Method for quality control of a fibrous tape during unwinding

Abstract

The invention relates to a method for monitoring a fiber strip (500) during continuous movement. The monitoring method comprises the steps of: periodically acquiring an image of at least a portion of the fiber strip (500); for each acquired image, analyzing the acquired image to: select several areas of interest within the acquired image, each area of ​​interest including the fiber strip (500); for each area of ​​interest, extracting the signal contained within the area of ​​interest, said signal being expressed in pixel intensity; and processing said signal to: determine the width of the fiber strip (500); determine, within the determined width of the fiber strip (500), the presence of holes and, if applicable, their length; and determine, within the determined width of the fiber strip (500), an average light intensity and a standard deviation of the light intensity. Figure for the abstract: Fig. 1

Description

Title of the invention: Method for controlling the quality of a fibrous strip during operation Technical field of the invention

[0001] The present invention falls within the field of quality control of a fibrous strip during unwinding.

[0002] A particular field of application of the invention is the real-time control of the quality of a fibrous web during its running in an automated resin impregnation machine to monitor potential defects in the fibrous web during a process of impregnating said fibrous web.

[0003] Another particular field of application of the invention is the real-time control of the quality of a fibrous strip being run through an automated deposition machine to monitor potential defects in the fibrous strip during a deposition process with a view to manufacturing a composite part. Prior art

[0004] The production of parts made of composite material, whether organic (for example based on carbon fibers reinforced by an epoxy resin) or ceramic (for example based on alumina fibers reinforced by an alumina resin) may require the use of pre-impregnated fibrous semi-finished products, whether in the form of pre-impregnated fabrics (commonly called "prepreg") or in the form of pre-impregnated wicks (commonly called "towpreg").

[0005] In a conventional resin impregnation process, the fibrous semi-finished product, generally in the form of a continuous dry fibrous strip wound and packaged in a reel, is unwound, impregnated with resin, dried, then wound again and packaged in the form of a reel. The impregnation of the fibrous semi-finished product can lead to defects, for example heterogeneity in the resin content (excess or lack of resin), whether in the length or in the width of the fibrous semi-finished product. In the particular case of the fibrous semi-finished product in the form of rovings, two additional defects can appear: variability in the width of the fibrous semi-finished product and the creation of slits, a slit separating the fibrous semi-finished product into two parts over a portion of its length.

[0006] In the resin impregnation process, for example, heterogeneity in the resin content across the width of the pre-impregnated fibrous semi-finished product may lead to difficulties in drying said pre-impregnated fibrous semi-finished product. The most impregnated areas will be less dry than the others, which may lead to fouling of the resin impregnation machine used to carry out the process. resin impregnation, and cause manufacturing shutdowns.

[0007] After resin impregnation, the pre-impregnated fiber semi-finished product is, for example, used in an automated deposition process to form a composite material part. An automated deposition process may, for example, be an automated placement process (known by the acronym AFP for "Automated Fiber Placement" in English terminology). A deposition machine is used to carry out the automated deposition process.

[0008] During the process of depositing the pre-impregnated fibrous semi-finished product on a substrate, the heterogeneity in the resin content along the length of said pre-impregnated fibrous semi-finished product can lead on the one hand to a modification of its flexibility and therefore to a difficulty in conveying the pre-impregnated fibrous semi-finished product onto the substrate, and on the other hand to a disturbance as to its capacity to adhere to the substrate, thus being able to cause manufacturing stoppages.

[0009] Similarly, variations in the width of the pre-impregnated fibrous semi-finished product may prove detrimental to the final properties of the composite material part. For example, too small a width of the pre-impregnated fibrous semi-finished product may generate gaps in the pre-impregnated fibrous semi-finished product at certain locations on the substrate at the time of deposition. Too large a width of the pre-impregnated fibrous semi-finished product may generate overthicknesses due to overlaps of strips of pre-impregnated fibrous semi-finished product at certain locations on the substrate at the time of deposition, or lead to jamming of the deposition machine, resulting in a stoppage of production.

[0010] Currently, checks of the pre-impregnated fibrous semi-finished product, during the resin impregnation process, are carried out by laser distance sensors measuring the time of flight. However, such sensors only allow the measurement of a single type of defect, that of the width. In addition, such sensors only allow point measurements, that is to say at a single point of the resin impregnation machine. Checks of the heterogeneity of the pre-impregnated fibrous semi-finished product, during the resin impregnation process, also exist. Light is emitted by a monochromatic source on the pre-impregnated fibrous semi-finished product and the signal transmitted by it is analyzed to deduce the resin content. Here again, only one type of defect, that of heterogeneity, is analyzed and the measurements are only point measurements. Presentation of the invention

[0011] The present invention aims to remedy the drawbacks of the solutions proposed by the prior art, in particular those set out above.

[0012] To this end, the present invention proposes an automatable, reliable and reproducible method for controlling, in real time, a fibrous web during decoupling. threading, allowing effective detection and measurement of defects, in particular the aforementioned defects, namely heterogeneity of the impregnation in the width and length of the fibrous strip, the number of slits, their size and position, and a change in the width of the fibrous strip along its length.

[0013] According to the invention, the control method comprises the steps of: - acquisition, periodically, of an image of at least part of the fibrous strip, - for each acquired image, analysis of said acquired image for: • select several areas of interest in the acquired image, each area of ​​interest including the fibrous band, • for each area of ​​interest, extraction of the signal contained in the area of ​​interest, called the useful signal, said useful signal being expressed in pixel intensity, and processing of said useful signal for: • determine the width of the fibrous strip, • determine, in the width of the determined fibrous strip, the presence of holes, and where appropriate their length, • determine, within the width of the determined fibrous strip, an average light intensity and a standard deviation on the light intensity.

[0014] Preferably, the fibrous strip is a fibrous semi-finished product intended to be used for the manufacture of composite material parts. The fibrous semi-finished product may be in the form of either pre-impregnated fabrics or pre-impregnated rovings.

[0015] The control method thus advantageously makes it possible to obtain, solely from images, an indication of the quality (width, presence of holes, homogeneity of the impregnation) of the fibrous strip and the appearance of any defects.

[0016] Determining the presence of holes allows the determination of a slot-type defect in the fibrous web.

[0017] Determining the average light intensity of the fibrous strip allows the determination of a heterogeneity type defect of the fibrous strip, in the length of said fibrous strip.

[0018] Determining the standard deviation of the light intensity of the fibrous strip allows the determination of a heterogeneity type defect of the fibrous strip, in the width of said fibrous strip.

[0019] The control method according to the invention is advantageously carried out in real time, that is to say during the running of the fibrous strip.

[0020] The control method advantageously allows, in a single image, control in several places of the fibrous strip, which existing sensors do not offer.

[0021] The control method is iterative, that is to say that the steps described above are repeated, as the fibrous strip advances, so as to control the entire strip over its entire length.

[0022] The areas of interest selected in each acquired image are preferably the same. The areas of interest in the images are previously defined, for example before the start of the scrolling of the fiber strip.

[0023] The testing method according to the invention is advantageously a non-destructive and non-contact testing method, that is to say that no degradation is made to the surface of the fibrous strip for the purposes of testing.

[0024] In particular embodiments, the method may further comprise one or more of the following characteristics, taken in isolation or in all technically possible combinations.

[0025] In particular embodiments, when each acquired image is a color image, the method may comprise a step prior to the step of analyzing each acquired image, of transforming the acquired image to provide an image whose pixel intensity is defined in gray levels.

[0026] In particular embodiments, the width of the fibrous band is determined by a two-kernel classification algorithm. Such an algorithm advantageously makes it possible to separate the fibrous band from the background.

[0027] The invention also relates to a device for controlling the quality of a treatment comprising: - an image acquisition device, - a control unit comprising: • a module for processing acquired images, • a storage module on which code instructions are recorded for the execution, by the acquired image processing module, of a control method according to one of its previous implementation modes.

[0028] The invention also relates to a use of a control device meeting one or more of the characteristics described above and / or below in a resin impregnation machine.

[0029] The invention also relates to a use of a control device meeting one or more of the characteristics described above and / or below in an automated depositing machine. Brief description of the figures

[0030] The invention will be better understood on reading the following description, given by way of non-limiting example, and made with reference to the following figures:

[0031] [Fig.l] illustrates a schematic side view of a machine for impregnating a fibrous web with resin, the resin impregnation machine being equipped with a device for controlling the quality of the fibrous web;

[0032] [Fig.2] represents a schematic top view of the resin impregnation machine of [Fig.l],

[0033] [Fig.3] represents a photograph of an image taken by an image acquisition device of the fiber web quality control device,

[0034] [Fig.4] illustrates the selection of areas of interest in the photograph of [Fig.3] and the curves of the signals associated with these areas of interest,

[0035] [Fig.5] illustrates the determination of a width of the fiber band in a curve of a signal associated with an area of ​​interest,

[0036] [Fig.6] illustrates the determination of holes in the fiber band in a curve of a signal associated with an area of ​​interest.

[0037] In these figures, identical reference numerals from one figure to another designate identical or similar elements. Furthermore, for reasons of clarity, the drawings are not to scale, unless otherwise indicated. Description of the embodiments

[0038] A method for controlling a fibrous web 500 during continuous running according to the invention is now described in the non-limiting context of use with a resin impregnation machine, called machine 200. The machine 200 is configured to impregnate the fibrous web 500, initially dry, with a resin 230.

[0039] Figures 1 and 2 schematically illustrate an example of the machine 200. [Fig.l] represents a side view of the machine 200 and [Fig.2] represents a top view of the machine 200 of [Fig.l].

[0040] In a non-limiting and conventional manner, the machine 200 comprises in particular a succession of rollers 210 and a tank 220 for receiving a resin 230 in liquid form.

[0041] At the start of the line, the fibrous strip 500 is dry and packaged in the form of a reel 510, as shown on the right of [Fig.l].

[0042] The fibrous strip 500 is unwound and conveyed to the tank 220, via the rollers 210, passes through the resin bath 230 and emerges impregnated with resin.

[0043] The fibrous strip runs continuously on the machine 200, in a direction of advance represented by the arrow A in FIGS. 1 and 2.

[0044] Preferably, the pre-impregnated fibrous strip forms a fibrous semi-finished product intended to be used for the manufacture of composite material parts.

[0045] During passage through the rollers 210 and through the resin bath 230, quality defects in the fibrous web 500 may appear. These defects may subsequently cause defects in the composite material part which will be produced from this fibrous strip.

[0046] In order to monitor defects in the fiber web, a control of the quality of the fiber web, during running, is ensured by the control method according to the invention, in real time.

[0047] [Fig. 1] schematically illustrates an example of a device for controlling the quality of the fibrous web 500, called control device 100, capable and configured to implement the control method. The control device 100 is preferably adapted to be integrated on the machine 200.

[0048] The control device 100 advantageously comprises an image acquisition device 110 configured to acquire, in real time, an image of at least a portion of the fiber strip 500 being run on the machine 200. The image acquisition device 110 is preferably arranged above the machine 200, as illustrated in [Fig.l].

[0049] The image acquisition device 110 is preferably a camera.

[0050] The camera is for example a conventional camera on the market. It is not necessary for the purposes of the invention to use very high frequency industrial cameras or cameras in spectra other than the visible spectrum.

[0051] The camera 110 may comprise a lens.

[0052] The camera 110 may include a filter, for example in the visible range.

[0053] The control device 100 may comprise several image acquisition devices 110 for each acquiring a portion of the fiber web. The image acquisition devices 110 may have a field of vision which slightly overlaps.

[0054] In the non-limiting example of [Fig. 1], the control device comprises two image acquisition devices 110, each image acquisition device being configured to acquire an image of a portion of the fiber strip.

[0055] The control device 100 further comprises a control unit 150, as illustrated in [Fig.l]. The control unit is connected to the image acquisition device(s) 110. The control unit 150 comprises: - a module 120 for processing the acquired images, for example of the microprocessor type, to implement the control method according to the invention, - a storage module 130, such as a computer memory, for example a hard disk, on which code instructions are recorded for the execution of the control method according to the invention.

[0056] The control method, in association with the control device 100, makes it possible to follow the movement of the fibrous strip 500 in order to detect, in real time, the appearance and variation of defects in said fibrous strip, from an image acquired by the image acquisition device. The defects detected include variation in the width of the fibrous strip, the presence of slits, and where applicable their length, heterogeneity in the resin content in the width and length of the fibrous strip, when the latter is impregnated with resin.

[0057] The control method is now described.

[0058] The control method comprises a step of acquiring, periodically, of an image of the fibrous band.

[0059] This step consists of acquiring an image of a portion of the fibrous strip 500 being run on the machine 200.

[0060] The image is acquired by the image acquisition device 110.

[0061] The image acquisition device 110 is for example configured so that the image that it acquires comprises the entire width of the fiber strip 500 and a section of its length.

[0062] [Fig. 3] illustrates a photograph of an image acquired by the image acquisition device 110. In this photograph are visible, from right to left of the image, a first roller 210, the tank 220 containing the resin 230, a second roller 210. The fibrous strip 500 runs from right to left, as illustrated by the arrow A.

[0063] To accentuate the visibility of the fibrous strip 500 (or the resin) in the acquired image, when the fibrous strip (or the resin) is rather light, a black background 600 can for example be arranged under the fibrous strip, as illustrated in [Fig. 3]. When the fibrous strip is rather dark, a white background can be arranged under the fibrous strip. The contrast between the fibrous strip and the background of the acquired image is thus increased.

[0064] In an exemplary implementation, the image acquisition device 110 acquires two images per second.

[0065] The control method then comprises, for each acquired image, a step of analyzing the acquired image.

[0066] The acquired image to be analyzed is a digital image composed of a matrix of pixels. Each pixel is associated with a color or gray level value.

[0067] When the image acquisition device 110 is a color camera, the control method comprises a step, prior to the analysis step, of transforming the image acquired by the image acquisition device 110 to provide an image whose pixel intensity is defined in gray levels. The intensity is thus distributed on a scale ranging from 0 (very dark pixel) to 255 (very bright pixel).

[0068] The analysis step comprises a first sub-step of selecting, in the acquired image, several areas of interest, each area of ​​interest including the fibrous band.

[0069] The areas of interest in the acquired image have been previously defined, for example before starting the impregnation process.

[0070] Such a first sub-step consists of selecting, in the image acquired by the image acquisition device, several distinct zones along the fiber strip being scrolled. Each selected zone, called zone of interest, of the acquired image corresponds to a useful part of the acquired image, that is to say a narrow part of the image where the fiber strip appears.

[0071] Preferably, the areas of interest in the acquired image are of the same dimension.

[0072] For example, the area of ​​interest is a segment incorporating at least the width of the fibrous band.

[0073] In the non-limiting example of [Fig.4], three areas of interest, each in the form of a segment, are defined in the acquired image, each materialized by a vertical line: - a first area of ​​interest, upstream of the first roller 210 located to the right of the acquired image, in the direction of movement of the fiber strip, - a second area of ​​interest, downstream of the first roller 210 located to the right of the acquired image, and upstream of the tank 220, in the direction of travel of the fibrous strip, - a third area of ​​interest, downstream of the tank 220 and upstream of the second roller located to the left of the acquired image, in the direction of travel of the fibrous strip.

[0074] In the first and second zones of interest, the fibrous web 500 is dry, and has not yet undergone resin impregnation. In the third zone of interest, the fibrous web 500 is impregnated with resin.

[0075] The analysis step then comprises, for each area of ​​interest defined in the acquired image, a sub-step of extracting the signal contained in the area of ​​interest, called the useful signal.

[0076] The useful signal is expressed in pixel intensity.

[0077] When the area of ​​interest is a segment, the useful signal is in the form of a two-dimensional signal.

[0078] View (a) of [Fig.4] represents the representative curve of the useful signal of the third area of ​​interest. View (b) of [Fig.4] represents the representative curve of the useful signal of the second area of ​​interest and view (c) represents the representative curve of the useful signal of the first area of ​​interest.

[0079] Each curve represents the pixel intensity values ​​along the segment (i.e. along a vertical axis in [Fig.4]). The abscissa axis thus represents the pixel coordinates and the ordinate axis represents the pixel values, in pixel intensity.

[0080] The analysis step then comprises, for each area of ​​interest defined in the image acquired, a useful signal processing sub-step.

[0081] In one embodiment, the sub-step of processing the useful signal comprises a first phase of determining the width of the fiber strip.

[0082] In an example of implementation of this first phase, the determination of the width of the fiber band can be carried out by means of a two-core classification algorithm (commonly known by the English terminology of “clustering”). The two-core classification algorithm is an algorithm known in itself. In summary, the two-core classification algorithm makes it possible to separate the useful signal into two.

[0083] The width of the fiber band is defined from a separation threshold of the useful signal, obtained by the two-core classification algorithm.

[0084] [Fig.5] is a reproduction of view (b) of [Fig.4], in which the separation threshold 700 of the useful signal, obtained by the two-core classification algorithm, is represented by a horizontal line. In the example of [Fig.5], the separation threshold 700 is at the intensity value 81.

[0085] [Fig.5] also illustrates the width (expressed in pixels) of the fibrous band deduced from said separation threshold.

[0086] The width of the fibrous band, expressed in pixels, is defined as the difference in pixels between: - the first pixel of the curve which exceeds the separation threshold and the last pixel of the curve which exceeds the separation threshold, when the fibrous band (or the resin) is clear, or - the first pixel of the curve which is located below the separation threshold and the last pixel of the curve which is located below the separation threshold, when the fibrous band (or the resin) is dark.

[0087] Thus, in the example of [Fig.5] (case of a fibrous strip or a clear resin), the first pixel above the separation threshold is the first point of the curve above the separation threshold, starting from the left of the curve, and the last pixel above the separation threshold is the first point of the curve above the separation threshold, starting from the right of the curve.

[0088] In the example of [Fig.5], the width of the fibrous strip is 56 pixels (first pixel: 55, last pixel: 111).

[0089] The width of the fibrous strip can then be expressed in mm.

[0090] The sub-step of processing the useful signal then comprises a second phase of determining, in the width of the fiber strip, holes.

[0091] In an example of implementation of this second phase, we determine, in the set of points of the useful signal, contained in the width defined in the first phase above: - the points located below the separation threshold when the fibrous band (or resin) is clear, - the points located above the separation threshold when the fibrous band (or resin) is dark.

[0092] [Fig.6] is a reproduction of [Fig.5], where, in the width of the fibrous strip, only the points located below the separation threshold 700 are represented, illustrating the presence of a hole in the width of the fibrous strip.

[0093] The useful signal processing sub-step then comprises a third phase of determining an average light intensity and a standard deviation on the light intensity.

[0094] In an example of implementation of this third phase, the average light intensity and the standard deviation of the light intensity are determined by calculating the average and the standard deviation of the values ​​of the points of all the points of the useful signal, contained in the defined width of the fiber strip, that is to say including the points representative of holes.

[0095] All the steps and sub-steps described above illustrating the control method are repeated for each image acquired by an image acquisition device and for each area of ​​interest in an acquired image.

[0096] The areas of interest selected in each acquired image are preferably the same, that is to say they are at the same locations in the images.

[0097] Thus, during the unwinding of the fiber web, for each image acquired by an image acquisition device, and for each area of ​​interest of an acquired image, the control method advantageously makes it possible to determine four criteria in real time. Each criterion is linked to a type of defect.

[0098] The first criterion is the width of the fibrous strip. The width is an important geometric criterion during the subsequent use of the fibrous strip, in particular for the manufacture of a part in composite material.

[0099] The second criterion is the detection of the presence of holes in the width of the fibrous strip. The holes in the width of the fibrous strip provide information on the presence and number of slots.

[0100] The detection of holes on several consecutive images, at the level of the same area of ​​interest, also makes it possible to determine the length of the slits.

[0101] The third criterion, corresponding to the determination of the average light intensity, makes it possible to obtain information on the variation of the resin rate over time. If the value of the average light intensity remains constant over time, that is to say over several consecutive images, at the level of the same area of ​​interest, it is possible to deduce a homogeneity of the resin rate over the length of the fibrous strip.

[0102] When the resin is light compared to the fibrous strip, if the value of the average light intensity decreases (respectively increases) over time, at the same area of ​​interest, one can deduce a heterogeneity of the resin rate over the length of the fibrous strip, in this case a decrease (respectively an increase) of the resin rate.

[0103] When the resin is dark compared to the fibrous strip, if the value of the average light intensity increases (respectively decreases) over time, at the same area of ​​interest, one can deduce a heterogeneity of the resin rate over the length of the fibrous strip, in this case a decrease (respectively an increase) in the resin rate.

[0104] The fourth criterion, corresponding to the determination of the standard deviation of the average light intensity, makes it possible to obtain information on the variation of the resin content across the width of the fibrous strip. The greater the standard deviation of the light intensity, the more heterogeneous the resin content across the width of the fibrous strip. This may mean, for example, that one edge of the fibrous strip, or both edges, has more or less resin than the center of the fibrous strip.

[0105] The control method thus advantageously makes it possible to monitor in real time the evolution of the four criteria on the one hand at different points of the fiber strip during movement, and on the other hand over time, by the repeated acquisition of images. The control method thus makes it possible to quickly detect manufacturing deviations in the fiber strip, during the resin impregnation process, and if necessary to stop the resin impregnation process if the detected defect is prohibitive.

[0106] The control device 100 and the associated control method have been described in the context of use on a resin impregnation machine. It is clear that without departing from the scope of the invention, the control device 100 and the associated control method can be adapted in any machine using a fibrous web during movement, and can in particular also be adapted in the context of use on an automated deposition machine.

Claims

Claims

1. Method for controlling a fibrous strip (500) during movement, said control method comprising the steps of: - acquiring, periodically, an image of at least a portion of the fibrous strip (500), - for each acquired image, analyzing said acquired image to: • select several areas of interest in the acquired image, each area of interest including the fibrous strip (500), • for each area of interest, extracting the signal contained in the area of interest, called the useful signal, said useful signal being expressed in pixel intensity, and processing said useful signal to: • determine a width of the fibrous strip (500), • determine, in the determined width of the fibrous strip (500), the presence of holes, and where appropriate their length, • determine, in the determined width of the fibrous strip (500), an average light intensity and a standard deviation on the light intensity.

2. Control method according to claim 1 comprising a step prior to the step of analyzing each acquired image, of transforming the acquired image to provide an image whose pixel intensity is defined in gray levels.

3. Control method according to one of the preceding claims in which the width of the fibrous strip is determined by a two-core classification algorithm.

4. Control method according to one of the preceding claims in which the determination of the presence of holes allows the determination of a slot-type defect in the fibrous web.

5. Control method according to one of the preceding claims in which the determination of the average luminous intensity of the strip fibrous allows the determination of a heterogeneity type defect of the fibrous band, in the length of said fibrous band.

6. Control method according to one of the preceding claims in which the determination of the standard deviation of the light intensity of the fibrous strip (500) allows the determination of a heterogeneity type defect of the fibrous strip, in the width of said fibrous strip.

7. Device (100) for controlling the quality of a fiber strip during movement comprising: - an image acquisition device (110), - a control unit (150) comprising: • a module (120) for processing the acquired images, • a storage module (130) on which code instructions are recorded for the execution, by the acquired image processing module, of a control method according to one of claims 1 to 6.

8. Use of the control device (100) according to claim 7 in a resin impregnation machine.

9. Use of the control device (100) according to claim 7 in an automated deposition machine.

Images