflected by the slit-shaped irradiation region H. That is, the entire area of the window W is illuminated by the regular reflection light R21.

When the end face C11 is cut with an inclination, if the inclination is within an allowable range, the window W11 is cut.
As shown in FIG. 21 (b), the regular reflection light R21 'reflected at the slit-shaped irradiation region H is set at a position H1' which is deviated in the width direction of the light receiving region H1 or as shown in FIG. 21 (c). The position is set as shown in FIG. However, also in this case, since the inclination is within the allowable range, the entire area of the window W is illuminated by the regular reflection light R21 '.

Further, when the end face C11 is distorted,
Specularly reflected light R21 'reflected by the slit-shaped irradiated area H
Has a curved shape as shown in FIG. 21D, for example. However, if the distortion is within an allowable range, the window W is set within the light receiving region H1 ', and the entire window W is illuminated by the regular reflection light R21'.

When there is a concave portion DP due to a bubble mark or the like on the end face C11, the reflected light R2 reflected at the concave portion DP is formed.
Are received at positions off the window W. Therefore, when the concave portion DP exists at a position corresponding to the window W, the luminance at that position decreases.

The third photometric device 3E measures the luminance level of all the pixels in the window W, and supplies the luminance level to the abnormality detection signal forming device 4 as a fifth analog image signal.

The operation of the A / D converter 401 and the image storage section 402 of the abnormality detecting signal forming device 4 when the fifth analog image signal is received is based on the above-described impermeability defect, thread disorder defect, and thread defect. This is the same as the case of the inspection for the peeling defect.

In the case of inspection for a foam defect, that is, when the abnormality detection signal forming unit 4 inputs the fourth analog image signal, the binarization processing unit 403 receives the image storage The digital image signal in the area of the window W shown in FIG. Then, this digital image signal is binarized by a predetermined threshold value S1P to obtain a fifth binary image signal.
Form a digitized image signal. That is, the binarization processing unit 403 replaces an image signal whose level is higher than the threshold S1P with a white level, and replaces an image signal whose level is lower than the threshold S1P with a black level.

In the conventional method of detecting a concave portion using light, the intensity of the light irradiated by the light irradiating device and the luminance, which is the amount of reflected light, are measured by the third light measuring device 3E. Is set so as not to be partially saturated at the photometric luminance level by the above. As described above, when the slit light having the light intensity according to the conventional method is applied, the brightness level in the window W is increased and decreased due to the unevenness due to the end portions of the many hollow fibers exposed on the end face C11 in FIG. As shown in FIG. Each time the hollow fiber module M to be inspected is replaced, the end face C
The eleven minute irregularities change, and the luminance level moves up and down accordingly. For this reason, the luminance level difference between the concave portion DP and the normal portion is small, and when the image signal is binarized by the binarization processing section 403, a portion other than the concave portion DP to be detected is replaced with a black level. Detection error is likely to occur.

In order to prevent the above detection error,
In the present invention, the light intensity of the slit light irradiated by the fourth light irradiation device 2E is set to be higher than the light intensity at the time when the luminance of each pixel starts to saturate in some pixels. FIG. 23 (c) shows the luminance level at the position corresponding to FIG. 23 (a) when the light intensity of the slit light is set to be about twice that at the time when a part of the end face C11 starts to be saturated at first. ing. As can be seen from FIG. 23 (c), since the light diffusing property of the end of the yarn bundle B exposed on the end face C11 is strong, the luminance of the end face C11 becomes higher as the light intensity of the slit light is increased. First, the yarn bundle B starts to be saturated from a certain central portion.

FIG. 23 (d) shows the luminance level at the position corresponding to FIG. 23 (a) when the light intensity of the slit light is set to be four times that at the time when the part of the end face C11 first begins to saturate. Is shown. In FIG. 23D, all the luminance levels other than the position corresponding to the concave portion DP have reached the saturation value. The number of times that the light intensity of the slit light is initially increased when a part of the end face C11 begins to saturate depends on the molding accuracy of the end face C11 and the end face C
If it is set to a value exceeding 4 times, the luminance level of the portion corresponding to the concave portion DP to be detected approaches the saturation value or reaches the saturation value, though it depends on the positional relationship between 11 and the photometric device 3. When a CCD camera is used, an abnormal phenomenon peculiar to the CCD camera, such as a smear phenomenon or a blooming phenomenon, may strongly occur. Therefore, according to the present invention, the light intensity of the slit light is first reduced by a part of the end face C11. It is specified within four times of the time when saturation starts.

On the other hand, it is effective to increase the light intensity as much as possible for the following reason described with reference to FIG.

FIG. 24 shows a hollow fiber module M having a concave portion DP having a diameter of 0.8 mm to be detected within a substantially ideal end face C11 and a hollow fiber module M having no concave portion DP serving as a defect but having a large surface distortion. Is for explaining the change of the binarized image due to the lamp voltage and the threshold value S1P.

A characteristic curve K1 shown in FIG.
The change of the limit threshold value S1P with respect to the lamp voltage for the binarization processing unit 403 to detect only the m concave portion DP alone is shown. That is, if the end surface C11 is an ideal surface, an arbitrary threshold value S1P above the characteristic curve K1 can be adopted. On the other hand, if a threshold value lower than the characteristic curve K1 is adopted, a low-luminance portion other than the concave portion DP is erroneously detected. However, if the end face C11 is an end face with strong distortion, the threshold value S1P
As the height is increased, a distorted portion other than the concave portion DP to be detected is detected, and a detection error starts to occur.

The characteristic curve K2 in FIG. 24 shows the change of the threshold value S1P at which the detection error starts to occur with respect to the lamp voltage, particularly for the hollow fiber module having a strong distortion. The threshold value S1P lower than the characteristic curve K2 is used. This indicates that the occurrence of the detection error can be avoided. That is, in order to detect the above-mentioned diameter 0.8 mm and the corresponding concave portion DP without error, it is preferable to use the region (hatched portion in FIG. 24) E sandwiched between the characteristic curves K1 and K2. Then, as is clear from FIG. 24, in the region E, as the light intensity increases, the width of the usable threshold value S1P increases. Therefore, in order to widen the selection range of the threshold value S1P, the light intensity of the slit
It can be said that it is preferable to set the light intensity to a light intensity within four times that at the time when a part of 11 starts to be saturated and to be large. As the light intensity increases, the image diameter of the binarized concave portion DP decreases, so it is not preferable to set the highest light intensity,
Increasing the light intensity makes it easier to detect the concave portion DP as described above, and improves the detection accuracy.

The method for determining the binarization threshold S1P is shown in FIG.
In addition to the method of inputting and reading out a constant value (may be changed depending on the type and lamp voltage) based on the above data, the standard value or range of the luminance distribution in the window imaged by the television camera is calculated. Alternatively, generally known methods such as determination may be applied.

In the inspection of the foaming defect, the binarized image display unit 404 displays only the image obtained by the image signal replaced with the white level and the black level by the binarization processing unit 403, that is, only the concave portion DP in black. Wind W indicated by
Display the image in.

In the discriminating device 5, the calculation judging section 502 calculates the area of the portion displayed in white on the binarized image display section 4 as follows.

The operation judging section 502 includes the binarizing section 4
03, the pixel replaced with the black level is detected, and based on the number of pixels constituting the concave portion DP here, the positional relationship, and the like,
The feature amounts such as the size and area of the concave portion DP formed by these pixels are calculated.

The storage unit 501 receives a command from the operation determination unit 502 and selects a reference value relating to the feature value for each product number of the hollow fiber module M from a reference value relating to the feature quantity by a data input device such as a keyboard. The reference value for foam defect inspection of the input product number is read out,
Give to 02. The reference value for foam defect inspection can be determined empirically.

The calculation judging section 502 also compares the characteristic quantities such as the size and area of the concave portion DP with the reference values for foam defect inspection output from the storage section 501 to judge the quality of the hollow fiber module M. A separating function is provided for separating the hollow fiber module M detached from the support device 1 and defective products based on the determination result. That is, when the comparison result satisfies the reference value> calculated feature value, it is determined to be non-defective, and when the reference value ≦ calculated feature value, the product is determined to be defective. Give 7

The display unit 6 displays the comparison result output by the calculation judging unit 502 for each hollow fiber module, that is, the result of judging the quality of the hollow fiber module with respect to the foaming defect.

The operation command unit 7 operates the hollow fiber module transfer mechanism 8 shown in FIG. 8 as described in the above description of the inspection for impervious defects and the like. In the inspection of the foaming defect, the operation instructing unit 7 determines whether the calculation
In response to the timing command from the second, a command signal is given to the drive control unit 108. As described above, upon receiving this command signal, the drive control unit 108 drives the drive motor 102 to rotate at a constant speed for a predetermined time or a predetermined angle about the axis C0, and then stops. Each time the rotation angle of the hollow fiber module M detected by the detector 109 reaches 180 degrees or more, the hollow fiber module transfer mechanism 8 described above is operated, and the hollow fiber module M mounted on the support device 1 is operated. Is removed, and a new hollow fiber module M is mounted on the support device 1.

The operation command section 7 determines the axis C of the hollow fiber module M of the product number based on the information on the product number of the hollow fiber module M given to the reference value reading section 502.
0 is coincident with the position corresponding to the window W,
A vertical movement mechanism (not shown) for the support device 1 is controlled.

The inspection apparatus configured as described above detects a concave portion DP due to a bubble mark or the like generated on the end surface of the resin agent layer C1 or C2 of the hollow fiber module M, and detects the hollow fiber module having the concave portion DP. It can be eliminated as defective.

With the above-described apparatus, the conditions A to C shown in Table 1 were used.
Inspection was performed to detect the concave portion DP in the above, and the following detection results were obtained.

[0179]

[Table 1]

Condition A is the width h of the irradiated area H / the width σ of the pixel.
Is set to “7”. Under this condition A, a concave portion DP having a diameter of 1 mm or less could be detected in the hollow fiber module M having a relatively good finish of the end face C11. However, when the end face C11 is cut obliquely or the end face C11 is distorted. In this case, when the degree became large, a detection error occurred because a part of the window W was not irradiated with the regular reflection light. However, this detection error is due to a problem in setting the width h of the irradiation area H and the width w of the window W.

In order to suppress such a detection error, it is conceivable to increase the width h of the irradiation area H and the width w of the window W while maintaining the relationship. However, in increasing the width h of the irradiation area H, it is necessary to note that the contrast of the image of the concave portion DP which is a defect image is likely to be deteriorated due to the strong light diffusion effect of the end face C11. On the other hand, as described above, the width w of the window W is set to be 1 / the width h of the irradiation region H in order to suppress the occurrence of the inspection error.
On the other hand, it is set to be as large as possible in consideration of the processing speed for inspecting the entire end face C11.

In the condition B, the width h of the irradiated area H / the width σ of the pixel is set to 50 in consideration of the above. In the experimental device, under this condition B, the concave portion DP having a diameter of 1 mm or less, which should be detected without any problem, was detected without any problem with respect to the inclined cut of the end face C11 or the displacement of the mounting position of the hollow fiber module M with respect to the support device 1. , A detection error occurred in some workpieces with strong surface distortion.

In condition C, the width h of the irradiation region H is increased and the width w of the window W is reduced as compared with the condition B. As a result, even in the hollow fiber module M having strong surface distortion in which a detection error has occurred in the condition B, the concave portion can be surely formed.
DP could be detected.

In the inspection of the foaming defect, in the above case, the third photometric device E3 drives the drive motor 102 to rotate the hollow fiber module M by a predetermined amount, whereby the brightness level of the entire end face C11 is obtained. The brightness level of the entire end face C11 may be measured by moving the hollow fiber module M upward or downward by the width h of the window W by a vertical movement mechanism (not shown). .

In the above example, the third photometric device 3E
The method of increasing the intensity of slit light is exemplified as a means for improving the detection accuracy such as increasing the amount of light measurement for each pixel, but the aperture of the camera lens attached to the third photometric device 3E is opened. Or you may replace it with a bright lens. Further, a method using amplification and a saturation characteristic function by a signal amplifier included in the third photometry device 3E may be used. However, a method of increasing the intensity of the slit light is more preferable because it is easy to use.

In the above example, the third photometric device 3E
Although a two-dimensional television camera that can obtain information on the luminance level of the pixels in the window W at one time is used, in other inspection devices, CCD elements are arranged one-dimensionally as a third photometric device 3E. Alternatively, a line sensor such as a one-dimensional CCD camera may be employed.

As the fourth light irradiation device 2E, a light irradiation device configured to convert laser light into slit light using an optical element may be used. However, even if the light intensity decreases, the light intensity can be periodically managed and adjusted by the lamp voltage. Therefore, it is more preferable to use the light irradiation device of the type used in the embodiment.

In the inspection apparatus shown in FIG. 1, only one end of the hollow fiber module M is inspected. However, the inspection apparatus for the hollow fiber module according to the invention is:
Needless to say, a device capable of inspecting both ends simultaneously is included.

In this embodiment, an example is shown in which a hollow fiber module is used for a dialyzer. However, this is merely an example. The present invention can be suitably applied to a foaming test of a water purifier module or the like.

[0190]

According to the hollow fiber module inspection apparatus of the first aspect, the color change of the end face of the hollow fiber module which has become whitish as compared with the normal one can be surely performed by the method of binarizing the image signal of this end face. By being able to capture, impervious defects can be easily and reliably inspected.

According to the hollow fiber module inspection apparatus of the second aspect, the end face of the hollow fiber module is illuminated by the first and second light emitting portions to flatten the light quantity distribution on the end face of the hollow fiber module. Thus, even if the image signal of the end face is binarized without correcting the image signal, the white microbubbles generated in the non-threaded portion can be reliably detected, and the automatic detection of the non-threaded defect in the hollow fiber module can be performed. This has the effect of enabling inspection. Also,
The first light emitting portion can make the lighting effect of the end face of the hollow fiber module uniform, and the combination of the first and second light emitting portions makes the lighting effect of a part slightly dived from the end face uniform. Therefore, it is not necessary to correct the illumination effect in local spot detection (shade detection in black and white, and color spot detection in color), and a broken thread defect can be inspected with a relatively simple device.

According to the inspection apparatus for the hollow fiber module according to the third aspect, by separating the defective hollow fiber module, it is possible to eliminate the irrationality of processing such a defective hollow fiber module in a later step.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of an inspection apparatus for a hollow fiber module according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating first and second light irradiation devices.

FIG. 3 is a diagram illustrating an illumination state by a first ring-shaped emission unit.

FIG. 4 is a diagram illustrating a light amount distribution during illumination by a first ring-shaped emission unit.

FIG. 5 is a view showing a modification of the first ring-shaped emission section.

FIG. 6 is a diagram illustrating a light amount detection unit on an end face of the hollow fiber module.

FIG. 7 is a diagram showing a light amount distribution when there is an impervious defect.

FIG. 8 is a schematic plan view showing a transfer mechanism of the hollow fiber module.

FIG. 9 is a view for explaining a ring-shaped emission section.

FIG. 10 is an explanatory cross-sectional view illustrating an irradiation state by a ring-shaped emission unit.

FIG. 11 is a diagram illustrating a light amount distribution in the inspection of a yarn disorder defect.

FIG. 12 is a diagram illustrating a light amount distribution in an inspection for a non-threading defect.

FIG. 13 is a detailed view of a portion used for inspecting a peeling defect in the inspection device of FIG. 1;

FIG. 14 is an explanatory diagram showing a reflection state at a peeling portion.

FIG. 15 is a diagram illustrating a minute region irradiated with slit light.

FIG. 16 is a diagram showing a binarized image of reflected light.

FIG. 17 is a diagram showing a configuration of a main part of another hollow fiber module inspection device for inspecting a peeling defect.

FIG. 18 is a detailed view of a portion used for inspection of a foaming defect in the inspection device of FIG. 1;

FIG. 19 is a diagram showing an irradiation state of slit light on an end face.

FIG. 20 is a diagram illustrating a reflection state at an end face.

FIG. 21 is a diagram illustrating a relationship between an irradiated portion and a window.

FIG. 22 is a diagram showing a relationship between a concave portion and a pixel.

FIG. 23 is a diagram showing a measured luminance level corresponding to light intensity.

FIG. 24 is a graph showing a relationship between light intensity and a threshold value for binarization.

FIG. 25 is a sectional view of a hollow fiber module.

FIG. 26 is a diagram illustrating a use state of a hollow fiber module.

FIG. 27 is a diagram illustrating a known illumination method.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Support device 2A 1st light irradiation device 2B 2nd light irradiation device 202A 1st ring-shaped light-emitting part 202B 2nd ring-shaped light-emitting part 3A 1st photometric device 403 Binarization processing part 502 Calculation determination part

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-6-273126 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01N 21/84-21/958 B01D 61 / 00-71/82

Claims (3)

(57) [Claims] 1. An apparatus for inspecting a hollow fiber module having a resin material layer in which ends of a plurality of hollow fibers are solidified with a resin material, comprising: a support means for supporting the hollow fiber module; and an end face of the hollow fiber module. A ring-shaped light emitting portion that is disposed substantially coaxially with the module at a position separated by a predetermined distance from the outer diameter of the module and has an outer diameter equal to or greater than the outer diameter of the module; At a position where the regular reflection light of the light irradiated from the end face of the hollow fiber module is not incident, and a photometric means for measuring the amount of light for each subdivided region of the end face of the hollow fiber module; and A threshold value is calculated by adding an arbitrary offset amount to the light measuring amount of one selected point or the average light measuring amount of two or more selected points, and the light measuring amount of each area measured by the light measuring unit is calculated. 2 in the threshold value
A binarization processing unit that outputs a binarized binarized signal; and a determination unit that determines whether or not the hollow fiber module is defective based on a processing result of the binarization processing unit. An inspection apparatus for a hollow fiber module. 2. An inspection apparatus for a hollow fiber module having a resin material layer in which ends of a plurality of hollow fibers are solidified with a resin material, wherein a supporting means for supporting the hollow fiber module, and an end face of the hollow fiber module A ring-shaped first light-emitting portion, which is disposed substantially coaxially with the module at a predetermined distance from and has an outer diameter equal to or greater than the outer diameter of the module, and the hollow at a position corresponding to the filling resin agent A ring-shaped second light emitting portion surrounding the resin material layer to irradiate the yarn module with light from the side; and an end face of the hollow fiber module and irradiated from the first light emitting portion. A photometer for measuring the amount of light in each of the subdivided areas of the end face of the hollow fiber module at a position where the regular reflection light of the light at the end face does not enter; and any one selected point on the end face of the hollow fiber module Thereby calculating the threshold value by adding an arbitrary offset amount to an average photometric quantity of photometric or 2 or more selected points, a metering of each region in which the photometric means is determined by the threshold value 2
A binarization processing unit that outputs a binarized binarized signal; and a determination unit that determines whether the hollow fiber module is defective based on a processing result of the binarization processing unit. In addition, the first light emitting unit and the second light emitting unit are arranged so that the end portion of the hollow fiber is not clogged with the resin agent and the distribution of the measured light amount is obtained by the imaging camera. Irradiates light that becomes flat in at least the hollow fiber region on the end face. 3. The hollow fiber module inspection apparatus according to claim 1, further comprising a classification unit that classifies the defective hollow fiber module based on a result of the determination by the determination unit.