Yarn anomaly detection and prediction system on a loom

Abstract

To provide an abnormality detection / prediction system of an original yarn on a loom, or a fuzz detection system of a filament yarn, which not only improves productivity and economic efficiency but also is useful as an ecology / green technology.SOLUTION: An abnormality detection system determines an abnormality degree of a weft yarn based on a variation degree of arrival timing of the weft yarn in weft insertion of a loom, and stops the loom when the abnormality degree is equal to or more than a predetermined value.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a raw yarn abnormality detection and prediction system on a loom, a hairiness detection system and method for filament yarn, and a weaving and processing system and method for filament yarn.

Background Art

[0002] A filament yarn is obtained by aggregating a plurality of fiber monofilaments obtained by spinning raw materials such as glass, synthetic resin, natural resin, carbon, and metal. Filament yarns are used, for example, in the manufacture of various members such as laminated plates, printed wiring boards, stent grafts, composite materials, reinforcing materials, and crack suppression materials for concrete. Among them, glass filament yarn is known as a raw material for printed wiring boards.

[0003] For example, glass fiber monofilaments of several μm to several tens of μm obtained by spinning molten glass are aggregated from several tens to several thousands to form a glass filament yarn, which is once wound around a drum to form a state called a cake. Then, the glass filament yarn is unwound from the cake and shipped in the form of a glass yarn (glass thread) with twist applied thereto. The glass yarn is woven and crosslinked, opened and / or pasted as needed, and compounded and laminated with a matrix resin such as an epoxy resin to be processed into a printed wiring board.

[0004] In addition, it may be shipped in the form of a bobbin unit in which glass yarn is wound around a bobbin, a glass roving in which several tens of glass filament yarns are bundled together, or a chopped strand in which glass filament yarns are cut into several mm to several tens of mm. The raw yarn can be used as warp and / or weft in the weaving process from the bobbin unit. Glass roving is used as a raw material for glass fiber reinforced plastics (FRP, FRTP), and chopped strands are used as reinforcing materials for thermoplastic resins.

[0005] In processes such as bundling multiple fiber monofilaments into a filament yarn, unwinding the filament yarn from a cake, weaving, or processing, fiber monofilaments come into contact with each other and rub against one another, which can cause fuzzing defects such as partial breakage or weaving of broken glass fiber monofilaments. Fuzzing defects not only worsen workability but can also affect the quality of the final product. Conventionally, methods for controlling looms have been studied from the viewpoint of suppressing weft insertion errors caused by yarn breakage in the weaving process (see, for example, Patent Document 1). [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2019-196556 [Overview of the Initiative] [Problems that the invention aims to solve]

[0007] For example, as described in Patent Document 1, a technique is known for detecting the arrival time of a weft thread propelled by an air jet.

[0008] However, conventional looms and their control methods have room for improvement from the standpoint of productivity, economy, ecology, and / or green technology.

[0009] Therefore, the present invention aims to provide an anomaly detection system, a yarn anomaly detection and prediction system on a loom, or a fuzz detection system for filament yarn, which are useful not only for improving productivity and economic efficiency but also as ecological and green technologies. [Means for solving the problem]

[0010] As a result of diligent research, the inventors have found that the above problem can be solved by detecting situations in the manufacturing process that produce low-quality intermediates or products, and by stopping the operation of the manufacturing equipment as necessary, and have completed the present invention. Accordingly, this disclosure encompasses the following aspects. <1> An abnormality detection system that determines the degree of abnormality of the weft threads based on the degree of variation in the timing of the weft thread arrival during weft insertion into a loom, and can stop the loom if the degree of abnormality exceeds a predetermined value. <2> An anomaly detection system according to item 1, wherein when the degree of anomaly is greater than or equal to a predetermined value, the determination result of the degree of anomaly is indicated by at least one selected from the group consisting of screen display, printing, sensory stimulation, and data transmission. <3> A filament yarn fluff detection system, A fluff detection system comprising: an abnormality calculation means that calculates the degree of abnormality of the weft yarn from the degree of variation in the arrival timing of the weft yarn during weft insertion of the loom using a calculation formula stored in an abnormality calculation formula storage means; and a loom abnormality stop means that sends an operation stop signal to the loom when the degree of abnormality is greater than or equal to a predetermined value. <4> The fluff detection system according to item 3, further comprising a loom abnormality alarm means that raises an alarm when the degree of abnormality exceeds a predetermined value. <5> The fluff detection system according to item 3 or 4, wherein the calculation formula includes a function or correspondence table for calculating the degree of abnormality of the weft yarn based on a series of sensor measurements during the weft insertion. <6> A fluff detection system according to any one of items 3 to 5, comprising a feature transformation means that transforms the degree of variation into a feature using a transformation formula stored in a feature transformation formula storage means. <7> The fluff detection system according to item 6, wherein the conversion formula includes a function or correspondence table that converts the sensor measurement sequence into features in the weft insertion. <8> The fluff detection system according to item 6 or 7, wherein the feature is the mean, variance, or distribution of the arrival timing. <9> A fluff detection system according to any one of items 3 to 8, comprising a weft abnormality determination means that determines whether or not the weft is abnormal using a determination formula stored in a weft abnormality determination formula storage means. <10> The fluff detection system according to item 9, wherein the determination formula includes a function or pass / fail map that determines whether the weft yarn meets quality standards during weaving based on the degree of abnormality or its progression. <11> A fluff detection system according to any one of items 3 to 10, comprising a means for collecting information on the surrounding environment of the location where the loom is installed. <12> The fluff detection system according to item 11, wherein the ambient environmental information is at least one selected from the group consisting of temperature, humidity, and wind direction. <13> A fluff detection system according to any one of items 3 to 12, comprising a weft pre-physical property information storage means that stores physical property information of a portion of the weft before the weft insertion. <14> A fluff detection system according to any one of items 3 to 13, comprising a loom characteristic information storage means that stores characteristic information of the loom. <15> The fluff detection system according to item 14, wherein the characteristic information of the loom is at least one selected from the group consisting of the model of the loom, the number of maintenance sessions, the date and time of the most recent maintenance, the maintenance method, and the model of the reed included in the loom. <16> A fluff detection system according to any one of items 6 to 8, comprising an abnormality calculation formula pre-modification means for accumulating the aforementioned feature quantities and modifying the calculation formula stored in the abnormality calculation formula storage means in batches so as to satisfy predetermined evaluation criteria based on the accumulated feature quantities. <17> Accumulate the feature quantities, sequentially learn the normal state and the abnormal state based on the accumulated feature quantities, and sequentially change the calculation formula stored in the abnormality degree calculation formula storage means based on the normal state and the abnormal state. The hairiness detection system according to any one of Items 6 to 8, comprising an abnormality degree calculation formula sequential change means. <18> The abnormality degree calculation formula sequential change means sequentially changes a parameter or a threshold value in the calculation formula. The hairiness detection system according to Item 17. <19> Detect the weft replacement by detecting a change in the column of the feature quantities and / or the stop time of the loom, and include weft change detection means for adjusting the conversion formula for the feature quantities and / or the calculation formula for the abnormality degree. The hairiness detection system according to any one of Items 6 to 8 and 16 to 18. <20> The hairiness detection system according to any one of Items 3 to 19, which is used offline. <21> The hairiness detection system according to any one of Items 3 to 19, which is used inline.

Advantages of the Invention

[0011] According to the present invention, an abnormality detection system, a raw yarn abnormality detection / prediction system on a loom, and a hairiness detection system for a filament yarn are provided. By these, it is possible to detect a situation where low-quality intermediates or products are continuously produced and stop the manufacturing apparatus as necessary. Therefore, not only productivity and economy can be improved, but also environmental considerations can be made as ecology and green technology.

Brief Description of the Drawings

[0012] [Figure 1] It is a schematic diagram schematically explaining the configuration of a hairiness detection system according to an embodiment of the present invention, showing a schematic diagram of a loom (a) and a schematic diagram of the hairiness detection system (b). [Figure 2] It is a flowchart of a hairiness detection method according to an embodiment of the present invention. [Figure 3]This is a schematic diagram illustrating the configuration of a fluff detection system according to a preferred embodiment of the present invention. [Figure 4] This is a flowchart illustrating an example of a predictive detection algorithm. [Figure 5] This is a schematic diagram illustrating the entire process flow, including a lint detection system, according to one embodiment of the present invention. [Modes for carrying out the invention]

[0013] The embodiments for carrying out the present invention (hereinafter referred to as "embodiments") will be described in detail below with reference to the drawings. However, the present invention is not limited to the embodiments or drawings described below, and various modifications are possible without departing from the spirit of the invention.

[0014] [First Embodiment] In the first embodiment, an abnormality detection system is provided that determines the degree of abnormality of the weft yarn based on the degree of variation in the timing of the weft yarn's arrival during weft insertion into the loom, and can stop the loom if the degree of abnormality exceeds a predetermined value. The abnormality detection system according to the first embodiment can detect abnormal situations that continue to produce low-quality intermediates or products, stop the loom as necessary, improve productivity and economic efficiency, and is also useful as an eco-green technology.

[0015] The anomaly detection system according to the first embodiment can be used as a device or method for detecting anomalies in fields where weft insertion is performed using warp and weft threads on a loom. For example, weft insertion is performed in weaving processes, filament yarn weaving processes, etc. When glass filament yarn is used, the anomaly detection system according to the first embodiment is preferably used in the manufacture of glass cloth and the manufacture of raw materials for printed circuit boards.

[0016] The timing of the weft yarn's arrival can be measured, for example, as pressure, pressure angle, or degree of laser beam shielding, by a device that positions the weft yarn relative to the warp yarns arranged on the reed, and a device that detects the start and end of weft insertion. The degree of variation in the arrival timing of multiple weft yarns and the determination of the degree of abnormality based thereon can be performed by a calculation component including an arithmetic processing unit and / or a storage medium. The calculation component and / or storage medium may include a display unit such as a monitor or a transmitting device such as a modem.

[0017] The predetermined values ​​for the degree of abnormality can be set in advance, one or more, and stored in a storage medium, for example, due to defects in the raw yarn, defects in the weft yarn, defects in the weft bobbin, defects in the loom, etc. The means for stopping the loom may be a part that physically stops the loom, a part that sends a stop signal to the loom, a part that transmits a stop signal to the loom, or a combination thereof, and may be positioned in contact with the loom or at a distance from the loom.

[0018] In the first embodiment of the anomaly detection system, if the degree of anomaly described above exceeds a predetermined value, it is preferable to display the determination result of the degree of anomaly and raise an alarm, and it is more preferable to display the determination result of the degree of anomaly by at least one selected from the group consisting of screen display, printing, sensory stimulation, and data transmission. Screen display and data transmission can be performed simultaneously or individually by the display unit and transmission device etc. described above. Printing is performed by printing the determination result onto a printing medium using a printing press. The sensory stimulation can be any stimulus as long as it stimulates human senses and raises an alarm, for example, sound, light emission, scent emission, sound change, color change, deformation, temperature change, etc.

[0019] The anomaly detection system according to the first embodiment can be used, for example, offline or inline.

[0020] [Second Embodiment] In the second embodiment, a filament yarn fluff detection system, An abnormality calculation means calculates the degree of abnormality of the weft yarn from the degree of variation in the arrival timing of the weft yarn during weft insertion into the loom, using a calculation formula stored in the abnormality calculation formula storage means. A loom malfunction stop means that sends a stop signal to the loom when the calculated abnormality level is above a predetermined value, and A lint detection system is provided that includes the following features.

[0021] The fluff detection system according to the second embodiment includes an abnormality calculation means and a loom abnormality stop means, which can detect situations in which the production of filament yarn or glass cloth with fluff defects continues, and stop the loom operation as necessary, thereby improving productivity and economic efficiency, and is also useful as an eco-green technology.

[0022] In the second embodiment, after converting the measured values ​​into features, the degree of abnormality can be calculated by comparing them with the features of past normal and abnormal behavior, and displayed numerically or graphically. Therefore, compared to not collecting or displaying measured values, it becomes possible to continuously detect abnormalities in the weft yarn, and if the degree of abnormality is high, the loom can be stopped, thereby minimizing the production of low-quality fabrics.

[0023] The filament yarn fluff detection system according to the second embodiment can be used, for example, offline or inline.

[0024] [Third Embodiment] In a third embodiment, a filament yarn fluff detection system is provided that combines the configurations of the first and second embodiments.

[0025] Configurations common to the first, second, and third embodiments, as well as other configurations and preferred embodiments, will be described below.

[0026] 〔composition〕 Figure 1 schematically shows the configuration of the loom 200(a) and the fluff detection system 100(b). Referring to Figure 1(a), the loom 200 includes a bobbin 210, a metal fiber detector 220, an FDP 240 with pins 230, a reed 250, a main nozzle 270 with an electro-pneumatic regulator 260, a sub-nozzle 280, a cutter 290, a feeler 300, etc. Weft insertion can be performed by unwinding the weft from the weft bobbin 210 and advancing the weft from the main nozzle 270 toward the warp threads positioned on the reed 250, for example, by an air jet. Weft insertion can be performed at any loom angle, for example, between 90° and 240°.

[0027] The main nozzle 270 and the sub-nozzle 280 are provided as nozzles for weft insertion. The main nozzle 270 is positioned upstream in the weft transport direction, while the sub-nozzle 280 is positioned downstream of the main nozzle 270 in the weft transport direction. The main nozzle 270 is provided as a single unit, while multiple sub-nozzles 280 are provided.

[0028] The main nozzle 270 injects air when the main valve (not shown) is open and stops injecting air when the main valve is closed. The main nozzle 270 or the main valve can be connected physically, electrically, or via data communication to the loom abnormality stop means of the detection system.

[0029] An electro-pneumatic regulator 260 can be connected to the main nozzle 270 to adjust the pressure of compressed air generated by an air compressor (not shown). The compressed air is supplied to the main nozzle 270 via a main valve (not shown) and used to propel the weft. The electro-pneumatic regulator 260 can be connected physically, electrically, or via data communication to a loom abnormality stop means of a detection system.

[0030] Multiple sub-nozzles 280 inject or stop air depending on the open / closed state of their respective sub-valves (not shown). The multiple sub-nozzles 280 can be arranged along the longitudinal direction of the reed 250 as shown in Figure 1, and / or the direction of the sub-nozzles can be set to form an angle of approximately 45° with respect to the injection direction of the main nozzle 270. Similar to the main nozzle 270, valves, air compressors, regulators, loom malfunction stop means of the detection system, etc., can be connected to the sub-nozzles 280.

[0031] The Feeler 300 detects whether the weft thread has reached a predetermined position when air is injected from the main nozzle 270 and the sub-nozzle 280 to insert the weft thread, and is sometimes called a reach sensor (13). The predetermined position can be set in the longitudinal direction of the reed 250, on the side furthest from the main nozzle 270, at the weft insertion end, according to the weave width of the fabric.

[0032] The feeler 300 is composed of, for example, an optical sensor. The feeler 300 can output a detection signal when the leading edge of the weft yarn, which is being conveyed in the longitudinal direction of the reed 250 by air jets from the weft insertion nozzle, reaches a predetermined position. Therefore, the timing of the weft yarn arrival, when the weft yarn reaches the predetermined position, can be the timing when the feeler 300 outputs a detection signal.

[0033] Referring to Figure 1(b), the abnormality or fluff detection system 100 includes an input device 1 equipped with a main nozzle sensor 11, a sub-nozzle sensor 12, and a reach sensor 13, a data processing device 2 that operates by program control, a storage device 3 that stores information, and an output device 4 such as a display device, a mobile terminal device, or a printing device.

[0034] In the input device 1, the main nozzle sensor 11, the sub-nozzle sensor 12, and the arrival sensor 13 can be physically, electrically, or data-communicated to the main nozzle 270, the sub-nozzle 280, and the feeler 300, respectively. Alternatively, the feeler 300 itself may be the arrival sensor 13.

[0035] The main nozzle sensor 11 detects, for example, the timing of the weft air jet start, the number of weft insertions, the weft insertion pressure, the type of weft, and the number of weft threads, and outputs them as signals. In particular, it is preferable for the main nozzle sensor 11 to detect and output pressure, the number of weft threads, or the number of weft insertions.

[0036] The sub-nozzle sensor 12 detects, for example, the number of weft insertions, weft insertion pressure, number of weft insertion direction corrections, type of weft thread, number of weft threads, and main nozzle status, and outputs them as signals. In particular, it is preferable for the sub-nozzle sensor 12 to detect and output pressure.

[0037] The arrival sensor 13 detects the timing of weft arrival when the weft reaches a predetermined position, the number of weft insertions, the weft insertion pressure, the type of weft, the number of weft threads, etc., and outputs them as signals. In particular, it is preferable that the arrival sensor 13 detects and outputs the timing of weft arrival using an optical sensor. Regarding weft insertion, the arrival timing can be understood as equal to the arrival angle.

[0038] The storage device 3 comprises a feature conversion formula storage unit 31, an abnormality calculation formula storage unit 32, and a weft abnormality determination formula storage unit 33.

[0039] The feature conversion formula storage unit 31 stores in advance conversion formulas, such as functions or correspondence tables, for converting the measurement value sequences of each sensor into features used to calculate the anomaly score. Examples of functions include basic statistics such as mean, variance, or distribution.

[0040] The abnormality calculation formula storage unit 32 stores in advance calculation formulas, such as pre-prepared functions or correspondence tables, used to calculate the abnormality of the weft yarn based on each feature quantity. Examples of functions include regression equations or decision trees based on if-then rules.

[0041] The weft abnormality determination formula storage unit 33 stores in advance determination formulas such as functions, pass / fail maps, or correspondence tables used to determine whether the weft yarn being woven meets quality standards based on the degree of abnormality or its progression. An example of a function is a sequential test that determines an abnormality only when the degree of abnormality exceeds a certain threshold for a certain period of time.

[0042] The data processing device 2 comprises a feature transformation means 21, an anomaly calculation means 22, and a weft anomaly determination means 23.

[0043] The feature conversion means 21 takes the sequence of measured values ​​provided by the input device 1 as input to a function or correspondence table and converts it into features using a conversion formula stored in the feature conversion formula storage unit 31. The features may be the mean, variance, or distribution of the arrival timing as described above.

[0044] The anomaly calculation means 22 takes the feature quantities of the measured values ​​converted by the feature quantity conversion means 21 as input and calculates the anomaly using the calculation formula stored in the anomaly calculation formula storage unit 32.

[0045] The weft abnormality determination means 23 takes the degree of abnormality of the measured value calculated by the degree of abnormality calculation means 22 as input and uses the determination formula stored in the weft abnormality determination storage unit 33 to determine whether or not the weft is abnormal.

[0046] [Operation] The operation of the detection system according to the first, second, or third embodiment will be described in detail with reference to Figures 1 and 2. Figure 2 shows a flowchart of the fluff detection method.

[0047] The sequence of measured values ​​provided by the input device 1 is supplied to the feature conversion means 21 (step A1).

[0048] The feature conversion means 21 inputs this sequence of measured values ​​into a conversion formula stored in the feature conversion formula storage unit 31. This conversion formula outputs the mean, variance, and distribution of the measured values ​​as features, provides these features to the anomaly score calculation means 22, and simultaneously outputs the changes in the features to the output device 4 as numerical values ​​or graphs (step A2).

[0049] Next, the abnormality calculation means 22 receives feature quantities such as the mean, variance, and distribution of the measured values ​​converted by the feature quantity conversion means 21, and calculates the abnormality using the calculation formula stored in the abnormality calculation formula storage unit 32. This calculation formula calculates a high abnormality when there is a high probability of abnormality, based on the feature quantities during past normal and abnormal behavior. This abnormality is provided to the weft abnormality determination means 23, and at the same time, the trend of the abnormality is output to the output device 4 as a numerical value or graph (step A3).

[0050] Next, the weft abnormality determination means 23 receives the abnormality score of the measured value calculated by the abnormality score calculation means 22 and determines whether the weft is abnormal or not using the determination formula stored in the weft abnormality determination storage unit 33. This determination formula determines that an abnormality has occurred when there is a high probability that an abnormality has occurred, such as when the abnormality score sequence shows a high index for a certain interval, and at the same time outputs the progress of the abnormality determination result to the output device 4 as a numerical value or graph (step A4).

[0051] [Preferred Embodiments] Figure 3 schematically shows the configuration of a detection system according to a preferred embodiment of the present invention. In addition to the configuration described above, the abnormality or fluff detection system 100 according to the preferred embodiment includes at least one selected from the group consisting of a loom installation environment sensor 14, a weft pre-physical property information storage unit 34, a loom characteristic information storage unit 35, an abnormality calculation formula pre-modification means 24, an abnormality calculation formula sequential modification means 25, a weft change detection means 5, a loom abnormality alarm device 41, and a loom abnormality stop device 42.

[0052] The loom installation environment sensor 14 collects data on the surrounding environment, such as temperature, humidity, and wind direction, at the location where the loom is installed. By using this data, the influence of the environment in which the loom is installed can be reflected, and if the environment changes, that influence can be incorporated into the calculation of the anomaly level.

[0053] The weft yarn pre-physical property information storage unit 34 stores physical property information, such as the results of a detailed inspection of a portion of the weft yarn using an inspection machine. Examples of inspection results include the number of weft yarn breaks and slack. By incorporating this pre-information as features, it becomes possible to calculate the degree of abnormality with greater accuracy.

[0054] The loom characteristic information storage unit 35 stores characteristic information of each loom, such as the manufacturer and model, maintenance methods and frequency, the date and time of the most recent maintenance, and the type and model of the reed included in the loom. By adjusting parameters or thresholds based on this information when calculating the degree of abnormality, it becomes possible to calculate the degree of abnormality with higher accuracy for each loom.

[0055] The abnormality calculation formula pre-modification means 24 stores feature quantities collected and transformed over a certain period in the past, and modifies the calculation formula of the abnormality calculation formula storage unit 32 based on these feature quantities so as to satisfy certain or predetermined evaluation criteria. It is preferable that this modification of the calculation formula be performed in batches. As evaluation criteria, some may output not only the abnormality score but also subdivided types of abnormalities.

[0056] The sequential anomaly calculation formula modification means 25 accumulates feature quantities collected and transformed over a certain period in the past, and sequentially learns normal and abnormal states based on these feature quantities. Accordingly, it sequentially modifies the parameters or thresholds of the calculation formula in the anomaly calculation formula storage unit 32. This makes it possible to calculate anomalies based on evaluation criteria that follow gradual changes in the surrounding environment, loom, and weft yarn.

[0057] The weft change detection means 5 detects the replacement of a weft yarn by detecting a change in the feature quantity sequence and / or the loom's stop time, and adjusts the conversion formula to the feature quantity and / or the calculation formula for the anomaly score. By using this calculation result, it is possible to automatically determine that a weft yarn has been replaced, and it becomes easy to retrieve the corresponding weft yarn information from the weft yarn pre-physical property information storage unit.

[0058] The loom malfunction alarm device 41 integrates the results collected by the output device 4 and, if it determines that the weft yarn is abnormal, raises an alarm to the loom operator, prompting them to decide whether or not to continue operating the loom.

[0059] The loom malfunction stop device 42 integrates the results collected by the output device 4 and, if it determines that the weft yarn is abnormal, sends a stop signal to the loom to forcibly stop its operation.

[0060] [Predictive detection algorithm] Figure 4 shows an example flowchart of the predictive detection algorithm. Assuming arrival timing = arrival angle, the predictive detection algorithm collects loom data such as the arrival angle of the weft yarn, converts the arrival angle of the weft yarn into statistics, and then predicts and detects the number of fluffs (n1, n2, n3, n4, n5, n6, n7…) based on the standard deviation, skewness, and kurtosis of the arrival angle. While a regression tree is given as an example of the predictive detection algorithm, it is not limited to this, and regression methods, classification methods, neural networks, SVMs, random forests, etc., as described in Trevor Hastie, Robert Tibshirani, and Jerome Friedman's "Fundamentals of Statistical Learning - Data Mining, Inference, and Prediction" (Kyoritsu Shuppan), may be used. Furthermore, the hierarchy of the predictive detection algorithm may be single-layered or multi-layered, and may include combinations, substitutions, or rearrangements of the standard deviation, skewness, and kurtosis of the arrival angle.

[0061] The predictive detection algorithm shown in Figure 4 is incorporated into the fluff detection system, preferably into at least one of the abnormality calculation means, the loom abnormality stop means, the abnormality calculation formula pre-modification means 24, the abnormality calculation formula sequential modification means 25, and the weft change detection means 5, so that the loom does not continue to produce low-quality or poor-quality cloth.

[0062] [Method for detecting fluffiness in filament yarn] The present invention's method for detecting fuzz in filament yarn is performed using a fuzz detection device equipped with the above-described abnormality calculation means and loom abnormality stop means. More specifically, the method for detecting fuzz in filament yarn involves the following steps: An abnormality calculation step that uses a calculation formula stored in an abnormality calculation formula storage means to calculate the abnormality of the weft yarn from the degree of variation in the timing of the weft yarn arrival during weft insertion into the loom, If the above abnormality level exceeds a predetermined value, a machine malfunction stop process is initiated, which sends a stop signal to the machine. Execute this.

[0063] A method for detecting fluffiness in filament yarn can be performed, for example, according to the flowchart shown in Figure 2.

[0064] [Overall process flow of weaving and finishing] Figure 5 schematically shows the overall weaving process flow including the fluff detection system described above. The surface inspection machine 400 (of the yarn), the sensing-enabled loom 201, and the fluff inspection machine 500 are arranged from upstream to downstream of the process flow, and the abnormality or fluff detection system according to the present invention can be incorporated into the inspection step of the process flow.

[0065] Alternatively, the memory device 3, data processing device 2, output device 4, etc. of the abnormality or fluff detection system may reside in a location isolated from the system, with data communication, or on the cloud.

[0066] The overall process flow is broadly divided into application infrastructure 6 and analysis infrastructure 7, and these two infrastructures are either overlapping or linked by data servers 678.

[0067] The application platform 6 automates data collection and processing to improve data accuracy and reduce the workload of operators. The application platform 6 transmits and receives data such as yarn diameter and yarn defects (ii) between the data server 678 and the surface inspection machine 400, transmits and receives data such as waveform data and weft data during weaving or weft insertion (iii), and fluff detection data (iv) between the data server 678 and the sensing-enabled loom 201, displays the transmitted and received data on the display device 9 as desired, and transmits and receives fluff inspection data (v) between the data server 678 and the fluff inspection machine 500.

[0068] The surface inspection machine 400 inspects the surface of the yarn or bobbin, for example using the MT method, and collects data (ii) such as yarn diameter and yarn defects, which it sends to the data server 679. On the other hand, it can receive signals regarding the pass / fail status of the yarn from the application base 6 and decide whether to use the yarn or change it.

[0069] The sensing-enabled loom 201 is equipped with, for example, a pressure sensor and a reach (flying) sensor, and uses these to collect data (iii) such as waveform data and weft data during weaving or weft insertion and transmit it to the data server 679. On the other hand, the sensing-enabled loom 201 receives the fluff detection data (iv) that has undergone preprocessing based on a prediction model stored in the data server, followed by inference processing, and can decide whether to operate or stop the loom.

[0070] The fluff inspection machine 500 measures the presence or details of fluff in the obtained fabric or cloth and sends fluff inspection data (v) to the data server. On the other hand, it receives a signal regarding the pass / fail status of the fabric or cloth from the application platform 6 and can further transmit the information to the operator or the surrounding environment, for example, via an enhanced voice service (EVS).

[0071] The analysis platform 7 accelerates the analysis cycle, improves detection performance, and adapts to changes in the state or environment of the inspection machine or loom. The analysis platform includes analysis means 10, which include a computer, a processing unit, and a storage medium. It receives various data (ii~v) from the data server 678, analyzes them, constructs a predictive model (i), and sends it to the data server 678. Through data transmission and reception between the data server 678 and the analysis means, and the associated machine learning and deep learning, the analysis platform 7 can contribute to accelerating the analysis cycle and improving detection performance.

[0072] According to the detection process described above, it is possible to change potentially defective yarn before use, stop the loom when situations that would continue to produce low-quality intermediates or products are detected, and instantly determine whether the resulting fabric or cloth is acceptable. This improves productivity and economic efficiency, and is also environmentally friendly from a resource perspective. [Explanation of Symbols]

[0073] 100 detection systems 1 Input device 11 Main nozzle sensor 12 Sub-nozzle sensor 13. Arrival Sensor 14. Loom installation environment sensor 2 Data Processing Devices 21 Feature transformation means 22 Abnormality degree calculation means 23. Weft Anomaly Detection Method 24. Pre-modification method for calculating abnormality level 25. Sequential modification means for calculating the degree of abnormality 3 Storage device 31 Feature Transformation Memory Unit 32 Abnormality degree calculation formula storage unit 33. Weft Anomaly Determination Formula Memory Unit 34. Weft yarn pre-physical property information storage unit 35. Loom characteristic information storage unit 4 Output device 41 Loom abnormality alarm device 42 Loom abnormal stop device 5. Weft change detection means

Claims

1. An anomaly detection system that determines the degree of abnormality of the weft yarn based on the degree of variation in the arrival timing of the weft yarn during weft insertion into a loom, and can stop the loom if the degree of abnormality is greater than or equal to a predetermined value, wherein the weft yarn is a glued glass filament, the weft insertion is of the air jet type, the arrival timing is detected based on whether or not the weft insertion end has reached at least the end side, and the system includes a loom installation environment information collection means for collecting information on the surrounding environment of the location where the loom is installed, and the surrounding environment information is at least one selected from the group consisting of temperature, humidity, and wind direction.

2. An anomaly detection system according to claim 1, wherein, when the degree of anomaly is greater than or equal to a predetermined value, the determination result of the degree of anomaly is indicated by at least one selected from the group consisting of screen display, printing, sensory stimulation, and data transmission.

3. The abnormality detection system according to claim 1 or 2, further comprising a loom abnormality alarm means that raises an alarm when the degree of abnormality exceeds a predetermined value.

4. An anomaly detection system according to any one of claims 1 to 3, comprising a feature transformation means that transforms the degree of variation into a feature using a transformation formula stored in a feature transformation formula storage means.

5. The anomaly detection system according to claim 4, wherein the conversion formula includes a function or correspondence table that converts the sensor measurement value sequence into feature quantities in the latitude input.

6. The anomaly detection system according to claim 4 or 5, wherein the feature quantity is the mean, variance, or distribution of the arrival timing.

7. An anomaly detection system according to any one of claims 1 to 6, comprising a weft anomaly determination means that determines whether or not the weft is abnormal using a determination formula stored in a weft anomaly determination formula storage means.

8. The abnormality detection system according to claim 7, wherein the determination formula includes a function or pass / fail map that determines whether or not the weft yarn meets quality standards during weaving based on the degree of abnormality or its progression.

9. An anomaly detection system according to any one of claims 1 to 8, further comprising a weft pre-physical property information storage means that stores physical property information of a portion of the weft before the weft insertion.

10. An anomaly detection system according to any one of claims 1 to 9, comprising a loom characteristic information storage means for storing characteristic information of the loom.

11. The abnormality detection system according to claim 10, wherein the characteristic information of the loom is at least one selected from the group consisting of the model of the loom, the number of maintenance sessions, the date and time of the most recent maintenance, the maintenance means, and the model of the reed included in the loom.

12. An anomaly detection system according to any one of claims 4 to 6, comprising: an anomaly calculation means for calculating the anomaly using a calculation formula stored in an anomaly calculation formula storage means; and an anomaly calculation formula pre-modification means for accumulating the feature quantities and modifying the calculation formula stored in the anomaly calculation formula storage means in batches so as to satisfy predetermined evaluation criteria based on the accumulated feature quantities.

13. An anomaly detection system according to any one of claims 4 to 6, comprising: an anomaly calculation means for calculating the anomaly using a calculation formula stored in an anomaly calculation formula storage means; an anomaly calculation formula sequential modification means for accumulating the feature quantities, sequentially learning normal and abnormal states based on the accumulated feature quantities, and sequentially modifying the calculation formula stored in the anomaly calculation formula storage means based on the normal and abnormal states.

14. The anomaly detection system according to claim 13, wherein the sequential modification means for the anomaly degree calculation formula sequentially modifies the parameters or thresholds in the calculation formula.

15. An anomaly detection system according to any one of claims 4 to 6 and 12 to 14, comprising a weft change detection means for detecting the change of the weft yarn by detecting a change in the column of the feature quantities and / or the stopping time of the loom, and adjusting the conversion formula to the feature quantities and / or the calculation formula for the degree of anomaly.

16. An anomaly detection system according to any one of claims 1 to 15, for use offline.

17. An anomaly detection system according to any one of claims 1 to 15, for use in line.

Images