Apparatus and method for detecting defects in a spinning mill and estimating one or more causes of the defects - Patents.com

Abstract

To detect a defect in a spinning mill and estimate one or more causes of the defect.SOLUTION: A spinning mill (M) comprises a plurality of textile machines for continuously processing a textile material. An electronic device is configured to receive parameter information of the one or more textile machines and the one or more textile materials, detect a defect (dF) and its location by identifying the parameter information of the textile materials that is different from reference information, access configuration information of the textile machines in the spinning mill (M), access knowledge base information on knowledge about possible causes of a possible defect in the spinning mill (M), and apply the parameter information, the configuration information, and the knowledge base information to one or more machine learning algorithms to estimate the one or more causes (eS1, eS2) of the defect (dF).SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present disclosure relates to an apparatus and method for detecting defects in a spinning mill and for estimating one or more causes of the defects.

[0002] 2. Description of Related Art A spinning mill typically has numerous textile machines for converting natural fibers, synthetic fibers, and their blends into yarns of the desired quantity and quality. Several types of textile machines may be arranged along the production line, such as textile machines for fiber preparation, textile machines for spinning preparation, ring spinning frames, compact spinning frames, rotor spinning frames, air-jet spinning frames, automated machines, winders, texturing machines, etc. A spinning mill may have hundreds of textile machines of different types and cover an area of ​​hundreds of square meters. Furthermore, textile machines may have hundreds or even thousands of components, particularly components for processing the parallel-running textile material. During the process of converting natural and synthetic fibers and their blends into yarns, the textile material may be opened and cleaned, converted into card slivers, processed by drawing frames and / or roving frames, spun into yarns, automatically transported from one textile machine to the next, rewound to ensure proper yarn performance in the subsequent production process, and textured to provide a textile appearance, such as for flat filament yarns. To provide a desired production capacity along a continuous production line, textile machines or textile machine components may include parallel configurations. Thus, a previous processing step may be consecutively followed by a subsequent processing step, in which case the previous processing step may require fewer or more parallel textile machines or components for processing the textile material than the subsequent processing step. For example, a fiber preparation step of delivering sliver may require fewer parallel machines or components than a step of spinning the sliver into yarn. Furthermore, for purposes of flexibility, redundancy, machine maintenance, etc., additional machines may be deployed that can provide alternative or additional production capacity during certain periods. Furthermore, depending on, for example, production schedules, fewer textile machines may be needed at night than during the day, and certain textile machines may be shut down for certain periods. For example, production during a later period may include a different set of textile machines than during an earlier period.

[0003] In the operation of a spinning mill, various parameters may be acquired. The acquired parameters may relate to parameters of the textile machine and / or parameters of the textile material, such as natural fibers, synthetic fibers and their blends, slivers, rovings, yarns, etc. The textile machine parameters may relate to physical parameters, chemical parameters, etc., such as rotation speed, power consumption, chemical properties of lubricants, etc. The textile material parameters may relate to physical parameters, chemical parameters, etc., such as thickness, density, chemical properties of fibers, etc. The spinning mill may have multiple sensors for acquiring the parameters of the textile machine and / or the textile material. Acquiring the parameters of the textile machine and / or the textile material may require taking samples from the textile machine and / or the textile material and testing the samples in a laboratory, in particular to determine the mechanical / physical and / or chemical properties of the samples.

[0004] In spinning mill operations, parameters of textile machines can be continuously or periodically acquired and monitored, and if the parameters differ from predetermined reference information, it can be concluded that a defect has occurred, and corrective measures can be taken automatically or manually. The cause of a defect related to a textile machine, for example, a difference in rotation speed from the reference information, can often be accurately and quickly determined and corrected. However, without being able to acquire every single parameter of the textile machines in a spinning mill, the spinning mill will not have a sufficient number of sensors to detect all defects that may occur in the textile machines, or if a sensor is broken, the detection of defects in the spinning mill by comparing the parameters of the textile machine with the reference information may be insufficient.

[0005] Therefore, in addition to or instead of acquiring parameters of the textile machine, parameters of the textile material may additionally or alternatively be acquired and monitored during production. If the parameters of the textile material differ from the reference information, it can be concluded that a defect has occurred. However, the cause of the defect may not only be related to the textile machine that produced the defective textile material, but also, alternatively or additionally, to one of the preceding upstream processing steps. For example, an upstream textile machine may not have the required sensor or may have a broken sensor, making it impossible to detect that a defect has occurred without cooperation with a subsequent processing step. Thus, the cause of the defect that caused the detected defect remains unknown. In today's spinning mills, experienced production staff and / or external consultations are required to investigate the cause of defects detected by comparing the textile material with the reference information. However, experienced staff and / or external consultations may not always be available, resulting in significant delays in correcting the cause of the defect.

[0006] It is of utmost importance to correct the cause of defects in the spinning mill as quickly as possible, especially to minimize production losses.

[0007] EP 3175025 A1 discloses a system for monitoring a spinning line. A detection device is associated with the textile machine and with main storage means for processing large amounts of data (big data) in order to realize predictive maintenance. Physical values ​​of the components of the textile machine, such as operating parameters, for example the temperature of the support structure, the pressure in the suction ducts, acceleration to detect vibrations, the distance between the various engines of the machine, the current absorbed by the electric motors, the force to detect belt tension, a webcam providing an image of the machine, are detected.

[0008] EP 3170778 A1 discloses a textile management system including a plurality of textile machines, the textile machines including a plurality of fiber processing units, a control device configured to control the fiber processing units, one or more common devices provided in all or some of the fiber processing units, and a management device configured to manage the textile machines. Information on the status of the fiber processing units / common devices is acquired, and a first status acquisition unit is configured to acquire information on the status of the fiber processing units, information on the status of the common devices, and information on conditions (e.g., temperature, humidity at the installation site), typical examples of which include a yarn monitoring device for detecting yarn thickness, a vibration sensor for detecting vibration in the winding section, a noise sensor for detecting noise in the yarn joining section, and a management device for managing the textile machine for analyzing the status of the fiber processing units (e.g., detecting defects) in order to manage periodic maintenance.

[0009] EP 2 352 867 A1 discloses the monitoring of a production process in a textile plant, such as a spinning mill, where raw materials are processed into intermediate and final products in several processing steps during the production process. Parameters of the raw materials, intermediate and / or final products are measured in at least two different processing steps and stored in a database and linked to an index file, so that the quality used for a lot is as close as possible to the lot quality, which may also be referred to as the lot quality required to provide the yarn quality provided to obtain a supply agreement.

[0010] EP 0 556 359 A1 discloses defect diagnosis based on a fundamental knowledge of signals indicative of the quality of the produced product, e.g., spectrograms obtained by inspecting the parameters of a fiber assembly or fiber filament. A first evaluation unit is used to detect characteristic deviations in the spectrogram and generate so-called defect descriptors for each deviation. A second evaluation unit determines possible faults for each defect descriptor based on the fundamental knowledge. Resulting defects can be detected automatically and diagnosed quickly and reliably. This system requires an accurate description of the structure of each textile machine. Subsystems may be subdivided until the component (e.g., cylinder, strap) responsible for the defect is finally reached. Each object is described by a node in a tree structure.

[0011] EP 0 685 580 A1 discloses recording the causes of product defects in yarns, rovings and slivers. The recorded defects are represented according to predetermined parameters to provide an error pattern. The error pattern is used to create a reference value for the cause of the defect.

[0012] https: / / towardsdatascience.com / top-10-algorithms-for-machine-learning-beginners-149374935f3c discloses machine learning algorithms and possible real-life applications.

[0013] Brief Summary of the Invention There may be a need for an apparatus and method for detecting defects in a spinning mill and estimating one or more causes of the defects.In particular, there may be a need for an apparatus and method for detecting defects in a spinning mill and estimating one or more causes of the defects that overcomes at least some of the shortcomings of the prior art.In particular, there may be a need for an apparatus and method for detecting defects in a spinning mill and estimating one or more causes of the defects that allows for rapid and accurate estimation of one or more causes of the defects.

[0014] Such a need can be met by the subject matter of the independent claims. Advantageous embodiments are set forth in the dependent claims.

[0015] One aspect of the present invention relates to an electronic device for detecting defects and estimating one or more causes of the defects in a spinning mill having a plurality of textile machines for continuously processing textile material, the electronic device being configured to perform the steps of receiving parameter information of one or more textile machines and one or more textile materials, detecting defects and their locations by identifying parameter information of the textile materials that differs from reference information, accessing configuration information of the textile machines of the spinning mill, accessing knowledge base information related to knowledge about possible causes of possible defects in the spinning mill, and applying the parameter information, configuration information, and knowledge base information to one or more machine learning algorithms to estimate one or more causes of the defects.

[0016] The parameter information received by the electronic device may include parameter values ​​of components of the textile machine, such as the rotation speed of the rotor, the power consumption of the textile machine, the quality of the lubricant, etc. Furthermore, the parameter information received by the electronic device may include parameter values ​​of the textile material, such as the diameter of the textile material, the fiber density of the textile material, etc. The parameter information received by the electronic device may relate to parameter values ​​obtained by electronic sensors, parameter values ​​determined in laboratory analysis results, etc.

[0017] The acquired textile material parameters may include, in particular, one or more of fiber moisture, fineness, fiber length, fiber length uniformity, fiber strength, fiber neps, fiber maturity, fiber color, fiber waste, sliver / roving count / hank, sliver / roving weight deviation, sliver / roving uniformity, thick spots, thin spots, twist, yarn strength, yarn elongation properties, tensile strength, twist and count, yarn hairiness, yarn friction resistance wear properties, yarn color, and others. Similarly, the textile material parameters may include molecular information, such as information used for molecular tagging (including DNA tagging). Textile machine parameters may be acquired, for example, by taking lubricant samples to determine the lubricant quality. Similarly, in variants of the invention, textile machine and / or textile material parameters may be based on external information sources, such as, for example, quality data of textile material suppliers (e.g., raw cotton quality information), type designations / product numbers and / or material properties and / or expiration dates of consumables such as lubricants.

[0018] In another variant of the invention, measured values ​​of climatic conditions (e.g., temperature, humidity, solar radiation, solar position (measured or based on an external source and / or date and / or time)) in raw material warehouses and / or intermediate product warehouses and / or finished product warehouses may be recorded as parameters of textile machines and / or textile materials or as environmental parameters. For example, climatic parameters (e.g., temperature, humidity, solar radiation, solar position (measured or based on an external source and / or date and / or time)) may be recorded in the area of ​​at least one department of a spinning mill, in particular in the area of ​​a specific textile machine and / or at least one individual textile machine and / or the smallest individual processing unit of at least one individual textile machine. This allows, for example, weather-related production effects to be taken into account. By acquiring such parameters in individual machines or individual processing units, general or intermittent problems in the climate management of a spinning mill can be revealed. In this way, incorrectly adjusted or arranged nozzles of the air conditioning system can be detected. Similarly, the possible negative effects on production of open exterior doors or sunlight through windows can be determined. It is also conceivable to obtain information regarding the presence and identity of textile workers as parameters.

[0019] If the parameter information of the textile material differs from the reference information, the electronic device detects that the textile material is no longer being produced based on the specified quantity and / or quality, and detects a defect in the spinning mill. The location of the defect corresponds to the specific parameter information. For example, if the diameter of a specific textile material produced by a specific textile machine differs from the reference diameter, the defect is located in this specific textile machine. However, the cause of the defect may not be located solely in this specific textile machine; the cause of the defect may also be located in any upstream textile machine that was involved in producing the specific textile material. In the prior art, determining the cause of the defect requires experienced production staff or external consultation. Because continued production at a reduced quantity or quality is unacceptable, the spinning mill or the specific textile machine must be shut down until the cause of the defect is determined and corrected.

[0020] The electronic device is configured to access configuration information of textile machines in the spinning mill, which may include, for example, information about the type of textile machine, the location of the textile machines, the transport systems between the textile machines, recent maintenance, etc.

[0021] The electronic device is configured to access knowledge base information relating to knowledge about possible causes of possible defects in the spinning mill, for example the knowledge base information may include information that for a particular erroneous variation in the diameter of the fiber material the cause of the defect is more likely to be in the drawing frame than in another textile machine, for a particular erroneous fiber density the cause of the defect is less likely to be in the drawing frame, etc.

[0022] The electronic device is configured to apply the information, i.e., parameter information, configuration information, and knowledge base information, to a machine learning algorithm to estimate the cause of the defect, which allows for a highly robust and powerful estimation of the cause of the defect based on the received and accessed information, thereby allowing for a fast and accurate estimation of the cause of the defect.

[0023] In some embodiments, the apparatus includes display means, such as a computer monitor, for displaying one or more causes of the defect, thus providing fast and accurate technical information about the spinning mill, thereby enabling fast and accurate corrective action to be taken.

[0024] In some embodiments, the device is further configured to automatically take corrective action, for example, the device includes a network interface and is configured to use information about the one or more causes of the defect to control the textile machine via a computer network and each network interface equipped on the textile machine, thereby allowing corrective action to be taken quickly and accurately.

[0025] In one preferred embodiment of the present invention, the device is further configured to perform the step of identifying parameter information of the fiber material that differs from the reference information by comparing the difference between one or more parameter values ​​and one or more reference values ​​with a threshold value. Defect detection is also possible in the case of complex quantity and / or quality requirements defined by multiple reference values, such as diameter and fiber density of the fiber material.

[0026] In one preferred embodiment of the invention, the apparatus is further configured to perform the step of using the configuration information to determine one or more possible or likely configurations of the textile machine from the location of the defects and one or more causes of the defects, whereby the machine learning algorithm can be limited to the possible or likely causes of the defects, thereby improving accuracy and speed.

[0027] In one preferred embodiment of the present invention, the apparatus is further configured to perform the step of using the configuration information to determine one or more unlikely or unprobable sequences of the textile machines from the locations of the defects and one or more causes of the defects. Improbable or unprobable causes of the defects can be excluded from the machine learning algorithm, thereby improving accuracy and speed. For example, a sequence including textile machines that have recently undergone maintenance is less likely than a sequence including textile machines that have undergone maintenance a long time ago.

[0028] In a preferred embodiment of the present invention, the apparatus is further configured to perform the step of using one or more of the time-dependent parameter information and the time-dependent configuration information. When a defect is detected, the apparatus can use previous parameter information and / or previous configuration information to deduce one or more causes of the defect. For example, a different set of textile machines may be involved at night than during the day, in which case the machine learning algorithm may filter out unlikely causes of the defect, thereby improving accuracy and speed.

[0029] In one preferred embodiment of the present invention, the apparatus is further configured to perform the step of applying the parameter information, configuration information, and knowledge base information to one or more machine learning algorithms selected from linear regression techniques, logistic regression techniques, support vector machines, decision trees, random forest techniques, K-nearest neighbors, K-means techniques, naive Bayes classifiers, and principal component analysis techniques, in particular taking into account one or more of insufficient parameter information, erroneous parameter information, insufficient configuration information, erroneous configuration information, insufficient knowledge base information, and erroneous knowledge base information. The prior art does not disclose or suggest applying spinning mill information to machine learning algorithms consisting of the listed techniques. The listed techniques have the advantage of enabling rapid and accurate estimation of the cause of defects. In particular, the listed techniques and algorithms have the advantage of being robust even in cases of insufficient information.

[0030] In one preferred embodiment of the present invention, the device is further configured to request and / or access supplemental configuration information and apply the supplemental configuration information to one or more machine learning algorithms. In particular, the machine learning algorithms enable designs that can provide information on whether the supplemental configuration information can improve the accuracy of estimating one or more causes of the defect. For example, parameter information and / or configuration information may be insufficiently received, and further information may be requested and / or accessed by the electronic device. For example, the additional information may relate to rotor wear or recent replacement that the device did not initially receive.

[0031] In one preferred embodiment of the invention, the device is further configured to perform the step of determining information on when replacement, repair, modification and / or different adjustment of the textile machine and / or components of the textile machine related to one or more causes of the defect is required, for example the device is configured to access a database storing causes of the defect associated with information on when replacement, repair, modification and / or different adjustment of the textile machine and / or components of the textile machine is required.

[0032] In one preferred embodiment of the present invention, the apparatus is further configured to perform the step of determining information regarding maintenance work on one or more additional textile machines and / or one or more additional components of the textile machines, particularly if downtime of one or more textile machines of the spinning mill occurs as a result of maintenance work related to one or more causes of defects. Maintenance work may result in downtime of the spinning mill or textile machine. For example, if a roving frame of the textile machine that receives textile material from the drawing and sliver coiler unit needs to be repaired, technical information related to the maintenance work of the drawing and sliver coiler unit can be provided to reduce the risk of future causes of defects in the drawing and sliver coiler unit.

[0033] In one preferred embodiment of the present invention, the apparatus is further configured to receive feedback information indicating whether a correction of the defect occurred as a result of a maintenance action related to one or more causes of the defect, and to update the knowledge base information accordingly. After a cause of the defect is successfully corrected, the knowledge base information can be updated to improve accuracy in cases of similar or identical defects.

[0034] Another aspect of the invention relates to a method for detecting defects and estimating one or more causes of the defects in a spinning mill having a plurality of textile machines for continuously processing textile material, the method comprising the steps of receiving, using an electronic device, parameter information of one or more textile machines and one or more textile materials, detecting defects and their locations by identifying, using the electronic device, parameter information of the textile materials that differs from reference information, accessing, using the electronic device, configuration information of the textile machines of the spinning mill, accessing, using the electronic device, knowledge base information related to knowledge about possible causes of possible defects in the spinning mill, and applying, using the electronic device, the parameter information, the configuration information, and the knowledge base information to one or more machine learning algorithms to estimate one or more causes of the defects.

[0035] In one preferred embodiment of the present invention, the method further comprises the step of identifying parameter information of the textile material that differs from the reference information by comparing the difference between one or more parameter values ​​and one or more reference values ​​with a threshold value.

[0036] In one preferred embodiment of the present invention, the method further comprises using the configuration information to determine one or more possible or likely configurations of the textile machine from the location of the defect and one or more causes of the defect.

[0037] In one preferred embodiment of the present invention, the method further comprises using the configuration information to determine one or more impossible or unlikely configurations of the textile machine from the location of the defects and one or more causes of the defects.

[0038] In a preferred embodiment of the present invention, the method further comprises using one or more of time-dependent parameter information and time-dependent configuration information.

[0039] In one preferred embodiment of the present invention, the method further comprises applying one or more machine learning algorithms including one or more of a linear regression technique, a logistic regression technique, a support vector machine, a decision tree, a random forest technique, a K-nearest neighbor technique, a K-means technique, a naive Bayes classifier, and a principal component analysis technique, in particular taking into account one or more of insufficient parameter information, erroneous parameter information, insufficient configuration information, erroneous configuration information, insufficient knowledge base information, and erroneous knowledge base information.

[0040] In a preferred embodiment of the present invention, the method further comprises requesting and / or accessing supplemental configuration information, and applying the supplemental configuration information to one or more machine learning algorithms.

[0041] In one preferred embodiment of the present invention, the method further comprises a step of determining information if replacement, repair, modification and / or different adjustment of the textile machine and / or components of the textile machine related to one or more causes of the defect is required.

[0042] In one preferred embodiment of the invention, in particular if downtime of one or more textile machines of the spinning mill occurs as a result of maintenance work related to one or more causes of the defect, the method further comprises a step of determining information regarding maintenance work of one or more additional textile machines and / or one or more additional components of the textile machines.

[0043] In one preferred embodiment of the present invention, the method further includes receiving feedback information indicating whether a correction of the defect has occurred as a result of a maintenance action related to one or more causes of the defect, and updating the knowledge base information accordingly. [Brief explanation of the drawings]

[0044] The invention will be better understood with the aid of the description of one embodiment given by way of illustrative example. [Figure 1] 1 is a schematic diagram of an exemplary textile machine in a spinning mill for processing input fibrous material into output fibrous material; FIG. [Figure 2] 1 is a diagram illustrating an example configuration of a spinning mill and an apparatus for detecting defects in the spinning mill and for estimating one or more causes of the defects; [Figure 3] 1 is a diagram illustrating an example configuration of a spinning mill, a detected defect, two possible causes of the defect, and an apparatus for detecting defects in a spinning mill and estimating two possible causes of the defect. [Figure 4] 1 shows a schematic diagram of possible method steps of a method for detecting that a spinning mill has a defect and for estimating one or more causes of the defect;

[0045] Detailed Description of the Invention FIG. 1 is a schematic diagram of exemplary textile machines 12, 23, 34, 45, 56, 67, 78 of a spinning mill for processing input fibrous material 1, 2, 3, 4, 5, 6, 7 into output fibrous material 2, 3, 4, 5, 6, 7, 8. Different types of textile machines 12, 23, 34, 45, 56, 67, 78 and / or different arrangements of textile machines 12, 23, 34, 45, 56, 67, 78 may be included, depending on the spinning mill. In the example shown in FIG. 1, one or more blowing chambers 12 of the textile machine are arranged to process raw cotton 1 into a chute mat 2. One or more carding machines 23 of the textile machine are arranged to process the chute mat 2 into carded sliver 3. One or more breaker drawing frames 34 of the textile machine are arranged to process the carded sliver 3 into breaker drawn sliver 4. One or more finisher drawing frames 45 of the textile machine are arranged to process the breaker drawn sliver 4 into finisher drawn sliver 5. One or more roving frames 56 of the textile machine are arranged to process the finisher drawn sliver 5 into roving 6. One or more ring spinning frames 67 of the textile machine are arranged to process the roving 6 into ring pipe yarn 7. One or more winding frames 78 of the textile machine are arranged to process the ring pipe yarn 7 into yarn cones 8.

[0046] A spinning mill enables the production of a desired quantity and / or quality of a desired fiber material 8 from an original fiber material 1. The textile machines 12, 23, 34, 45, 56, 67, 78 shown in Figure 1 are configured to process input fiber material 1, 2, 3, 4, 5, 6, 7 into output fiber material 2, 3, 4, 5, 6, 7, 8, respectively, so as to obtain the desired quantity and / or quality. The quantity and / or quality of the output fiber material 2, 3, 4, 5, 6, 7, 8 depends on the parameters of the textile machines 12, 23, 34, 45, 56, 67, 78 and / or the parameters of the input fiber materials 1, 2, 3, 4, 5, 6, 7. The quantity and / or quality of the output fibrous materials 2, 3, 4, 5, 6, 7, 8 may depend not only on the direct input fibrous materials 1, 2, 3, 4, 5, 6, 7 of a particular textile machine 12, 23, 34, 45, 56, 67, 78, but also on the parameters of any upstream fibrous materials 2, 3, 4, 5, 6 produced by any upstream textile machine 12, 23, 34, 45, 56, as well as the parameters of the original fibrous material 1. For example, the direct input fibrous material 4 of the finisher drawing frame 45 is breaker drawing sliver 4, while the upstream fibrous materials include raw cotton 1, chute mat 2, and card sliver 3. Therefore, not only may the quantity and / or quality of the directly input fibrous materials 1, 2, 3, 4, 5, 6, 7 not match the desired quantity and / or quality, but also the quantity and / or quality of the output fibrous materials 2, 3, 4, 5, 6, 7, 8 may be reduced if the quantity and / or quality of the upstream input fibrous materials 1, 2, 3, 4, 5, 6 do not match the desired quantity and / or quality. For example, if the quantity and / or quality of one or more of the raw cotton 1, chute mat 2, and card sliver 3 does not match the desired quantity and / or quality, the quantity and / or quality of the finisher drawn sliver 5 produced by the drawing frame 45 of the textile machine may be reduced.

[0047] One exemplary configuration of a spinning mill M is shown schematically in Figure 2. Also shown in Figure 2 is a device E for detecting defects in the spinning mill and estimating one or more causes eS1, eS2 of the defects. The spinning mill M shown in Figure 2 includes different and / or additional textile machines than those shown in Figure 1. The present disclosure is not limited to the exemplary configuration of the spinning mill M shown in Figure 2, but also applies to any other configuration of a spinning mill.

[0048] The spinning mill M shown in Figure 2 has a bale opener 12 that processes raw material 1 into so-called microtufts 2. The bale opener 12 is followed by a pre-washer 23 that pre-washes the microtufts 2 into pre-washed fiber material 3. The pre-washer 23 is followed by a uniform mixer 34 that processes the pre-washed fiber material 3 into a mixed fiber material 4. The uniform mixer 34 is followed by a storage and feeder 45 that deep-cleans the mixed fiber material 4 into a highly cleaned fiber material 5. The storage and feeder 45 is followed by a condenser 56 that further cleans the highly cleaned fiber material 5 into further cleaned fiber material 6. The condenser 56 is followed by four drawing and sliver coilers 67.1, 67.2, 67.3, 67.4 of a textile machine that processes the further cleaned fiber material 6 into carded sliver coils 7. The four drawing and sliver coiler units 67.1, 67.2, 67.3, 67.4 are followed by eight roving frames 78.1, 78.2, 78.3, 78.4, 78.5, 78.6, 78.7, 78.8 of textile machinery which process the carded sliver coils 7 into roving yarn 8. The eight roving frames 78.1, 78.2, 78.3, 78.4, 78.5, 78.6, 78.7, 78.8 are followed by two ring spinning frames 89.1, 89.2 of textile machinery which process the roving yarn 8 into ring pipe yarn (spinning cops) 9.

[0049] As shown in FIG. 2, the conveying system is arranged to transport the fibrous material from an upstream textile machine to the next textile machine. The conveying system may include a pipe system D, a trolley (or cart) system T, a rail system R, etc. For example, in the pipe system D, the fibrous material is transported by an air flow generated by a ventilator. For example, in the trolley system T, the fibrous material is transported by a container placed on a trolley. For example, in the rail system R, the fibrous material is transported by a conveying device placed on a rail. Other conveying systems may also be included. In the example shown in FIG. 2, the pipe system D is arranged to transport the fibrous material through a series of textile machines, including a bale opener 12, a pre-washer 23, a uniform mixer 34, a storage and feeding machine 45, a condenser 56, and drawing and sliver coilers 67.1, 67.2, 67.3, and 67.4 of the textile machines. A trolley system T is arranged to transport the textile material from the drawing and sliver coiler units 67.1, 67.2, 67.3, 67.4 to the rovers 78.1, 78.2, 78.3, 78.4, 78.5, 78.6, 78.7, 78.8 of the textile machine. A rail system R is arranged to transport the textile material from the rovers 78.1, 78.2, 78.3, 78.4, 78.5, 78.6, 78.7, 78.8 of the textile machine to the ring spinning frames 89.1, 89.2 of the textile machine. The transport systems may be arranged in different ways.

[0050] As shown in Figure 2, parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 is obtained. The parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 may include parameter values ​​obtained from textile machines 12, 23, 34, 45, 56, 67, 78, and 89 and / or parameter values ​​obtained from textile materials 1, 2, 3, 4, 5, 6, 7, 8, and 9. Figure 2 schematically illustrates exemplary locations for obtaining the parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9. However, due to the complexity and / or costs, particularly in the typical configuration of a spinning mill M, one or more of the textile machines 12, 23, 34, 45, 56, 67, 78, 89 and / or textile materials 1, 2, 3, 4, 5, 6, 7, 8, 9 may only be partially acquired or not acquired at all.

[0051] The parameter values ​​of the parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 may relate to the rotation speed, power consumption, maintenance date, etc. of each textile machine 12, 23, 34, 45, 56, 67, 78, and 89. In addition, the parameter values ​​of the parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 may relate to the thickness, weight, etc. of each textile material 1, 2, 3, 4, 5, 6, 7, and 8.

[0052] To obtain the parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9, various electronic sensors are arranged to obtain the respective parameter values. The respective electronic sensors may relate to an electronic sensor for obtaining the rotational speed, an electronic sensor for obtaining the power consumption, etc. To obtain the parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9, laboratory analysis results may be required to obtain the respective parameter values. The laboratory analysis results may relate to the density of the fibers, the quality of the lubricant, etc.

[0053] The parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 may be acquired periodically or irregularly. The parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 may be acquired at short intervals at high speeds or at long intervals at low speeds. For example, the parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 acquired by electronic sensors may be acquired periodically at high speeds, such as every minute or every second. For example, the parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 requiring laboratory analysis results may be acquired at low speeds, such as every Monday or Thursday, after machine maintenance, and so on.

[0054] As shown in Figure 2, parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9 are acquired at specific locations. Thus, specific parameter information, such as the parameter information marked with reference sign p3, relates to a specific location in the spinning mill, in particular to a specific rotation speed of the rotor of a textile machine, a specific power consumption of a specific textile machine, a specific diameter of the textile material produced by a specific textile machine, etc.

[0055] FIG. 2 shows a schematic diagram of an electronic device E for detecting defects in a spinning mill and for estimating one or more causes eS1, eS2 of the defects according to the present disclosure.

[0056] The electronic device E may be in the form of a computer, such as a computer typically used in a single location (such as a traditional desktop computer, workstation, or server) and a computer typically portable (such as a laptop, notebook, tablet, or handheld computer). The electronic device E may include a machine-readable medium having stored thereon instructions that program a processor of the electronic device E to perform some or all of the operations and functions described in this disclosure. The machine-readable medium may include any mechanism for storing or transmitting information in a form readable by a machine (e.g., a computer), such as a hard disk drive (HD), a solid-state disk drive (SSD), a compact disc read-only memory (CD-ROM), a read-only memory (ROM), a random access memory (RAM), an erasable programmable read-only memory (EPROM), or the like. In another embodiment, some of these operations and functions may be performed by specific hardware components having hardwired logic. Alternatively, the operations and functions may be performed by any combination of programmable computer components and fixed hardware circuitry components. In one embodiment, the machine-readable medium has stored thereon a plurality of instructions that, when executed by a processor, cause the processor to perform a method in the electronic device E as described in this disclosure.

[0057] 2, the electronic device E is configured to receive parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 of the textile machines 12, 23, 34, 45, 56, 67, 78, and 89 and / or the textile materials 1, 2, 3, 4, 5, 6, 7, 8, and 9. For example, electronic sensors configured to acquire each of the parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 are disposed at respective positions within the spinning factory M, and electronic signals representing the acquired parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 are transmitted from the electronic sensors to the electronic device E via a computer network. Furthermore, the parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9 acquired in the results of laboratory analysis may be transmitted to the electronic device E via the computer network. That is, the electronic device E is configured to receive parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9 including different types of parameters acquired at different locations.

[0058] The electronic device E is configured to detect defects in the spinning mill by identifying parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9 of the textile materials 1, 2, 3, 4, 5, 6, 7, 8, 9 that differ from the reference information. For example, a difference between one or more parameter values ​​and one or more reference information is compared to a threshold, and a defect in the spinning mill M is detected if the difference exceeds the threshold. For example, a defect in the spinning mill M is detected if the diameter of the textile material exceeds a maximum diameter or if the diameter of the textile material is below a minimum diameter. Along with the defect detection, the location of the detected defect is also detected, because each of the parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9 is acquired at a specific location in the spinning mill M.

[0059] As shown in Fig. 2, the electronic device E is configured to access configuration information cI relating to the layout configuration of the spinning mill M. As shown in Fig. 2, the configuration information cI can be stored externally to the electronic device E, such as in a database on a computer server that stores information about the spinning mill M. Alternatively or additionally, the configuration information cI can be stored on a machine-readable medium of the electronic device E. The configuration information cI can include information about the type of textile machine, model information about the textile machine, the location of the textile machine, the arrangement of the textile machine, operating hours of the textile machine, etc.

[0060] The configuration information cI enables the determination or estimation of the arrangement of textile machines required to produce a particular textile material at a particular location within the spinning mill M. For example, the set of textile machines required to produce a particular boll 9 may include the ring spinning frame of the textile machine with reference number 89.1 that produced the particular boll 9 and the roving frame of the textile machine with reference number 78.3, because the configuration information cI stores information that the textile material produced by the roving frame of the textile machine with reference number 78.3 was transported to the ring spinning frame of the textile machine with reference number 89.1 during the relevant period. Furthermore, in this example, the configuration information cI indicates that the set of textile machines for producing the particular boll 9 includes the drawing and sliver coiler unit with reference number 67.4, because the configuration information cI stores information that the textile material produced by the drawing and sliver coiler unit with reference number 67.4 was transported to the roving frame of the textile machine with reference number 78.3 during the relevant period. Furthermore, in the configuration shown in FIG. 2, for example, since there are no alternative sets of textile machines in spinning mill M, the set of textile machines required to produce a particular boll 9 would include condenser 56, storage and feeding machine 45, uniform mixer 34, pre-washer 23, and bale opener 12. Thus, when a defect is detected in a particular bobbin 9 produced by the ring spinning frame with reference number 89.1, one or more causes eS1, eS2 of the defect will be limited to the ring spinning frame with reference number 89.1, the roving frame of the textile machine with reference number 78.3, the drawing and sliver coiler unit with reference number 67.4, the condenser 56, the storage and feeding machine 45, the uniform mixer 34, the pre-washer 23 and the bale opener 12, while the ring spinning frame with reference number 89.2, the roving frames of the textile machine with reference numbers 78.1, 78.2, 78.4, 78.5, 78.6, 78.7, 78.8 and the drawing and sliver coiler units with reference numbers 67.1, 67.2, 67.3 will not be considered as possible causes of the defect.

[0061] Furthermore, the configuration information cI may contain time-dependent information, for example based on a production plan which requires different textile machines at night and during the day, or during maintenance work and during normal production. Thus, for a first point in time it may be necessary to include or exclude a different set of textile machines as possible causes of a defect than for a second point in time. The configuration information cI may relate to current values ​​as well as to past values. The configuration information cI may also contain information about machine settings, such as mechanical settings, technical settings, software settings, etc.

[0062] The configuration information cI may include an indication of whether the stored information is incomplete or inaccurate. For example, as shown schematically in FIG. 2, a trolley system T may be arranged to transport textile material from drawing and sliver coilers 67.1, 67.2, 67.3, and 67.4 of a textile machine to roving frames 78.1, 78.2, 78.3, 78.4, 78.5, 78.6, 78.7, and 78.8 of the textile machine. The trolley system may require manual operation by an operator, who may be instructed to manually record the origin and destination. The operator may forget to record the origin and destination or may record an inaccurate origin or destination. Thus, the configuration information cI may include incomplete or inaccurate information.

[0063] As shown in FIG. 2, electronic device E is configured to access knowledge base information kI related to knowledge indicative of the causes of defects in spinning mill M. Knowledge about the causes of defects in spinning mill M may relate to longitudinal information gathered by experts in the spinning mill field. For example, knowledge may include that if boll 9 contains yarns with thicknesses that vary by more than a predetermined reference variation, the cause of the defect is more likely to be in, for example, the drawing and sliver coilers 67.1, 67.2, 67.3, 67.4 of the textile machine than in the bale opener 12 or the storage and feeding machine 45. Knowledge may include information gathered from computer algorithms and artificial intelligence for a particular spinning mill or multiple spinning mills. The multiple spinning mills may include spinning mills with essentially the same layout, spinning mills with similar layouts, spinning mills with different layouts, etc.

[0064] The electronic device E is configured to apply the parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9, the configuration information cI, and the knowledge base information kI to one or more machine learning algorithms to infer one or more causes of the defects. Thus, information including the parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9, the configuration information cI, and the knowledge base information kI is processed using one or more machine learning algorithms.

[0065] The machine learning algorithms may include one or more of linear regression techniques, logistic regression techniques, support vector machines, decision trees, random forest techniques, k-nearest neighbors (k-NN), k-means techniques, naive Bayes classifiers, and principal component analysis techniques.

[0066] Linear regression techniques can detect dependencies in the information. When a defect is detected, or in other words, when one of the parameter values ​​differs from the reference value, linear regression techniques can determine other parameter values ​​that are still not different from the reference value but that are correlated with the parameter value that differs from the reference value, thereby allowing one or more causes of the defect to be inferred.

[0067] Similar to linear regression techniques, logistic regression techniques can detect dependencies within information about binary values.

[0068] Support vector machines can classify complex information by defining hyperplanes within the space defined by the information. For example, it can be inferred that if the information currently being analyzed is on one side of the hyperplane, then a particular cause of the defect is more likely than if the information is on the other side of the hyperplane.

[0069] The decision tree may be applied to the information to deduce one or more causes of the defect. Based on the decision tree, the information is analyzed in a step-by-step manner to deduce one or more causes of the defect.

[0070] Random forest technology can be applied to information to estimate one or more causes of defects. Based on random forest technology, information is applied to a set of decision trees to estimate one or more causes of defects. Random forests are based on partitioning features or information into subsets of trees. The model considers only a small subset of features or information, rather than all of the features or information. Random forests can be run in parallel on various computational engines, such as various cores of a processor, various processors of a computer system, various computer systems, etc., thereby enabling estimation of one or more causes of defects within a given time frame, especially in cases of high complexity. Random forests can process various classes of features or information, such as binary features or information, categorical features or information, numerical features or information, etc. Random forests enable estimation of one or more causes of defects with little or no preprocessing of the features or information, and no rescaling or transformation of the features or information is required. Random forests are particularly well suited for high-dimensional data, features, or information because execution is limited to a subset of the data, features, or information.

[0071] K-nearest neighbors can be applied to the information to estimate one or more causes of the defect. Under K-nearest neighbors, the information is applied to classes defined based on the sample information and the sample's K nearest neighbors.

[0072] A K-means technique can be applied to the information to estimate one or more causes of the defects. Based on the K-means technique, the information is applied to clusters defined based on sample information and optimization of centroid locations.

[0073] A naive Bayes classifier may be applied to the information to estimate one or more causes of the defect, where the information is classified based on Bayes' theorem.

[0074] Principal component analysis techniques can be applied to the information to estimate one or more causes of defects. Based on principal component analysis techniques, a set of correlated variables is transformed into a set of uncorrelated variables to eliminate redundancy.

[0075] Machine learning algorithms may include common algorithms such as artificial neural networks (best for data patterns), convolutional neural networks (best for images), recurrent neural networks (best for audio signals), self-organizing maps (best for feature detection), deep Boltzmann machines (best for system modeling), autoencoders (best for property detection), etc.

[0076] Depending on the class or type of spinning mill, one or more machine learning algorithms may be applied to the information.

[0077] An artificial neural network (ANN) can be applied to the information to deduce one or more causes of the defect by performing regression and classification where patterns and sequences are recognized. Based on the artificial neural network, the information is analyzed in stages to deduce one or more causes of the defect.

[0078] Convolutional neural networks (CNNs) can be applied to still images, video streams, visual data, and other two-dimensional data to infer one or more causes of defects. Based on convolutional neural networks, a mathematical operation called convolution is performed to find specific features that infer one or more causes of defects.

[0079] Recurrent neural networks (RNNs) can be applied to time series data, sequence modeling, or audio signals (including noise patterns) to infer one or more causes of defects. Based on recurrent neural networks, long short-term memories (LSTMs) are used to process variable-length sequences of inputs and process the data in an internal state (memory) to infer one or more causes of defects.

[0080] Self-organizing maps (SOMs) can be applied to the information for data dimensionality reduction and visual representation. Based on the SOMs, competitive learning can be applied to approximate the data distribution to infer one or more causes of defects.

[0081] The K-means technique can be applied to the separately partitioned cluster data to estimate one or more causes of defects. The K-means technique is used to learn features, and the information is applied to the clusters defined based on the sample information and the optimization of the centroid positions.

[0082] Deep Boltzmann machines (DBMs) can be applied to model and monitor the behavior of a spinning mill or its subsystems, including climatic conditions, to infer one or more causes of defects. Based on deep Boltzmann machines, internal representations are learned to represent and solve difficult combinatorial problems to infer one or more causes of defects.

[0083] Autoencoders can be applied to information for dimensionality reduction (decoding) and information retrieval (encoding) to infer one or more causes of defects. Based on the autoencoder, various input representations are learned to assume useful characteristics to infer one or more causes of defects.

[0084] Artificial neural networks, convolutional neural networks, and recurrent neural networks belong to the class of supervised machine learning algorithms. Self-organizing maps, deep Boltzmann machines, and autoencoders belong to the class of unsupervised machine learning algorithms. Artificial neural networks can be used for regression and classification. Convolutional neural networks can be used for computer vision. Recurrent neural networks can be used for time series analysis. Self-organizing maps can be used for feature detection. Deep Boltzmann machines and autoencoders can be used for recommendation systems.

[0085] The cause of the defect may be related to a misadjusted textile machine 12, 23, 34, 45, 56, 67, 78, 89, worn components of the textile machine 12, 23, 34, 45, 56, 67, 78, 89, etc.

[0086] The estimation of the cause of the defect may include providing information as to whether the textile machine 12, 23, 34, 45, 56, 67, 78, 89 and / or components of the textile machine 12, 23, 34, 45, 56, 67, 78, 89 need to be replaced, repaired, modified, adjusted differently, etc.

[0087] FIG. 3 shows a schematic diagram of one exemplary configuration of a spinning mill, a detected defect, two possible causes of the defect, and an apparatus for detecting defects in the spinning mill and estimating the two possible causes of the defect.

[0088] The electronic device E receives parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9 relating to one or more parameters of the textile machines 12, 23, 34, 45, 56, 67, 78, 89 and one or more textile materials 1, 2, 3, 4, 5, 6, 7, 8, 9. For example, upon receiving a new parameter value, the parameter value is compared to a reference value. For example, the diameter of the textile material is compared to a reference diameter. If the parameter differs from the reference information, a defect is detected. Based on information relating to the location or origin of the parameter value, such as the location of a sensor, the location of the defect is also detected. Thus, the electronic device E detects defects dF and their locations in the spinning mill M by identifying parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9 relating to one or more parameters of the one or more textile materials 1, 2, 3, 4, 5, 6, 7, 8, 9 that differ from the reference information. In the exemplary configuration shown in FIG. 3, the defect dF concerns a bobbin 9 produced by a ring spinning frame of a textile machine with reference number 89.1.

[0089] The electronic device E accesses configuration information cI regarding the layout configuration of the spinning mill M and knowledge base information kI regarding knowledge about causes of defects within the spinning mill M, and applies the parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9, the configuration information cI, and the knowledge base information kI to one or more machine learning algorithms to deduce one or more causes of defects eS1, eS2. In the example shown in Figure 3, two causes of defects eS1, eS2 have been deduced, where a first cause of the defect, referenced eS1, is related to the drawing and sliver coiler of the textile machine referenced 67.1, and a second cause of the defect, referenced eS2, is related to the roving frame of the textile machine referenced 78.4. For example, when the random forest technique was applied to parameter information p1, p2, p3, p4, p5, p6, p7, p8, and p9, configuration information cI, and knowledge base information kI, the first cause eS1 and the second cause eS2 were estimated.

[0090] 4 shows schematic diagrams of possible method steps of a method for detecting that a spinning mill has a defect dF and estimating one or more causes eS1, eS2 of the defect. In step 1, an electronic device E receives parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9 of one or more textile machines (12, 23, 34, 45, 56, 67, 78, 89) and one or more textile materials 1, 2, 3, 4, 5, 6, 7, 8, 9. In step 2, the electronic device E detects the defect dF and its location by identifying parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9 of textile materials 1, 2, 3, 4, 5, 6, 7, 8, 9 that differ from the reference information. In step 3, the electronic device E accesses configuration information cI of textile machines 12, 23, 34, 45, 56, 67, 78, 89 of the spinning factory M. In step 4, the electronic device accesses knowledge base information kI regarding knowledge about possible causes eS' of possible defects dF' in the spinning factory M. In step 5, the electronic device applies the parameter information p1, p2, p3, p4, p5, p6, p7, p8, p9, the configuration information cI, and the knowledge base information kI to one or more machine learning algorithms to estimate one or more causes eS1, eS2 of the defect dF. [Explanation of symbols]

[0091] M Spinning Factory 1,2,3,4,5,6,7,8,9 Fiber materials 12, 23, 34, 45, 56, 67, 78, 89 Textile machinery p1,p2,p3,p4,p5,p6,p7,p8,p9 Parameters of textile machine and / or textile material D Pipeline System T Trolley System R-rail system E. Electronic device for detecting defects in spinning mills and for predicting the cause of the defects cI Spinning mill configuration information kI Spinning Mill Knowledge Base Information dF Detected defects Cause of eS1, eS2 defects

Claims

1. 1. An electronic device (E) for detecting defects (dF) in a spinning factory (M) having a plurality of textile machines (12, 23, ..., 89) for continuously processing textile material (1, 2, ..., 9) and for estimating one or more causes (eS1, eS2) of said defects (dF), said electronic device (E) comprising: receiving parameter information (p1, p2, ..., p9) of one or more of said textile machines (12, 23, ..., 89) and one or more of said textile materials (1, 2, ..., 9); detecting said defects (dF) and their locations by identifying said parameter information (p1, p2, ..., p9) of said fiber materials (1, 2, ..., 9) that differ from reference information; accessing configuration information (cI) of the textile machines (12, 23..., 89) of the spinning factory (M); accessing a knowledge base (kI) of information about possible causes (eS') of possible defects (dF') in said spinning mill (M); applying the parameter information (p1, p2, ..., p9), the configuration information (cI), and the knowledge base information (kI) to one or more machine learning algorithms to estimate one or more of the causes (eS1, eS2) of the defect (dF); An electronic device (E) configured to execute the above.

2. 2. The device according to claim 1, further configured to perform the step of identifying the parameter information (p1, p2, ..., p9) of the fiber material (1, 2, ..., 9) that differs from the reference information by comparing a difference between one or more parameter values ​​and one or more reference values ​​with a threshold value.

3. 3. The device according to claim 1 or 2, further configured to perform a step of using said configuration information (cI) to determine one or more possible or probable configurations of said textile machines (12, 23, ..., 89) from the location of said defects (dF) and one or more of said causes (eS1, eS2) of said defects.

4. 4. The device according to claim 1, further configured to perform a step of using the configuration information (cI) to determine one or more impossible or unlikely configurations of the textile machines (12, 23, ..., 89) from the location of the defect (dF) and one or more of the causes (eS1, eS2) of the defect.

5. 5. The device according to claim 1, further configured to perform the step of using one or more of time-dependent parameter information (p1, p2, ..., p9) and time-dependent configuration information (cI).

6. 6. The apparatus of claim 1, further configured to perform the step of applying one or more of the machine learning algorithms selected from a linear regression technique, a logistic regression technique, a support vector machine, a decision tree, a random forest technique, a k-nearest neighbor technique, a k-means technique, a naive Bayes classifier, and a principal component analysis technique, wherein one or more of the insufficient parameter information (p1, p2, ..., p9), the insufficient configuration information (cI), and the insufficient knowledge base information (kI) are applied to the machine learning algorithms.

7. 7. The apparatus of claim 1, further configured to perform the steps of requesting and / or accessing supplemental configuration information (scI) and applying the supplemental configuration information (scI) to one or more of the machine learning algorithms.

8. 8. The device according to claim 1, further configured to perform a step of determining information when replacement, repair, modification and / or different adjustment of the textile machine (12, 23, ..., 89) and / or components of the textile machine (12, 23, ..., 89) related to one or more of the causes (eS1, eS2) of the defect (dF) is required.

9. 9. The device according to claim 1, further configured to perform a step of determining information about maintenance work on one or more additional textile machines (12, 23, ..., 89) and / or one or more additional components of the textile machines (12, 23, ..., 89) if downtime of one or more of the textile machines (12, 23, ..., 89) of the spinning factory (M) occurs as a result of maintenance work related to one or more of the causes (eS1, eS2) of the defect (dF).

10. 10. The device according to claim 9, further configured to perform the steps of receiving feedback information indicating whether the maintenance action related to one or more of the causes (eS1, eS2) of the defect resulted in a correction of the defect (dF) and updating the knowledge base information (kI) accordingly.

11. 1. A method for detecting defects (dF) and estimating one or more causes (eS1, eS2) of the defects (dF) in a spinning factory (M) having a plurality of textile machines (12, 23, ..., 89) for continuously processing textile material (1, 2, ..., 9), the method comprising: receiving, by means of an electronic device (E), parameter information (p1, p2, ..., p9) of one or more of said textile machines (12, 23, ..., 89) and one or more of said textile materials (1, 2, ..., 9); detecting said defects (dF) and their locations by identifying, using said electronic device (E), said parameter information (p1, p2, ..., p9) of said textile material (1, 2, ..., 9) that differs from reference information; accessing, using said electronic device (E), configuration information (cI) of said textile machines (12, 23..., 89) of said spinning mill (M); accessing, using said electronic device (E), a knowledge base of information (kI) relating to knowledge about possible causes (eS') of possible defects (dF') in said spinning mill (M); and applying, using the electronic device (E), the parameter information (p1, p2, ..., p9), the configuration information (cI), and the knowledge base information (kI) to one or more machine learning algorithms to deduce one or more of the causes (eS1, eS2) of the defect (dF). A method comprising:

12. 12. The method of claim 11, further comprising the step of identifying the parameter information (p1, p2, ..., p9) of the fiber material (1, 2, ..., 9) that differs from the reference information by comparing a difference between one or more parameter values ​​and one or more reference values ​​with a threshold.

13. 13. The method according to claim 11 or 12, further comprising the step of using said configuration information (cI) to determine one or more possible or probable configurations of said textile machines (12, 23, ..., 89) from the location of said defects (dF) and one or more of said causes (eS1, eS2) of said defects.

14. 14. The method according to any one of claims 11 to 13, further comprising the step of using said configuration information (cI) to determine one or more impossible or unlikely configurations of said textile machines (12, 23, ..., 89) from the location of said defects (dF) and one or more of said causes (eS1, eS2) of said defects.

15. 15. The method according to any one of claims 11 to 14, further comprising the step of using one or more of time-dependent parameter information (p1, p2, ..., p9) and time-dependent configuration information (cI).

16. 16. The method of claim 11, further comprising applying one or more of the machine learning algorithms selected from a linear regression technique, a logistic regression technique, a support vector machine, a decision tree, a random forest technique, a k-nearest neighbor technique, a k-means technique, a naive Bayes classifier, and a principal component analysis technique, wherein one or more of the insufficient parameter information (p1, p2, ..., p9), the insufficient configuration information (cI), and the insufficient knowledge base information (kI) are applied to the machine learning algorithms.

17. 17. The method of any one of claims 11 to 16, further comprising the steps of requesting and / or accessing supplemental configuration information (scI) and applying the supplemental configuration information (scI) to one or more of the machine learning algorithms.

18. 18. The method according to any one of claims 11 to 17, further comprising a step of determining information when replacement, repair, modification and / or different adjustment of the textile machine (12, 23, ..., 89) and / or components of the textile machine (12, 23, ..., 89) related to one or more of the causes (eS1, eS2) of the defect (dF) is required.

19. 19. The method according to any one of claims 11 to 18, further comprising a step of determining information about maintenance work on one or more additional textile machines (12, 23, ..., 89) and / or one or more additional components of the textile machines (12, 23, ..., 89) if downtime of one or more of the textile machines (12, 23, ..., 89) of the spinning factory (M) occurs as a result of maintenance work related to one or more of the causes (eS1, eS2) of the defect (dF).

20. 20. The method of claim 19, further comprising the step of receiving feedback information indicating whether the maintenance action related to one or more of the causes (eS1, eS2) of the defect resulted in a correction of the defect (dF), and updating the knowledge base information (kI) accordingly.

Images