Claims:I/We claim:

1. A wear determination system (100) that determines, in a fiber processing unit (10) that forms a package (P) by winding a yarn (Y) around a winding tube (T), among components constituting a cradle (31) that rotatably supports the package (P), a worn-out state of a wear-prone component (37a, 41a) that wears out during a period in which the package (P) is rotated by a driving section (47), comprising:
an acquiring section (81A) that acquires as a state variable, vibrations generated in the cradle (31), and number of rotations or time from a time point at which a braking device (39) that brakes rotation of the package (P) starts operation to a time point at which the rotation of the package (P) stops; and
a determining section (81B) that determines the worn-out state of the wear-prone component (37a, 41a) based on the state variable acquired by the acquiring section (81A).

2. The wear determination system (100) as claimed in Claim 1, further comprising a storage section (83A) that stores therein an analytical model (M) generated by machine learning, wherein
the determining section (81B) inputs the state variable in the analytical model (M) and outputs the worn-out state of the wear-prone component (37a, 41a) as a determination result.

3. The wear determination system (100) as claimed in Claim 2, wherein the analytical model (M) is a pre-trained model obtained by learning by using teacher data in which the state variable and the worn-out state of the wear-prone component (37a, 41a) when the state variable is acquired are mapped with each other.

4. The wear determination system (100) as claimed in Claim 3, further comprising:
a receiving section (81E) that receives the teacher data; and
a learning section (81F) that learns the worn-out state of the wear-prone component (37a, 41a) according to the teacher data received by the receiving section (81E).

5. The wear determination system (100) as claimed in Claim 4, wherein
the receiving section (81E) is communicably connected to a collecting section (71A) that collects information on an operating state from each of a plurality of the fiber processing units (10), and
the learning section (81F) uses a part of the information on the operating state collected by the collecting section (71A) as the teacher data.

6. The wear determination system (100) as claimed in any one of Claims 1 to 5, wherein the state variable includes winding speed of the package (P) in the fiber processing unit (10).

7. The wear determination system (100) as claimed in any one of Claims 1 to 6, wherein the state variable includes information on amount of yarn wound around the package (P).

8. The wear determination system (100) as claimed in any one of Claims 1 to 7, wherein
the vibrations generated in the cradle (31) include vibrations in a first direction coinciding with an axial direction of the winding tube (T) and vibrations in a second direction orthogonal to the first direction, and
the state variable includes the vibrations in the first direction and the vibrations in the second direction.

9. The wear determination system (100) as claimed in any one of Claims 1 to 8, wherein the state variable includes information on a state of the winding tube (T).

10. The wear determination system (100) as claimed in any one of Claims 1 to 9, wherein
the cradle (31) includes a cradle arm (33, 34) for holding the winding tube (T),
the wear-prone component (37a) is a winding bush (37a) that constitutes an axis receiving member that rotatably supports the winding tube (T) held by the cradle arm (33, 34), and
the determining section (81B) determines a worn-out state of the winding bush (37a).

11. The wear determination system (100) as claimed in any one of Claims 1 to 9, wherein
the cradle (31) includes
a cradle arm (33, 34) that holds the winding tube (T); and
a lever (40) that rotates the cradle arm (33, 34) so that the cradle arm (33, 34) is in a closed state in which the cradle arm (33, 34) holds the winding tube (T) or in an open state in which the hold of the cradle arm (33, 34) on the winding tube (T) has been released,
the wear-prone component (41a) is a lever bush (41a) that constitutes a part of an axis receiving member that rotatably supports the lever (40), and
the determining section (81B) determines a worn-out state of the lever bush (41a).

12. The wear determination system (100) as claimed in any one of Claims 1 to 9, wherein
the cradle (31) includes
a cradle arm (33, 34) that holds the winding tube (T); and
a lever (40) that rotates the cradle arm (33, 34) so that the cradle arm (33, 34) is in a closed state in which the cradle arm (33, 34) holds the winding tube (T) or in an open state in which the hold of the cradle arm (33, 34) on the winding tube (T) has been released,
the wear-prone component (37a, 41a) includes a winding bush (37a) constituting an axis receiving member that supports rotation of the winding tube (T) held by the cradle arm (33, 34) and a lever bush (41a) that constitutes a part of an axis receiving member that supports the lever (40), and
the determining section (81B) determines a worn-out state of each of the winding bush (37a) and the lever bush (41a).

13. The wear determination system (100) as claimed in any one of Claims 1 to 12, further comprising a notifying section (71D) that notifies of a determination result obtained by the determining section (81B).

14. The wear determination system (100) as claimed in any one of Claims 1 to 13, further comprising a managing section (71E) that manages a maintenance schedule for each of the wear-prone components (37a, 41a), wherein
the managing section (71E) updates the schedule based on the determination result obtained by the determining section (81B).

15. A textile machine system (200) comprising:
the wear determination system (100) as claimed in any one of Claims 1 to 14;
a plurality of textile machines (1) each including a plurality of the fiber processing units (10) and a control device (2) that collectively controls the fiber processing units (10);
a managing device (70) that manages the textile machines (1); and
a server device (80) capable of communicating with the managing device (70), wherein
the acquiring section (81A) and the determining section (81B) are included in the server device (80),
the control device (2) collects the state variable from each of the fiber processing units (10) and sends the acquired state variables to the server device (80), and
the server device (80) sends the determination result obtained by the determining section (81B) to the managing device (70).

16. A textile machine system (200) comprising:
the wear determination system (100) as claimed in any one of Claims 1 to 14;
a plurality of textile machines (1) each including a plurality of the fiber processing units (10) and a control device (2) that collectively controls the fiber processing units (10); and
a managing device (70) that manages the textile machines (1), wherein
the acquiring section (81A) and the determining section (81B) are included in the managing device (70),
the control device (2) collects the state variable from each of the fiber processing units (10) and sends the acquired state variables to the managing device (70), and
the managing device (70) includes an output section (81D) that outputs the determination result obtained by the determining section (81B).

17. A textile machine system (200) comprising:
the wear determination system (100) as claimed in any one of Claims 1 to 14; and
a textile machine (1) including a plurality of the fiber processing units (10) and a control device (2) that collectively controls the fiber processing units (10), wherein
the acquiring section (81A) and the determining section (81B) are included in the control device (2),
the control device (2) collects the state variable from each of the fiber processing units (10), and
the control device (2) includes an output section (81D) that outputs the determination result obtained by the determining section (81B).
, Description:BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a wear determination system and a textile machine system.

2. Description of the Related Art
A textile machine including a fiber processing unit that winds a yarn around a winding tube to form a package is known in the art. Such a fiber processing unit of the textile machine includes a cradle that rotatably supports the package.
When a part of a component constituting the cradle wears out, a behavior different from the desired one occurs, and this may affect the quality of the package. In addition, since such a behavior often is not reflected in operation data, such as production efficiency, even if the wear amount of the component exceeds a reference wear amount indicative of the replacement timing of the component, this behavior cannot be noticed. Thus, there is a possibility that a failure will occur if the replacement of the component is delayed.
For example, Japanese Patent Application Laid-Open No. H6-127833 discloses a conventional wear determination system (cradle failure diagnosis device). This wear determination system includes an acceleration sensor that detects the acceleration of a cradle that supports a package, and a determining section that determines an abnormality of the cradle based on the acceleration detected by the acceleration sensor, or a vibration speed or an amplitude value calculated from the acceleration.
In the above-mentioned wear determination system, it is possible to obtain an objective inspection result, unlike the conventional visual inspection. In recent years, with the improvement in information processing technology, processing based on a lot of pieces of information has become possible, and a more accurate wear determination result is required.

SUMMARY OF THE INVENTION
In view of the above discussion, one object of the present invention is to provide a wear determination system and a textile machine system that can more accurately determine a worn-out state of a wear-prone component that wears out while a package is being rotated by a driving unit.
According to one aspect of the present invention, a wear determination system that determines, in a fiber processing unit that forms a package by winding a yarn around a winding tube, among components constituting a cradle that rotatably supports the package, a worn-out state of a wear-prone component that wears out during a period in which the package is rotated by a driving section, includes an acquiring section that acquires as a state variable, vibrations generated in the cradle, and number of rotations or time from a time point at which a braking device that brakes rotation of the package starts operation to a time point at which the rotation of the package stops; and a determining section that determines the worn-out state of the wear-prone component based on the state variable acquired by the acquiring section.
According to another aspect of the present invention, a textile machine system includes the above wear determination system; a plurality of textile machines each including a plurality of the fiber processing units and a control device that collectively controls the fiber processing units; a managing device that manages the textile machines; and a server device capable of communicating with the managing device. The acquiring section and the determining section are included in the server device, the control device collects the state variable from each of the fiber processing units and sends the acquired state variables to the server device, and the server device can send the determination result obtained by the determining section to the managing device.
According to still another aspect of the present invention, a textile machine system includes the above wear determination system; a plurality of textile machines each including a plurality of the fiber processing units and a control device that collectively controls the fiber processing units; and a managing device that manages the textile machines. The acquiring section and the determining section are included in the managing device, the control device collects the state variable from each of the fiber processing units and sends the acquired state variables to the managing device, and the managing device can include an output section that outputs the determination result obtained by the determining section.
According to still another aspect of the present invention, a textile machine system includes the above wear determination system; and a textile machine including a plurality of the fiber processing units and a control device that collectively controls the fiber processing units. The acquiring section and the determining section are included in the control device, the control device collects the state variable from each of the fiber processing units, and the control device can include an output section that outputs the determination result obtained by the determining section.
The above and other objects, features, advantages and technical and industrial significance of this invention will be better understood by reading the following detailed description of presently preferred embodiments of the invention, when considered in connection with the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic configuration diagram of an automatic winder including a wear-prone component whose worn-out state is determined by a wear determination system according to one embodiment.
FIG. 2 is a side view of a winder unit included in the automatic winder shown in FIG. 1.
FIG. 3 is a front view of a cradle included in the winder unit shown in FIG. 2.
FIG. 4 is a side view of the cradle included in the winder unit shown in FIG. 2.
FIG. 5 is a schematic configuration diagram of an automatic winder and a managing device provided in a spinning factory.
FIG. 6 is an overall system diagram of an example of the wear determination system according to one embodiment.
FIG. 7 is a functional block diagram showing a functional configuration of the managing device shown in FIG. 5.
FIG. 8 is a functional block diagram showing a functional configuration of a server device shown in FIG. 6.
FIG. 9 is a flowchart showing a process procedure until an analytical model M that is a pre-trained model is generated.
FIG. 10 is a flowchart showing a process procedure until the wear-prone component is determined by using the analytical model M that is the pre-trained model.
FIG. 11A is a diagram showing a relationship between a wear amount of a lever bush, a wear amount of a winding bush, and a magnitude of vibrations in a first direction.
FIG. 11B is a diagram showing a relationship between the wear amount of the lever bush, the wear amount of the winding bush, and a magnitude of vibrations in a second direction.
FIG. 11C is a diagram showing a relationship between the wear amount of the lever bush, the wear amount of the winding bush, and number of rotations before rotation of a package stops.

DETAILED DESCRIPTION
Exemplary embodiments of a wear determination system 100 will be described below with reference to the drawings. In the description of the drawings, the same or similar elements will be denoted by the same reference symbols and a redundant description thereof will be omitted.
The wear determination system 100 is a system that determines a worn-out state of a component of a winder unit (fiber processing unit) 10 shown in FIG. 2 that winds a yarn Y around a winding tube T to form a package P. More particularly, the wear determination system 100 determines a worn-out state of a wear-prone component that wears out during a period in which the package P is rotated by a driving section 47 among the components that constitute a cradle 31 that rotatably supports the package P.
In the present embodiment, the driving section 47 rotates a drum 45 that contacts the package P thereby causing the package P to rotate. Therefore, the period during which the package P is rotated by the driving section 47 refers to the period during which the drum 45 contacts the package P and the drum 45 is rotated by the driving section 47. For example, a braking device 39 described later operates in a state in which the drum 45 and the package P are separated from each other; therefore, the components constituting the braking device 39 are not subjected to the determination by the wear determination system 100. In the present embodiment, examples of the wear-prone component include a lever bush 41a (see FIG. 4) and a winding bush 37a (see FIG. 3).
To begin with, an automatic winder (textile machine) 1 including the winder unit 10 will be described. As shown in FIG. 1, the automatic winder 1 includes a machine control device 2, a plurality of the winder units 10, and a doffing device 50. The machine control device 2 includes a setting section 3, a display section 4, and a control section 5. The setting section 3 is a device operated by a worker to input predetermined setting values or select appropriate control methods for performing setting of each winder unit 10. The display section 4 is a device that displays the winding status of the yarn Y in each winder unit 10, the content of any problem that may have occurred, the worn-out state of the wear-prone components, schedule information related to a maintenance schedule, or the like. The setting section 3 and the display section 4 may be configured by a touchscreen.
The winder unit 10 forms the package P by unwinding the yarn Y from a yarn feeding bobbin B and then winding the yarn Y around the winding tube T (see FIG. 2) while traversing the yarn Y. The winding tube T and the package P come in various sizes and shapes. For example, the winding tube T and the package P can have a truncated cone shape (cone shape) or a cylindrical shape. In the present embodiment, the winding tube T and the package P are positioned exactly above the yarn feeding bobbin B in the machine height direction, and the yarn Y is caused to run from a lower side to an upper side. However, the present invention can be applied to a winder unit in which the yarn Y runs from the upper side to the lower side. When the winder unit has such a configuration, in the present specification, "lower/down/downward" should be read as "upper/up/upward" or "upstream (with respect to a running direction of the yarn)", and "upper/up/upward" should be read as "lower/down/downward" or "downstream (with respect to the running direction of the yarn)".
When the package P has become full in a particular winder unit 10, for example, the doffing device 50 moves to the position of this winder unit 10, doffs the full package P from the winder unit 10, and sets an empty winding tube T in the winder unit 10. The package P doffed by the doffing device 50 is discharged to a placing section (not shown) arranged on a back side (a side of the automatic winder 1 on which a passage for the workers is provided is referred to as a front side, and a side on the opposite side of the front side is referred to as the back side) of the winder unit 10, and thereafter the package P is collected by various means. The doffing device 50 can discharge not only the full package P but also a non-full package P or an empty winding tube T as appropriate. Note that, it is allowable that the automatic winder 1 is devoid of the doffing device 50. In this case, it is preferable that the worker performs the doffing work manually.
As shown in FIG. 2, the winder unit 10 includes a unit controller 11 and a unit main body 12. The unit controller 11 is accommodated in a unit frame 10a. The unit controller 11 is an electronic control unit including a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), an I/O port, a communication port, and the like. A computer program that controls each component of the unit main body 12 is stored beforehand in the ROM. The components of the unit main body 12 and the machine control device 2 are connected to the I/O port and the communication port. Accordingly, the unit controller 11 can control the operation of each component of the unit main body 12 while communicating with the machine control device 2.
In the unit main body 12, along the running path of the yarn Y from the yarn feeding bobbin B to the winding tube T and the package P, the following components are arranged in this order: a yarn unwinding assisting device 13, a tension applying device 14, a yarn joining device 15, a yarn monitoring device 16, and a winding device 30. The yarn unwinding assisting device 13, the tension applying device 14, the yarn joining device 15, the yarn monitoring device 16, and the winding device 30 are supported by the unit frame 10a. A yarn supplying unit 17 is arranged below the unit main body 12. The yarn supplying unit 17 holds the yarn feeding bobbin B at a predetermined position. The yarn feeding bobbin B can be conveyed to the yarn supplying unit 17 by a not-shown bobbin conveying system, or can be manually set on the yarn supplying unit 17 by the worker.
The yarn unwinding assisting device 13 assists in unwinding the yarn Y by controlling a size of a balloon of the yarn that is formed above the yarn feeding bobbin B when the yarn Y is unwound from the yarn feeding bobbin B. The tension applying device 14 applies a predetermined tension to the running yarn Y. As the tension applying device 14, for example, a gate type, a disk type, or the like can be used. The yarn joining device 15 joins a lower yarn on the yarn feeding bobbin B side and an upper yarn on the package P side when the yarn Y is cut by the yarn monitoring device 16 or when the yarn Y from the yarn feeding bobbin B is disconnected. The yarn monitoring device 16 detects a yarn defect such as a slub by detecting the thickness of the yarn Y. The yarn monitoring device 16 includes a cutter that cuts the yarn Y when the yarn defect is detected, for example. Note that, the yarn monitoring device 16 may detect a foreign matter contained in the yarn Y as the yarn defect.
A lower yarn capturing and guiding device 18 that captures the lower yarn on the yarn feeding bobbin B side and guides the same to the yarn joining device 15 is arranged below the yarn joining device 15. The lower yarn capturing and guiding device 18 includes a pipe arm 18b that rotates vertically about an axis 18a. A lower yarn suction port 18c is provided at a distal tip of the pipe arm 18b. A not-shown negative pressure source is connected to the pipe arm 18b. A suction flow for capturing the lower yarn on the yarn feeding bobbin B side is generated in the lower yarn suction port 18c.
An upper yarn capturing and guiding device 19 that captures the upper yarn on the package P side and guides the same to the yarn joining device 15 is arranged above the yarn joining device 15. The upper yarn capturing and guiding device 19 includes a pipe arm 19b that rotates vertically around an axis 19a. An upper yarn suction port 19c is provided at a distal tip of the pipe arm 19b. A not-shown negative pressure source is connected to the pipe arm 19b. A suction flow for capturing the upper yarn on the package P side is generated in the upper yarn suction port 19c.
The winding device 30 winds the yarn Y on the package P while traversing the yarn Y. The winding device 30 includes the cradle 31 and the drum 45.
The cradle 31 holds the package P by pinching the winding tube T of the package P. The cradle 31 is swingable between a state in which the held package P contacts the drum 45 and a state in which the held package P is separated from the drum 45.
The drum 45 transmits the driving force of the driving section 47 to the package P that is in contact with the drum 45 to rotate the package P and traverses the yarn Y on the surface of the package P. A spiral traverse groove 45a is formed on an outer peripheral surface of the drum 45. The drum 45 is rotationally driven by the driving section 47. The driving section 47 is provided on one side of the drum 45, and is accommodated in the unit frame 10a. Examples of the driving section 47 include, for example, a servo motor, a step motor, or the like. The driving section 47 is controlled by the unit controller 11.
The yarn Y unwound from the yarn feeding bobbin B is wound on the surface of the package P while being traversed for a certain width by the traverse groove 45a formed on the outer peripheral surface of the drum 45. Accordingly, the package P having a certain winding width is formed. When the cradle 31 holds the winding tube T around which the yarn Y has been wound, the drum 45 contacts the outer peripheral surface of the package P. On the other hand, when the cradle 31 holds an empty winding tube T around which no yarn Y has been wound, the drum 45 contacts the outer peripheral surface of the winding tube T itself. In the following description, the package P and the winding tube T are also collectively referred to as the package P.
As shown in FIGS. 3 and 4, the cradle 31 includes a first cradle arm 33, a second cradle arm 34, and a connecting beam 35 that connects a base end 33a of the first cradle arm 33 and a base end 34a of the second cradle arm 34. The cradle 31 is swingably supported by the unit frame 10a. The cradle 31 is urged so that the held package P contacts the drum 45 at a predetermined pressure.
The first cradle arm 33 is attached to the connecting beam 35 so that a distal end 33b thereof is swingable inward (toward a distal end 34b of the second cradle arm 34) and outward. The first cradle arm 33 is biased inward so that it can pinch and hold the package P. The first cradle arm 33 is provided with a cradle lever (lever) 40 for controlling attachment/detachment and the like of the package P.
The cradle lever 40 is coupled to the first cradle arm 33 via an axis receiving member 41. The axis receiving member 41 includes the lever bush 41a. The lever bush 41a is a ring-shaped member formed of metal or the like, for example. As described above, the lever bush 41a is one of the components that wears out during the period when the drum 45 contacts the package P and the package P is rotated by the driving force of the driving section 47.
A first bearing unit (axis receiving member) 37 is mounted inside the distal end 33b of the first cradle arm 33. A second bearing unit (axis receiving member) 38 is mounted inside the distal end 34b of the second cradle arm 34. The package P is rotatably held by the cradle 31 by fitting respective ends of the winding tube T in the first bearing unit 37 and the second bearing unit 38. In the present embodiment, the first bearing unit 37 is a bearing unit that rotatably supports a large-diameter end (first end) of the cone-shaped package P. The second bearing unit 38 is a bearing unit that rotatably supports a small-diameter end (second end) of the cone-shaped package P.
The first bearing unit 37 includes the winding bush 37a. The winding bush 37a is a ring-shaped member formed of resin or the like, for example. As described above, the winding bush 37a is one of the components that wares out during the period in which the drum 45 contacts the package P and the package P is rotated by the driving force of the driving section 47. In the present embodiment, the winding bush 37a and the lever bush 41a are made from different material. Therefore, the winding bush 37a and the lever bush 41a wear out differently due to the difference in the installation location and the material thereof.
The cradle 31 includes an angle sensor 36a. The angle sensor 36a includes a rotary encoder, a resolver, or the like. The angle sensor 36a detects a swing angle of the cradle 31 with respect to the frame. As the winding of the yarn Y around the package P proceeds whereby the diameter of the package P increases, the cradle 31 swings with respect to the frame. Accordingly, the diameter (thickness) of the package P can be calculated based on the swing angle detected by the angle sensor 36a. The swing angle detected by the angle sensor 36a is acquired by the machine control device 2.
The winder unit 10 includes an acceleration sensor 36b. The acceleration sensor 36b detects vibrations generated in the cradle 31. The acceleration sensor 36b according to the present embodiment detects vibrations in a first direction coinciding with an axial direction of the winding tube T and vibrations in a second direction that is orthogonal to the first direction. The vibrations detected by the acceleration sensor 36b are acquired by the machine control device 2. In the present embodiment, the first direction is the axial direction of the winding tube T, and coincides with the width direction of the machine frame (the width direction of the winder unit 10). The second direction coincides with the front-back direction of the machine frame (the front-back direction of the winder unit 10).
The acceleration sensor 36b is specifically provided on a control board installed on the unit frame 10a of the winder unit 10. This control board functions as the unit controller 11. The acceleration sensor 36b is not limited to the one that is provided on the control board constituting the unit controller 11, and may be provided at an appropriate location of the cradle 31, for example.
The braking device 39 brakes the rotation of the package P. The braking device 39 includes a brake shoe that can be pressed against the first bearing unit 37, a variable air pressure unit that controls the operation of the brake shoe, and the like. The braking device 39 is controlled by the unit controller 11. The braking device 39 operates in a state in which the drum 45 and the package P are separated from each other (lift-up). In the present embodiment, it is possible to acquire the number of rotations of the package P from a time point at which the braking device 39 starts operation to a time point at which the rotation of the package P stops.
The operation timing of the braking device 39 can be determined, for example, by inputting and outputting a command signal for operating the braking device 39. The first bearing unit 37 includes a sensor that detects the rotation of the package P. It is possible to determine that the rotation of the package P has stopped based on a change in the value detected by the sensor.
As shown in FIG. 5, for example, one spinning factory is provided with a plurality of the automatic winders 1 and provided with one managing device 70 for managing the automatic winders 1. The managing device 70 can communicate with the machine control device 2 provided in each of the automatic winders 1. The managing device 70 collects various operation data from the machine control device 2 and monitors the status of each automatic winder 1. Examples of the operation data include information on the vibrations of the cradle 31, the number of rotations of the package P from the time point at which the braking device 39 starts operation to the time point at which the package P stops, and information on the diameter of the package P acquired from the angle sensor 36a. Also, examples of the operation data include number of yarn defect removals, operation efficiency, and rate of yarn joining failures. Further, examples of the operation data include a winding speed of the package P.
The winding speed is a winding condition set by a worker by using the setting section 3. Therefore, the winding speed can be said to be the setting data. In the specification and the claims of the present application, the information including the operation data and the setting data is also referred to as "information on the operating state".
The managing device 70 is configured as a computer system including a CPU, a main storage unit such as a RAM and a ROM, an auxiliary storage unit exemplified by a hard disk, a flash memory, or the like. The present embodiment has a storage section 71F as shown in FIG. 7 as the auxiliary storage unit. The functions of an operation data input section (collecting section) 71A, an operation data output section 71B, a determination result input section 71C, a notifying section 71D, and a managing section 71E, which will be described in detail later, are performed under the control of the CPU by reading a predetermined computer software on the hardware such as the CPU and the main storage unit.
The storage section 71F stores therein a schedule related to maintenance of each automatic winder 1. In the present embodiment, the storage section 71F stores therein the next scheduled maintenance date for each winder unit 10 for each of the wear-prone components (the lever bush 41a and the winding bush 37a) constituting the automatic winder 1. In addition, the storage section 71F stores therein replacement time information in which a replacement time, which is a guide for the replacement time of the wear-prone component, is mapped with each of the worn-out states of the wear-prone component (the lever bush 41a and the winding bush 37a). The replacement time information is, for example, information indicating that the replacement time comes after one year if the worn-out state of the lever bush 41a is "wear amount: small". The replacement time information is, for example, information indicating that the replacement time comes after three months if the worn-out state of the lever bush 41a is "wear amount: medium". The replacement time information is, for example, information indicating that the replacement time comes within one week if the worn-out state of the lever bush 41a is "wear amount: large".
The operation data input section 71A acquires operation data from the machine control devices 2 of the automatic winders 1. The operation data input section 71A can temporarily store the acquired operation data in a not-shown storage section.
The operation data output section 71B transmits the operation data acquired by the operation data input section 71A to a server device 80. The operation data output section 71B can transmit the operation data to the server device 80 each time the operation data is acquired by the operation data input section 71A. Alternatively, the operation data output section 71B can transmit the operation data stored in the not-shown storage section to the server device 80 at a predetermined timing.
The determined status of the wear-prone component determined by an analysis section (determining section) 81B of the server device 80 described in detail below, that is, the worn-out state of the lever bush 41a and the winding bush 37a, is input into the determination result input section 71C. In the present embodiment, the determination result input section 71C acquires information transmitted from the server device 80.
The managing section 71E determines whether the replacement time has come for the lever bush 41a and the winding bush 37a based on the worn-out state of the lever bush 41a and the winding bush 37a. For example, the managing section 71E can determine whether the replacement time has come based on the replacement time information stored in advance in the storage section 71F or the like.
The managing section 71E manages the scheduled maintenance date of each wear-prone component (the lever bush 41a and the winding bush 37a) constituting the automatic winder 1 for each winder unit 10. The managing section 71E manages the scheduled maintenance date input from a not-shown input section of the managing device 70, and based on the worn-out state of the lever bush 41a and the winding bush 37a input from the determination result input section 71C modifies (updates) the scheduled maintenance date. For example, if the condition of the wear-prone component is not bad (the wear amount is small), the managing section 71E postpones the scheduled maintenance date, and if the condition of the wear-prone component is bad (the wear amount is large), the managing section 71E prepones the scheduled maintenance date.
The notifying section 71D reports the determined status of the wear-prone component obtained by the determination result input section 71C to the worker of the automatic winder 1 ("worker" may include "manager" and "maintenance staff"). More specifically, the determination result input section 71C displays the worn-out state of the lever bush 41a and the winding bush 37a on at least one of a not-shown display unit that can communicate with the managing device 70, the display section 4 of the machine control device 2 (see FIG. 1), and a display section of a portable terminal 73 such as a tablet that can communicate with the managing device 70 or the machine control device 2 (see FIG. 6).
Also, the notifying section 71D displays the scheduled maintenance date of the lever bush 41a and the winding bush 37a managed by the managing section 71E on at least one of the not-shown display unit that can communicate with the managing device 70, the display section 4 of the machine control device 2 (see FIG. 1), and the display section of the portable terminal 73 such as the tablet that can communicate with the managing device 70 or the machine control device 2 (see FIG. 6).
That is, the notifying section 71D displays, for example, as the maintenance schedule, information that indicates whether the replacement time determined by the managing section 71E for the lever bush 41a and the winding bush 37a has come, on at least one of the not-shown display unit that can communicate with the managing device 70, the display section 4 that can communicate with the machine control device 2 (see FIG. 1), and the display section of the portable terminal 73 such as the tablet that can communicate with the managing device 70 or the machine control device 2 (see FIG. 6).
The notifying section 71D may cause the display section to display a message prompting to place an order for a wear-prone component based on the worn-out state. Wear of the lever bush 41a and the winding bush 37a often progresses over a span of several months. Also, each spinning factory does not always have the stock of those bushes. Accordingly, when the worker sees the message prompting the order displayed on the display section and places an order of the component when the wear amount is "medium", the worker can obtain the component before the wear amount becomes "large". Therefore, components can be replaced without delay when the wear amount changes to "large". For this reason, it is possible to prevent production efficiency from being lowered due to waiting for the components to be obtained (waiting for replacement). That is, it is preferable that the notifying section 71D performs, as a first stage, a notification that prompts to place an order of the components (notification of predictive maintenance), and performs, as a second stage, a notification that prompts to replace the components (abnormality notification).
The notification as the second stage can be realized, for example, by updating the schedule when the maintenance has a regular scheduled schedule (maintenance schedule). In the schedule, a scheduled maintenance date is set for each winder unit 10, or for each maintenance target item (for example, the wear-prone component), or for each maintenance content. The maintenance content includes, for example, "replacement of the lever bush 41a", "replacement of the winding bush 37a", "replacement of the brake shoe", "lubrication", "inspection", "cleaning", or the like. It is preferable that the maintenance schedule is displayed, for example, in a calendar format, and that each piece of the maintenance information is indicated on each scheduled maintenance date.
Next, the wear determination system 100 will be described. As shown in FIG. 6, the wear determination system 100 includes the server device 80 that can communicate with the managing device 70 provided in each of spinning factories U1, U2, U3. That is, the wear determination system 100 is constructed as a part of a textile machine system 200 that includes the automatic winders 1 and the managing device 70, which are respectively provided in the respective spinning factories U1, U2, and U3, and the server device 80. The server device 80 is configured as a computer system including a CPU, a main storage unit such as a RAM and a ROM, an auxiliary storage unit exemplified by a hard disk, a flash memory, or the like. The present embodiment includes, as the auxiliary storage unit, an analytical model storage section 83A and a teacher data storage section 83B as shown in FIG. 8.
The functions of an acquiring section 81A, the analysis section 81B, an output section 81D, a teacher data input section (receiving section) 81E, and an analytical model learning section (learning section) 81F, which will be described in detail later, are performed under the control of the CPU by reading a predetermined computer software on the hardware such as the CPU and the main storage unit. Note that hardware for performing matrix calculation and product-sum calculation at high speed for performing arithmetic processing at high speed may be used. The processing in this case is also executed under the control of the CPU.
The analytical model storage section 83A stores therein an analytical model M generated by machine learning. The analytical model M is the pre-trained model trained by using teacher data (later-explained winding bush learning data, winding bush verification data, lever bush learning data, and lever bush verification data) in which are mapped the later-explained state variable (may also be called explanatory variable) and the worn-out state of the lever bush 41a and the winding bush 37a at the time when the state variable is acquired. That is, the analytical model storage section 83A stores therein two analytical models M: an analytical model M1 for the lever bush and an analytical model M2 for the winding bush. When the later-explained state variable is input, the analytical model M1 for the lever bush outputs information indicating the worn-out state of the lever bush 41a, and the analytical model M2 for the winding bush outputs information indicating the worn-out state of the winding bush 37a.
The state variable of the wear determination system 100 according to the present embodiment includes the vibrations generated in the cradle 31, and the number of rotations from the time point at which the braking device 39 that brakes the rotation of the package P starts operation until the time point at which the rotation of the package P stops (hereinafter, "number of package rotations before stopping"). The vibrations generated in the cradle 31 include vibrations in the first direction coinciding with the axial direction of the winding tube T and vibrations in the second direction that is orthogonal to the first direction.
The teacher data storage section 83B stores therein the data in which is mapped, with the state variable, which includes the vibrations generated in the cradle 31 and the number of package rotations before stopping, the worn-out state of the lever bush 41a and the wound-out state of the winding bush 37a, respectively. The teacher data storage section 83B stores therein the data processed by the teacher data input section 81E. The details of the data processed by the teacher data input section 81E will be described later. Further, the teacher data stored in the teacher data storage section 83B is read out by the analytical model learning section 81F and used to generate (update) the pre-trained model.
The acquiring section 81A acquires the state variable that includes the vibrations generated in the cradle 31 and the number of package rotations before stopping. Specifically, the state variables are acquired as a part of the operation data by the machine control device 2 provided in each automatic winder 1. The operation data obtained by the machine control devices 2 is collected in the managing device 70. The acquiring section 81A of the server device 80 acquires the operation data from the managing device 70 deployed in each place, and from the operation data extracts the vibrations generated in the cradle 31 (vibrations in the first direction and vibrations in the second direction) and the number of package rotations before stopping.
In the present embodiment, the operation data used for machine learning is extracted from the various types of the operation data acquired by the server device 80 from the managing device 70; however, the entity that extracts the operation data used for machine learning is not limited to the server device 80. For example, it is allowable that the managing device 70 extracts (selects) the operation data to be used for the machine learning from various types of the operation data and transmits the extracted operation data to the server device 80.
The analysis section 81B inputs the state variable acquired by the acquiring section 81A in the analytical model M stored in the analytical model storage section 83A, and acquires the worn-out state of the wear-prone component output from the analytical model M (that is, outputs the worn-out state as a determination result). More specifically, the analysis section 81B inputs the state variable acquired by the acquiring section 81A in each of the analytical model M1 for the lever bush and the analytical model M2 for the winding bush whereby the information indicating the worn-out state of the lever bush 41a and the information indicating the worn-out state of the winding bush 37a is output.
It is preferable that the analysis section 81B displays the worn-out state stepwise, for example, by a magnitude or degree of wear. Specifically, it is preferable that the analysis section 81B categorizes the worn-out state into three categories: "wear amount: large", "wear amount: medium", and "wear amount: small". The category "wear amount: small" also includes a state in which the wear amount is zero. However, the manner of outputting the worn-out state by the analysis section 81B is not limited to this. For example, the analysis section 81B can output a concrete value of the wear amount.
The output section 81D sends the worn-out state of the lever bush 41a and the worn-out state of the winding bush 37a obtained by the analysis section 81B to the managing device 70 that manages the sender of the operation data.
The teacher data input section 81E acquires the teacher data used by the analytical model learning section 81F for learning. The teacher data input section 81E acquires data (hereinafter, "operation data" and "correct label data") relating to the worn-out state of the lever bush 41a and the worn-out state of the winding bush 37a when operating under each operating condition.
The worn-out state can be grasped by the worker by manually measuring the wear amount. To do this, the worker disassembles the cradle 31 of the automatic winder 1 operated under the operation status indicated by each operation data, measures the wear amount after removing the lever bush 41a and the winding bush 37a, and inputs the measurement result in the teacher data input section 81E as the worn-out state. The worn-out state is stored by linking with, for example, a measurement date, a factory code of the automatic winder 1, a machine number (serial number) of the automatic winder 1, a unit number, or the like. Because the operation data is also recorded by being mapped with the operation date and time, the factory code of the automatic winder 1, the machine number of the automatic winder 1, the unit number, or the like, the teacher data storage section 83B can store therein the operation data by linking with the worn-out state.
The analytical model learning section 81F generates the analytical model M that is the pre-trained model by performing the machine learning by using the teacher data stored in the teacher data storage section 83B. The analytical model M obtained by such machine learning is the pre-trained model configured to output a predicted result of the worn-out state of the lever bush 41a and the worn-out state of the winding bush 37a when the state variable that includes the vibrations generated in the cradle 31 (vibrations in the first direction and vibrations in the second direction) and the number of rotations until the rotation of the package P stops is input. The analytical model M generated (updated) by the analytical model learning section 81F is stored in the analytical model storage section 83A.
Next, a process procedure until the analytical model M that is the pre-trained model is generated by the wear determination system 100 will be described mainly with reference to FIGS. 8 and 9.
The teacher data input section 81E of the wear determination system 100 acquires the teacher data used for learning by the analytical model learning section 81F and stores the acquired data in the teacher data storage section 83B. In the present embodiment, the teacher data input section 81E acquires the operation data transmitted from the managing devices 70 deployed at various places (Step S1). In addition, the teacher data input section 81E acquires data on the worn-out state of the lever bush 41a and the worn-out state of the winding bush 37a when operating under the operating condition (hereinafter, "correct label data") (Step S2).
Next, the teacher data input section 81E maps the operation data acquired at Step S1 with the correct label data acquired at step S2. In the present embodiment, the teacher data input section 81E acquires the teacher data generated by merging the operation data and the correct label data (Step S3) as the winding bush learning data and the winding bush verification data (Step S4). The winding bush learning data and the winding bush verification data have the same format. More specifically, the teacher data (a plurality of records) generated by merging the operation data and the correct label data is divided into the winding bush learning data and the winding bush verification data. For example, if there are 10,000 pieces of the teacher data, then, 8000 pieces are used as the winding bush learning data and 2,000 pieces are used as the winding bush verification data.
Similarly, the teacher data input section 81E merges the operation data and the correct label data to generate the lever bush learning data and the lever bush verification data (Step S7).
In the present embodiment, the teacher data input section 81E stores, in the teacher data storage section 83B, as the teacher data, the winding bush learning data, the winding bush verification data, the lever bush learning data, and the lever bush verification data generated by the teacher data input section 81E as described above.
Next, the analytical model learning section 81F executes the machine learning by using the teacher data accumulated in the teacher data storage section 83B (Steps S5 and S8) thereby acquiring the respective analytical models M that is the pre-trained model (Steps S6 and S9). That is, the analytical model learning section 81F acquires the two analytical models M: the analytical model M1 for the lever bush that outputs the worn-out state of the lever bush 41a, and the analytical model M2 for the winding bush that outputs the worn-out state of the winding bush 37a.
The method of machine learning executed by the analytical model learning section 81F is not limited to any specific method. That is, as the method of machine learning, various methods such as the neural network, the k-nearest neighbor method, and the support vector machine (SVM) can be used. When the analytical model M is configured by the neural network, for example, the pre-trained model in which the parameters of the intermediate layer of the neural network are tuned by the teacher data is obtained as the analytical model M.
Once the analytical model M1 for the lever bush and the analytical model M2 for the winding bush are generated, the analytical model learning section 81F verifies the performance of the generated analytical model M1 for the lever bush and the generated analytical model for the winding bush. Specifically, the analytical model M1 for the lever bush is verified by using the lever bush verification data that has the same format as the lever bush learning data generated in the teacher data input section 81E and that is unknown data not included in the lever bush learning data.
For example, after the analytical model M1 for the lever bush is trained by 8000 pieces of the lever bush learning data out of 10,000 pieces of the teacher data, the analytical model learning section 81F inputs the operation data (state variable) of 2000 pieces of the lever bush verification data in the analytical model M1 for the lever bush whereby the worn-out state of the lever bush is output as the determination result. Then, the analytical model learning section 81F compares the determination result with the correct answer data (worn-out state) of 2000 pieces of the lever bush verification data one by one to calculate the correct answer rate. As a result of such verification, the analytical model learning section 81F determines that the verification result is "acceptable" if the correct answer has a preset acceptable level, and otherwise determines "unacceptable". If the analytical model learning section 81F determines "unacceptable", it employs a method of tuning the algorithm, increasing the learning data, or the like. The analytical model learning section 81F repeatedly executes such learning and verification until the correct answer rate falls within the pre-set acceptable range.
Next, in reference to the wear determination system 100 described above, the process procedure until the determination of the wear-prone component by using the analytical model M that is the pre-trained model will be explained mainly while referring to FIGS. 8 and 10.
The acquiring section 81A acquires the state variable that includes the vibrations generated in the cradle 31 and the number of package rotations before the rotation stops. In the present embodiment, the acquiring section 81A acquires the operation data from the managing devices 70 deployed at various places (Step S11). Next, the acquiring section 81A extracts from the operation data the vibrations generated in the cradle 31 (vibrations in the first direction and vibrations in the second direction) and the number of rotations of the package P until the package P stops (Step S12). Then, the acquiring section 81A acquires the state variable for the winding bush (Step S13) and the state variable for the lever bush (Step S16).
The analysis section 81B inputs the state variable acquired by the acquiring section 81A in the analytical model M2 for the winding bush stored in the analytical model storage section 83A, and acquires the worn-out state of the winding bush 37a output from the analytical model M2 for the winding bush to perform analysis (Step S14). Further, the analysis section 81B inputs the state variable acquired by the acquiring section 81A in the analytical model M1 for the lever bush stored in the analytical model storage section 83A, and acquires the worn-out state of the lever bush 41a output from the analytical model M1 for the lever bush to perform analysis (Step S17).
Next, the output section 81D outputs the result of the analysis obtained by the analysis section 81B, that is, the worn-out state of the winding bush 37a and the worn-out state of the lever bush 41a acquired by the analysis section 81B, to the managing device 70 (Steps S15 and S18).
The operation and advantageous effects of the wear determination system 100 according to the above embodiments will be described below. In the above wear determination system 100, the worn-out state of the wear-prone component (the lever bush 41a and the winding bush 37a) is determined based on, not only the vibrations generated in the cradle 31, but also the number of rotations from the time point at which the braking device 39 that brakes the rotation of the package P starts operation until the time point at which the rotation of the package P stops. As a result, because the worn-out state of the wear-prone component is determined by taking into account "the vibrations generated in the cradle 31" and "the number of package rotations before stopping" that better reflect the worn-out state of the wear-prone component, a more appropriate determination can be made. As a result, the worn-out state of the wear-prone component in which wear out occurs while the package P is being rotated by the driving force of the driving section 47 can be determined with higher accuracy.
FIG. 11A is a diagram showing a relationship between the wear amount of the lever bush 41a, the wear amount of the winding bush 37a, and the magnitude of the vibrations in the first direction. FIG. 11B is a diagram showing a relationship between the wear amount of the lever bush 41a, the wear amount of the winding bush 37a, and the magnitude of the vibrations in the second direction. FIG. 11C is a diagram showing a relationship between the wear amount of the lever bush 41a, the wear amount of the winding bush 37a, and the number of package rotations before stopping.
The present inventors noticed that there is a uniqueness in the relationship between the wear amount of the lever bush 41a and the wear amount of the winding bush 37a, and the vibrations in the first direction, the vibrations in the second direction, or the number of package rotations before stopping. That is, with respect to the vibrations (vibrations in the first direction and vibrations in the second direction), it was found that the vibrations (for example, vibration acceleration) do not increase even if the lever bush 41a wears but the vibrations increase when the winding bush 37a wears (see FIGS. 11A and 11B). It was also found that, when the lever bush 41a wears, the number of rotations before the rotation of the package P stops increases, but when the winding bush 37a wears, the number of rotations before the rotation of the package P stops decreases (See FIG. 11C).
As a result, it was found that the combinations of the degree of wear of each of the lever bush 41a and the winding bush 37a have an interaction. Then, it was found that the accuracy of determining the worn-out state of the wear-prone component can be improved by combining these combinations with the vibrations or the number of rotations of the package P until the package P stops. In the wear determination system 100 according to the above embodiments, the vibrations and the number of rotations of the package P until the package P stops are taken into consideration, so that the accuracy of determining the worn-out state of the wear-prone component can be increased.
In the wear determination system 100, highly accurate determination of the worn-out state can be realized by machine learning. Further, once the analytical model M is generated by machine learning, the worn-out state can be easily and accurately estimated. The worker can easily grasp whether it is the time to replace the wear-prone component (whether the component should be replaced).
In the above wear determination system 100, because the analytical model M that calculates the worn-out state of the wear-prone component is a pre-trained model that is trained by using the teacher data in which the state variable and the worn-out state of the wear-prone component when the state variable is acquired are mapped, the determination accuracy of the worn-out state is high. Therefore, the worn-out state of the wear-prone component can be determined with higher accuracy.
Because the wear determination system 100 includes the teacher data input section 81E, the analytical model M can be updated as needed by using the teacher data input in the teacher data input section 81E. Accordingly, it is possible to generate the analytical model M capable of deriving a more accurate worn-out state, and it is possible to determine the worn-out state of the wear-prone component by using the highly accurate analytical model M.
The operation data can be continuously collected in the teacher data input section 81E of the wear determination system 100. Accordingly, the analytical model M can be generated (updated) based on a large number of the teacher data, so that the accuracy of the analytical model M can be improved. Further, by performing the verification by using the verification data, for example, erroneous learning, insufficient learning, or over-learning can be prevented.
In the above wear determination system 100, as the vibrations generated in the cradle 31, the vibrations in the first direction that matches the axial direction of the winding tube T and the vibrations in the second direction that is orthogonal to the first direction are acquired, and the acquired vibrations are used as the factors for determining the worn-out state. As a result, because the vibrations in the first direction and the vibrations in the second direction, which are the vibrations generated in the cradle 31, are used to determine the worn-out state of the wear-prone component, the worn-out state of the wear-prone component in which wear out occurs while the package P is being rotated by the driving section 47 can be determined with higher accuracy.
The notifying section 71D of the wear determination system 100 displays the determination result obtained by the analysis section 81B on at least one of the not-shown display unit that can communicate with the managing device 70, the display section 4 of the machine control device 2 (see FIG. 1), and the display section of the portable terminal 73 such as the tablet that can communicate with the managing device 70 or the machine control device 2 (see FIG. 6). As a result, the worn-out state of the wear-prone component is reported to the worker or the like, and the worker can appropriately take some action on the wear-prone component.
The above wear determination system 100 includes the managing section 71E that manages the maintenance schedule for each wear-prone component, and updates the schedule based on the determination result obtained by the analysis section 81B. As a result, even when the maintenance is scheduled at regular intervals, for example, if the worn-out state of the wear-prone component is bad, the scheduled maintenance date can be preponed, or if the worn-out state is not so bad, the scheduled maintenance date can be postponed. That is, the schedule can be updated appropriately according to the worn-out state of the wear-prone component. As a result, a more effective maintenance schedule can be presented.
Although the embodiments have been described above, the present invention is not limited to the above embodiments.
In the above embodiment, an example in which the worn-out state of the wear-prone component (the lever bush 41a and the winding bush 37a) is determined based on the multiple types of the state variables including the vibrations generated in the cradle 31 and the number of rotations from the time point at which the braking device 39 that brakes the rotation of the package P starts operation until the time point at which the rotation of the package P stops has been described, but the present invention is not limited to this. For example, in the wear determination system 100 according to First Modification, the winding speed of the package P in the winder unit 10 may be included in the state variable in addition to the state variables mentioned in the above embodiment. The winding speed of the package P can be acquired, for example, by referring to the setting value set by using the setting section 3 (or by referring to the operation data).
The analytical model M in this case is a pre-trained model that is pre-trained by using the teacher data in which is mapped the state variable that includes the vibrations generated in the cradle 31, the number of rotations from the time point at which the braking device 39 that brakes the rotation of the package P starts operation until the time point at which the rotation of the package P stops, and the winding speed of the package P, with the worn-out state of the lever bush 41a and the winding bush 37a at the time when the state variable is acquired. To generate the pre-trained model, the teacher data in which the winding speed of the package P is mapped as the state variable is used. In the wear determination system 100 according to First Modification, more appropriate wear determination can be performed.
In the wear determination system 100 according to Second Modification, information on the amount of yarn wound around the package P may be included in the state variable in addition to the state variables mentioned in the embodiment or First Modification. The "information on the amount of the yarn wound around the package P" includes least one among "the total length of the yarn Y wound around the winding tube T", "the total weight of the yarn Y wound around the winding tube T", and "the diameter (thickness) of the package P".
For example, the diameter (thickness) of the package P can be calculated based on the total length of the yarn Y wound around the package P (or the elapsed time from the start of winding of the yarn Y around the package P), the winding speed of the yarn Y, and the type (thickness or the like) of the yarn Y. Alternatively, the traverse angle is calculated based on the traverse speed and the running speed of the yarn Y in the running path of the yarn Y between the yarn feeding bobbin B and the drum 45, the peripheral speed of the package P is calculated based on the running speed of the yarn Y and the calculated traverse angle, and the diameter of the package P can be calculated based on the rotation speed of the package P and the calculated peripheral speed of the package P. In the wear determination system 100 according to Second Modification, more appropriate wear determination can be performed.
In the wear determination system 100 according to Third Modification, information on the state of the winding tube T may be included in the state variable in addition to the state variables mentioned in the embodiment, First Modification, or Second Modification. The "information on the state of the winding tube T" includes the degree of degradation of the function or performance of the winding tube T itself due to use or due to a defect. The state of the winding tube T may be determined by the worker by visual evaluation or may be determined based on a photograph of the winding tube T taken by a predetermined imaging method. In the wear determination system 100 according to Third Modification, more appropriate wear determination can be performed.
In the above embodiment, an example was described in which the number of rotations from the time point at which the braking device 39 that brakes the rotation of the package P starts operation until the time point at which the rotation of the package P stops is used as one of the state variables; however, the time until the rotation of the package P stops can be used as one of the state variables.
In the above embodiments and modifications, an example in which the wear determination system 100 is built with the server device 80 has been described; however, it can be built with the managing device 70, or with the machine control device 2 of the automatic winder 1. Further, the function realized by the managing section 71E or the like in the managing device 70 may be configured in the server device 80.
In the above embodiment, as the vibrations generated in the cradle 31, the vibrations in the first direction coinciding with the axial direction of the winding tube T and the vibrations in the second direction orthogonal to the first direction are detected, and the second direction is the front-back direction of the machine; however, this should not be taken as limiting. For example, the second direction can be inclined by 45 degrees in a vertical direction with respect to the front-back direction of the machine. Alternatively, a sensor that can measure vibrations along three directions can be mounted on the winder unit 10, vibrations in three directions, that is, the vibrations in the width direction (left-right direction) of the machine as a first direction coinciding with the axial direction of the winding tube T, the vibrations in the front-back direction of the machine that is a second direction, and the vibrations in a third direction that is orthogonal to the first direction and the second direction can be detected, and the vibrations along three directions can be used as the state variable.
In the embodiment, an example has been described in which the winding bush 37a provided at the large-diameter end (first end) of the cone-shaped package P is taken as a target for determining the worn-out state; however, instead of or in addition, the winding bush 37a provided at the small-diameter end (second end) may be taken as a target for determining the worn-out state. Also, when the package P supported by the cradle is cylindrical, at least one of the winding bushes 37a provided in the axial direction may be taken as a target for determining the worn-out state. When there is a plurality of winding bushes, the worn-out state can be represented by a collective worn-out state of all the winding bushes, or the worn-out state of each winding bush can be determined independently.
In one aspect of the present invention, the component that is the target for determining the worn-out state can be any component that can be worn out in the period during which the package P is rotated by the driving section, and such component should not be limited to the winding bush 37a and the lever bush 41a.
In the above embodiments and modifications, the determining section determined the worn-out state of the wear-prone component by using the analytical model generated by the machine learning; however, the method adopted by the determining section is not limited to using the analytical model generated by the machine learning. For example, the determining section can use a statistical method, artificial intelligence, or a threshold value obtained in advance by a numerical method to perform the determination. In the embodiments and the modifications, as an example of the machine learning method, supervised learning, in which learning is performed based on the teacher data, has been described; however, the machine learning can be performed by reinforcement learning.
There is no particular limitation on the timing at which the analysis section 81B determines the worn-out state of the wear-prone component. Such determination may be made each time the state variable is acquired by the acquiring section 81A. Alternatively, the state variable acquired by the acquiring section 81A is stored, and the determination may be made once a month, once in three months, or the like, for example. Alternatively, the analysis section 81B may determine the worn-out state in accordance with the standard maintenance period. For example, if the standard maintenance period is two years, the determination may be made, to be on the safer side, six months before the standard maintenance period expires.
At least some of the embodiments and the modifications described above can be arbitrarily combined.
According to one aspect of the present invention, a wear determination system that determines, in a fiber processing unit that forms a package by winding a yarn around a winding tube, among components constituting a cradle that rotatably supports the package, a worn-out state of a wear-prone component that wears out during a period in which the package is rotated by a driving section, includes an acquiring section that acquires as a state variable, vibrations generated in the cradle, and number of rotations or time from a time point at which a braking device that brakes rotation of the package starts operation to a time point at which the rotation of the package stops; and a determining section that determines the worn-out state of the wear-prone component based on the state variable acquired by the acquiring section. Note that, the acquiring section and the determining section here do not mean a sensor such as an element that collects target information and converts it into a signal that can be processed by an arithmetic device or the like, but mean a portion provided as a part of the function of the arithmetic device.
In the above wear determination system, the worn-out state of the wear-prone component is determined based on, not only the vibrations generated in the cradle, but also the number of rotations or the time from the time point at which the braking device that brakes the rotation of the package starts operation until the time point at which the rotation of the package stops. As a result, because the worn-out state of the wear-prone component is determined by taking into account a plurality of factors including at least "the vibrations generated in the cradle" and "the number of rotations of the package or the time before the rotation of the package stops", more appropriate determination can be performed. The present inventors noticed that "the vibrations generated in the cradle" and "the number of rotations of the package or the time before the rotation of the package stops" better reflect the worn-out state of the wear-prone component. As a result, by determining based on at least these two state variables the worn-out state of the wear-prone component in which wear out occurs while the package is being rotated by the driving section can be determined with higher accuracy.
The above wear determination system can further include a storage section that stores therein an analytical model generated by machine learning. The determining section can input the state variable in the analytical model and output the worn-out state of the wear-prone component as a determination result. According to this wear determination system, highly accurate determination of the worn-out state can be realized by machine learning. Further, once the analytical model is generated by machine learning, the worn-out state can be easily and accurately determined.
In the above wear determination system, the analytical model can be a pre-trained model obtained by learning by using teacher data in which the state variable and the worn-out state of the wear-prone component when the state variable is acquired are mapped with each other. In the analytical model of this wear determination system, because the worn-out state is derived from the analytical model obtained by learning by using the teacher data, a more accurate worn-out state can be derived. Therefore, the worn-out state of the wear-prone component can be determined with higher accuracy.
The above wear determination system can further include a receiving section that receives the teacher data; and a learning section that learns the worn-out state of the wear-prone component according to the teacher data received by the receiving section. In this wear determination system, it is possible to generate the analytical model capable of deriving a more accurate worn-out state, and the determining section can determine the worn-out state of the wear-prone component based on the worn-out state derived by using the highly accurate analytical model.
In the above wear determination system, the receiving section can be communicably connected to a collecting section that collects information on an operating state from each of a plurality of the fiber processing units, and the learning section can use a part of the information on the operating state collected by the collecting section as the teacher data. In this wear determination system, the information on the operating state can be continuously collected from each fiber processing unit. Accordingly, the analytical model can be generated (updated) based on a large number of the teacher data, so that the accuracy of the analytical model can be improved.
In the above wear determination system, the state variable can include winding speed of the package in the fiber processing unit. In this wear determination system, because the worn-out state of the wear-prone component can be determined based on three factors including the vibrations, the number of rotations or the time, and the winding speed, a more appropriate determination can be made. As a result, the worn-out state of the wear-prone component in which wear out occurs while the package is being rotated by the driving section can be determined with higher accuracy.
In the above wear determination system, the state variable can include information on amount of yarn wound around the package. The "information on the amount of the yarn wound around the package" includes least one among "the total length of the yarn wound around the winding tube", "the total weight of the yarn wound around the winding tube", and "the diameter (thickness) of the package". In this wear determination system, because the worn-out state of the wear-prone component can be determined based on three factors including the vibrations, the number of rotations or the time, and the information on the amount of the yarn, a more appropriate determination can be made. As a result, the worn-out state of the wear-prone component in which wear out occurs while the package is being rotated by the driving section can be determined with higher accuracy.
In the above wear determination system, the vibrations generated in the cradle can include vibrations in a first direction coinciding with an axial direction of the winding tube and vibrations in a second direction orthogonal to the first direction, and the state variable can include the vibrations in the first direction and the vibrations in the second direction. In this wear determination system, because the vibrations in the first direction and the vibrations in the second direction, which are the vibrations generated in the cradle, are used to determine the worn-out state of the wear-prone component, the worn-out state of the wear-prone component in which wear out occurs while the package is being rotated by the driving section can be determined with higher accuracy.
In the above wear determination system, the state variable can include information on a state of the winding tube. The "information on the state of the winding tube" includes the degree of degradation of the function or performance of the winding tube itself due to use or due to a defect. For example, as the vibrations increase when a winding tube in a bad state is used, it is possible to reflect the relationship between the state of the winding tube and the vibrations. As a result, the worn-out state of the wear-prone component in which wear out occurs while the package is being rotated by the driving section can be determined with higher accuracy.
In the above wear determination system, the cradle can include a cradle arm for holding the winding tube, the wear-prone component can be a winding bush that constitutes an axis receiving member that rotatably supports the winding tube held by the cradle arm, and the determining section can determine a worn-out state of the winding bush. In this wear determination system, the worn-out state of the winding bush can be determined with higher accuracy.
In the above wear determination system, the cradle can include a cradle arm that holds the winding tube; and a lever that rotates the cradle arm so that the cradle arm is in a closed state in which the cradle arm holds the winding tube or in an open state in which the hold of the cradle arm on the winding tube has been released, the wear-prone component can be a lever bush that constitutes a part of an axis receiving member that rotatably supports the lever, and the determining section can determine a worn-out state of the lever bush. In this wear determination system, the worn-out state of the lever bush can be determined with higher accuracy.
In the above wear determination system, the cradle can include a cradle arm that holds the winding tube; and a lever that rotates the cradle arm so that the cradle arm is in a closed state in which the cradle arm holds the winding tube or in an open state in which the hold of the cradle arm on the winding tube has been released, the wear-prone component can include a winding bush constituting an axis receiving member that supports rotation of the winding tube held by the cradle arm and a lever bush that constitutes a part of an axis receiving member that supports the lever, and the determining section can determine a worn-out state of each of the winding bush and the lever bush. In this wear determination system, the worn-out state of each of the winding bush and the lever bush can be determined with higher accuracy.
The above wear determination system can further include a notifying section that notifies of a determination result obtained by the determining section. In this wear determination system, because the worn-out state of the wear-prone component is reported to the worker or the like, the worker can appropriately take some action on the wear-prone component.
The above wear determination system can further include a managing section that manages a maintenance schedule for each of the wear-prone components. The managing section can update the schedule based on the determination result obtained by the determining section. In this wear determination system, even when the maintenance has a regular scheduled schedule, the schedule can be updated appropriately according to the worn-out state of the wear-prone component. As a result, a more effective maintenance schedule can be presented.
According to another aspect of the present invention, a textile machine system includes the above wear determination system; a plurality of textile machines each including a plurality of the fiber processing units and a control device that collectively controls the fiber processing units; a managing device that manages the textile machines; and a server device capable of communicating with the managing device. The acquiring section and the determining section are included in the server device, the control device collects the state variable from each of the fiber processing units and sends the acquired state variables to the server device, and the server device can send the determination result obtained by the determining section to the managing device. In this textile machine system, for example, the state variable collected from each of the fiber processing units arranged in a spinning factory in each region is sent to the server device and the worn-out state of the wear-prone component included in each of the fiber processing units is determined based on the sent state variable. As a result, the worn-out state of the wear-prone component in which wear out occurs while the package is being rotated by the driving section can be determined with higher accuracy.
According to still another aspect of the present invention, a textile machine system includes the above wear determination system; a plurality of textile machines each including a plurality of the fiber processing units and a control device that collectively controls the fiber processing units; and a managing device that manages the textile machines. The acquiring section and the determining section are included in the managing device, the control device collects the state variable from each of the fiber processing units and sends the acquired state variables to the managing device, and the managing device can include an output section that outputs the determination result obtained by the determining section. In this textile machine system, for example, the state variable collected from each of the fiber processing units arranged in a spinning factory is sent to the server device and the worn-out state of the wear-prone component included in each of the fiber processing units is determined based on the sent state variable. As a result, the worn-out state of the wear-prone component in which wear out occurs while the package is being rotated by the driving section can be determined with higher accuracy.
According to still another aspect of the present invention, a textile machine system includes the above wear determination system; and a textile machine including a plurality of the fiber processing units and a control device that collectively controls the fiber processing units. The acquiring section and the determining section are included in the control device, the control device collects the state variable from each of the fiber processing units, and the control device can include an output section that outputs the determination result obtained by the determining section. In this textile machine system, based on the state variable collected from each of the fiber processing units arranged in a spinning factory, the worn-out state of the wear-prone component included in each of the fiber processing units is determined. As a result, the worn-out state of the wear-prone component in which wear out occurs while the package is being rotated by the driving section can be determined with higher accuracy.
According to the present invention, the worn-out state of the wear-prone component in which wear out occurs while the package is being rotated by the driving section can be determined with higher accuracy.
In the above explanation, the meaning of "a plurality of" also includes "a predetermined number of".
Although the invention has been explained with respect to specific embodiments for a complete and clear disclosure, the appended claims are not to be thus limited but are to be construed as embodying all modifications and alternative constructions that may occur to one skilled in the art that fairly fall within the scope of the claims.