WO2020245915A1

WO2020245915A1 - Learning model generation method, computer program, learning model, control device, and control method

Info

Publication number: WO2020245915A1
Application number: PCT/JP2019/022172
Authority: WO
Inventors: 智浩北村
Original assignee: ホソカワミクロン株式会社
Priority date: 2019-06-04
Filing date: 2019-06-04
Publication date: 2020-12-10
Also published as: JPWO2020245915A1

Abstract

Provided is a learning model generation method, a computer program, a learning model, a control device, and a control method.　Using a computer, a powder processing apparatus equipped with at least a function of pulverizing powder raw materials acquires measurement data that indicates the operation state of the powder processing apparatus and particle diameter data of powder obtained from the powder processing apparatus, and generates a learning model that outputs a computation result for a control parameter for controlling the operation state of the powder processing apparatus in accordance with input of the particle diameter desired by the user by using the acquired measurement data and particle diameter data as training data.

Description

Learning model generation method, computer program, learning model, control device, and control method

The present invention relates to a learning model generation method, a computer program, a learning model, a control device, and a control method.

The powder processing process is composed of a combination of various processes such as storage, supply, transportation, pulverization, classification, and mixing (see, for example, Patent Document 1). In order to stably obtain a product or intermediate having the quality desired by the user, it is necessary to appropriately set the operating conditions of the powder processing apparatus.

Japanese Unexamined Patent Publication No. 2008-194592

However, the powder handled in the powder processing process exhibits a wide variety of properties depending on the type, size, shape, etc. of the powder. Therefore, there is a problem that the operating conditions set in the powder processing apparatus often depend on the experience and intuition of field engineers.

The present invention provides a learning model generation method, a computer program, a learning model, a control device, and a control method capable of determining control parameters related to a powder processing apparatus without depending on the experience and intuition of a field engineer. The purpose.

The method for generating a learning model according to one aspect of the present invention includes measurement data indicating an operating state of the powder processing apparatus with respect to the powder processing apparatus having at least a function of crushing the powder raw material using a computer. The powder particle size data obtained from the powder processing apparatus is acquired, and the acquired measurement data and the particle size data are used as teacher data, and the powder is input according to the input of the particle size desired by the user. Generate a learning model that outputs the calculation results for the control parameters that control the operating state of the processing device.

The computer program according to one aspect of the present invention includes measurement data indicating the operating state of the powder processing apparatus and the powder processing apparatus for the powder processing apparatus having at least a function of crushing the powder raw material in the computer. The particle size data of the powder obtained from the above is acquired, and the acquired measurement data and the particle size data are used as the training data, and the operating state of the powder processing apparatus is performed according to the input of the particle size desired by the user. It is a computer program for executing a process of generating a learning model that outputs a calculation result for a control parameter that controls.

The learning model according to one aspect of the present invention includes an input layer in which a particle size desired by the user is input, an output layer that outputs calculation results for control parameters that control the operating state of the powder processing apparatus, and the powder. An intermediate layer in which the relationship between the particle size and the control parameter is learned by using the measurement data indicating the operating state of the processing device and the particle size data of the powder obtained from the powder processing device as training data. When the particle size desired by the user is input to the input layer, the computer calculates the intermediate layer and outputs the calculation result for the control parameter that controls the operating state of the powder processing apparatus. Make it work.

The control device according to one aspect of the present invention shows a receiving unit that receives an input of a particle size desired by a user, particle size data of powder obtained from the powder processing device, and an operating state of the powder processing device. A learning model in which the relationship between the particle size of the powder and the control parameters that control the operating state of the powder processing apparatus is learned by using the measurement data as the teacher data, and the reception unit accepts the learning model. A calculation processing unit that inputs the data of the particle size to the learning model and executes the calculation by the learning model, and a control unit that controls the operation of the device including the powder processing device based on the calculation result by the learning model. And.

In the control method according to one aspect of the present invention, the input of the particle size desired by the user is received, and the received particle size data is the particle size data of the powder obtained from the powder processing device and the powder processing device. Using the measurement data indicating the operating state of the powder as the teacher data, the relationship between the particle size of the powder and the control parameters that control the operating state of the powder processing apparatus is input to the learning model. , The operation by the learning model is executed, and the operation of the apparatus including the powder processing apparatus is controlled based on the calculation result obtained from the learning model.

According to the present application, control parameters related to the powder processing apparatus can be determined without depending on the experience and intuition of field engineers.

It is a schematic diagram which shows the whole structure of the powder processing system which concerns on Embodiment 1. FIG. It is a schematic cross-sectional view which shows the structure of the powder processing apparatus which concerns on Embodiment 1. FIG. It is a block diagram which shows the internal structure of the control device which concerns on Embodiment 1. FIG. It is a block diagram which shows the internal structure of a terminal device. It is a schematic diagram which shows the structural example of the learning model in Embodiment 1. FIG. It is a conceptual diagram which shows an example of the data which a control device collects. It is a flowchart explaining the generation procedure of the learning model by a control device. It is a flowchart explaining the control procedure by a control device. It is a schematic cross-sectional view which shows the structure of the powder processing apparatus which concerns on Embodiment 2. It is a schematic diagram which shows the structural example of the learning model in Embodiment 2. It is a schematic cross-sectional view which shows the structure of the powder processing apparatus which concerns on Embodiment 3. It is a schematic cross-sectional view which shows the structure of the powder processing apparatus which concerns on Embodiment 4. FIG. It is a schematic diagram which shows the structural example of the learning model in Embodiment 4. It is a schematic cross-sectional view which shows the structure of the powder processing apparatus which concerns on Embodiment 5. It is a schematic diagram which shows the structural example of the learning model in Embodiment 5. It is a schematic diagram which shows the structural example of the learning model in Embodiment 6. It is a schematic diagram which shows the structural example of the learning model in Embodiment 7. It is a conceptual diagram which shows the example of collecting data in Embodiment 8. It is a flowchart explaining the generation procedure of the learning model which concerns on Embodiment 8. It is a flowchart explaining the control procedure by a control device. It is a flowchart explaining the re-learning procedure of a learning model. It is a flowchart explaining the adjustment procedure of a control parameter. It is a flowchart explaining the procedure of the process executed by a terminal device and a control device. It is a schematic diagram which shows an example of an input screen. It is a schematic diagram which shows an example of a monitoring screen.

Hereinafter, the present invention will be specifically described with reference to the drawings showing the embodiments thereof.
(Embodiment 1)
FIG. 1 is a schematic view showing the overall configuration of the powder processing system 1 according to the first embodiment. The powder processing system 1 according to the first embodiment is, for example, a raw material supply machine 2, a hot air generator 3, a powder processing device 4A, a cyclone 5, a dust collector 6, a blower 7 (or a pump), and a control device 100 (FIG. 3).

The raw material supply machine 2 is a device for supplying the powder raw material to the powder processing device 4A. The powder raw material supplied by the raw material supply machine 2 to the powder processing device 4A is, for example, a raw material for fine powdered toner used for coloring paper in a copier or a laser printer. Not limited to toner raw materials, powder paints, battery materials, magnetic materials, dyes, resins, waxes, polymers, pharmaceuticals, catalysts, metal powders, silica, solder, cement, foods, inorganic materials, organic materials, or metal materials It may be a powder raw material for producing powder.

The raw material supply machine 2 is connected to the powder processing device 4A via the raw material supply path TP1. A screw feeder 21 (see FIG. 2) for transporting the powder raw material is provided in the raw material supply path TP1. The screw feeder 21 is preferable when the powder raw material is a solid and the powder raw material is continuously charged at a constant speed. A double damper, a rotary valve, or the like may be used instead of the screw feeder 21. Further, the hot air generated by the hot air generator 3 may be introduced into the raw material supply path TP1 and the powder raw material may be supplied to the powder processing apparatus 4A together with the hot air. Further, the raw material feeder 2 performs weight control using a weight sensor S1 (see FIG. 3) such as a load cell, and the amount of retention in the device is constant even when continuous processing is performed in the powder processing device 4A. The supply amount of the powder raw material may be adjusted so as to be. Further, the raw material supply machine 2 supplies the powder raw material per unit time (that is, the supply speed) based on the output of the built-in timer (not shown) and the supply amount of the powder raw material measured by the weight sensor S1. May be measured. The weight sensor S1 may be provided not only in the raw material supply machine 2, but also in the powder processing device 4A, the cyclone 5, and the dust collector 6.

The hot air generator 3 is a device for generating hot air to be introduced into the powder processing device 4A, and includes a heat source such as a heating heater, a blower, and a control device for controlling the temperature and air volume of the hot air. The hot air generator 3 is not limited to the above configuration, and a known configuration may be used. For example, a part of the hot air introduced into the powder processing apparatus 4A may be recovered and circulated between the hot air generator 3 and the powder processing apparatus 4A.

The hot air generator 3 is connected to the powder processing device 4A via the gas introduction path TP2. The hot air generated by the hot air generator 3 is introduced into the powder processing apparatus 4A via the gas introduction path TP2. The temperature of the hot air generated by the hot air generator 3 is appropriately set according to the powder processed by the powder processing apparatus 4A. For example, hot air of about 200 ° C. to 600 ° C. may be generated in order to dry the powder processed by the powder processing apparatus 4A.

The powder processing apparatus 4A is an apparatus for producing powder having a particle size smaller than a predetermined particle size by crushing the powder raw material supplied from the raw material supply machine 2 and classifying the obtained powder. The powder processing apparatus 4A according to the present embodiment mainly performs a pulverization treatment for crushing a powder raw material and a classification treatment for classifying the pulverized powder (for example, ACM Parberizer manufactured by Hosokawa Micron Co., Ltd. (registered trademark). )).

The powder processed by the powder processing apparatus 4A is transported to the cyclone 5 via the powder transport path TP3. In the present embodiment, the particle size sensor S2 is installed in the middle of the powder transport path TP3 from the powder processing device 4A to the cyclone 5, and the particle size of the powder passing through the powder processing device 4A by the particle size sensor S2. Is measured at regular or regular timings (for example, every 5 seconds). The powder collected by the cyclone 5 is taken out to the product tank 8 and collected as a product.

The particle size sensor S2 is a device that measures the particle size distribution using, for example, a laser diffraction / scattering method, and outputs values of D10, D50, and D90. Here, D10, D50, and D90 represent particle diameters corresponding to cumulative 10%, 50%, and 90% from the small diameter side of the cumulative volume distribution in the particle size distribution, respectively. The cumulative volume distribution is a distribution representing the relationship between the particle size (μm) of the powder and the integration frequency (volume%) from the small diameter side. D50 is also generally referred to as an average particle diameter (median diameter). Instead of D10, D50, D90, a mode diameter representing the particle diameter having the largest appearance ratio in the frequency distribution of the particle diameter may be used, and various arithmetic average values (number average, length average, area average, volume average, etc.) may be used. ) May be used.
In the present embodiment, the particle size sensor S2 is installed in the powder transport path TP3, but the raw material supply machine 2, the path from the cyclone 5 to the product tank 8, and the dust collector 6 to the dust collector tank 9 are reached. It may be installed on a route or the like.

A dust collector 6 is connected to the cyclone 5 via a dust collection path TP4. The dust collector 6 includes a bug filter for collecting fine powder or the like that has passed through the cyclone 5. The gas that has passed through the bug filter of the dust collector 6 flows to the blower 7 through the exhaust air passage TP5 and is discharged from the discharge port of the blower 7. On the other hand, the fine powder or the like collected by the bug filter of the dust collector 6 is taken out to the dust collection tank 9 and collected.

A blower 7 is connected to the dust collector 6 via an exhaust air passage TP5. By driving the blower 7, the gas flow from the powder processing apparatus 4A to the cyclone 5 (that is, the gas flow for extracting the powder from the powder processing apparatus 4A) and the gas flow from the cyclone 5 to the dust collector 6 Form a flow. In the present embodiment, a flow rate sensor S3 (see FIG. 3) is installed in the middle of the exhaust passage TP5 from the dust collector 6 to the blower 7, and the discharge / suction flow rate when taking out the powder from the powder processing apparatus 4A is measured. Measure at regular or regular timing (for example, every 5 seconds). Instead of installing the flow rate sensor S3 in the exhaust air passage TP5, the flow rate sensor S3 is installed in the raw material supply path TP1, the gas introduction path TP2, the powder transport path TP3, the dust collection path TP4, the discharge port of the blower 7, and the like. You may.

The control device 100 is a device that controls the operation of devices including the powder processing device 4A, and is connected to and from these devices so that various data can be exchanged. Here, the device including the powder processing device 4A includes the powder processing device 4A, and may further include at least one of the raw material supply machine 2, the hot air generator 3, the cyclone 5, the dust collector 6, and the blower 7. It represents that. In the following description, the device including the powder processing device 4A is also described as a device to be controlled.

FIG. 2 is a schematic cross-sectional view showing the configuration of the powder processing apparatus 4A according to the first embodiment. The powder processing apparatus 4A includes a cylindrical casing 410 that performs powder processing inside the powder processing apparatus 4A. The casing 410 is provided with a raw material input port 411, a gas introduction port 412, a crushing rotor 413, a guide ring 414, a classification rotor 415, a powder outlet 416, and the like.

As the material of the casing 410, a known material conventionally used for the casing of the powder processing apparatus may be used. Specifically, iron-based steel materials such as SS400, S25C, S45C, SPHC (Steel Plate Hot Commercial), stainless steel materials such as SUS304 and SUS316, iron casting materials such as FC20 and FC40, and stainless casting materials such as SCS13 and 14 Metal, ceramics, glass, etc. may be used. Further, aluminum, other wood, or synthetic resin may be used as long as an abrasion resistant material is attached to the inner wall surface.

In order to improve the durability of the equipment, the inner surface of the casing 410 is subjected to plating treatment such as hard chrome plating treatment, abrasion resistant thermal spraying material treatment such as tungsten carbide spraying, metal vapor deposition under vacuum, carbon vapor deposition of diamond structure, etc. Abrasion resistant treatment may be applied.

Further, when powder treatment of a low melting point resin component such as toner is performed in the casing 410, if the fixed component is mixed in the product, the quality becomes poor. Therefore, in order to prevent turbulence of the air flow due to adhesion or fixation of the treated powder or blockage in the casing 410, the inner surface of the casing 410 is buffed, electrolytically polished, coated with PTFE (Polytetrafluoroethylene), or plated with nickel or the like. Treatment may be applied.

The casing 410 is provided with a raw material input port 411 for charging the powder raw material supplied from the raw material supply machine 2 into the casing 410. The raw material input port 411 is preferably provided at a position above the rotary disk 413A of the crushing rotor 413. The powder raw material supplied from the raw material supply machine 2 is conveyed by the screw feeder 21 in the raw material supply path TP1 and is charged into the casing 410 from the raw material input port 411.

The casing 410 is provided with a gas introduction port 412 for introducing hot air (gas) from the hot air generator 3 into the casing 410. The gas introduction port 412 is connected to the hot air generator 3 via the gas introduction path TP2. The position of the gas introduction port 412 is not particularly limited, but it is preferably provided at a position lower than the crushing rotor 413 so that the gas is introduced into the casing 410 via the rotating crushing rotor 413. In the present embodiment, the gas is introduced from the direction intersecting the rotation direction of the crushing rotor 413, but the gas may be introduced along the rotation direction of the crushing rotor 413.

The gas introduced from the gas introduction port 412 forms an air flow that circulates while swirling inside the casing 410, and reaches the cyclone 5 and the dust collector 6 from the powder outlet 416 via the classification rotor 415 from the inside of the casing 410. To do. The airflow in the casing 410 may be formed by suction by the blower 7 connected via the cyclone 5 and the dust collector 6, or may be formed by blowing (pressurizing) from the gas introduction port 412 side. The type of gas introduced into the casing 410 may be appropriately determined according to the target processed product. For example, air may be used, or an inert gas such as nitrogen or argon may be used to prevent oxidation.

Depending on the powder to be treated, the temperature rise in the casing 410 may cause the powder to soften, and the powders may be fused to each other to cause variations in particle size or decrease in yield. .. Therefore, temperature sensors S4 (see FIG. 3) may be provided at one or a plurality of locations in the casing 410 to control the temperature inside the casing 410. For example, when the powder to be treated is toner having a low melting point, for example, the temperature of the gas introduced into the casing 410 may be adjusted so that the exhaust temperature at the powder outlet 416 is 35 to 55 ° C. ..
Further, the temperature sensor S4 may be provided in the cyclone 5, the dust collector 6, the product tank 8, the dust collecting tank 9, and the like.

Further, in the present embodiment, the hot air from the hot air generator 3 is introduced into the casing 410, but using a cold air generator (not shown in the figure), the temperature is about −20 ° C. to 5 ° C. in the casing 410. It may be configured to introduce the cold air of. In this case, the gas introduced into the casing 410 is preferably a dehumidified gas in order to prevent dew condensation. In addition, when processing foods that lose their flavor due to heat or are easily deteriorated, cold air adjusted to 0 to 15 ° C. may be used.

Further, in order to adjust the internal temperature of the casing 410, a jacket portion may be provided around the casing 410. The jacket portion adjusts the internal temperature of the casing 410 by circulating and supplying a heating fluid or a cooling fluid from a tank provided separately.

The crushing rotor 413 is a rotor including a rotary disk 413A and a plurality of hammers 413B protruding upward from the upper peripheral edge of the rotary disk 413A, and the rotation speed is adjusted to a desired speed by the power of the crushing motor 413M (see FIG. 3). It is configured to rotate. Here, the crushing motor 413M is one of the driving units included in the powder processing system 1. The rotation speed of the crushing motor 413M is measured by the rotation speed sensor S5 (see FIG. 3) at regular or periodic timings (for example, at 5-second intervals). A plurality of hammers 413B of the crushing rotor 413 are arranged at equal intervals in the circumferential direction on the upper peripheral edge of the rotary disk 413A. The shape, size, number, and material of the hammer 413B are appropriately designed according to the required particle size, circularity, and the like of the product powder. For example, although FIG. 2 shows a rod-shaped hammer 413B, it may be a rectangular parallelepiped hammer or a trapezoidal hammer in a plan view. Further, instead of the hammer 413B, a blade-like structure may be used.

The crushing rotor 413 is rotated by the power of the crushing motor 413M to generate a swirling airflow in the casing 410, and the powder raw material introduced into the casing 410 is impacted, compressed, and ground by the action of the hammer 413B. The powder raw material is crushed by giving mechanical energy such as shearing.

As the material of the crushing rotor 413, a known material conventionally used for the crushing rotor of the powder processing apparatus may be used. For example, SS400, S25C, S45C, SUS304, SUS316, SUS630 and the like can be used. Further, for the hammer 413B, a cemented carbide tip may be attached so that the hammer 413B can withstand an impact force, or a composite of a metal such as ceramics or cermet having wear resistance and toughness and ceramics may be used.

Further, the surface of the crushing rotor 413 is subjected to plating treatment such as hard chrome plating, abrasion resistant thermal spraying material treatment such as tungsten carbide spraying, metal vapor deposition under vacuum, and carbon vapor deposition of diamond structure in order to improve the durability of the device. Abrasion resistant treatment such as SUS630 and quench hardening treatment of SUS630 may be performed. Further, the surface of the crushing rotor 413 may be buffed, electrolytically polished, coated with PTFE, or plated with nickel or the like.

In the present embodiment, the configuration of the hammer 413B protruding upward from the upper peripheral edge portion of the rotating disk 413A has been described, but a hammer protruding downward from the lower surface peripheral edge portion of the rotating disk 413A may be used. The hammer protruding downward in this way does not directly pulverize the powder raw materials in the casing 410, but can form a strong swirling airflow in the casing 410, so that the powder raw materials collide with each other. The powder raw material can be indirectly crushed.

Further, a crushing liner may be provided on the inner peripheral surface of the casing 410 at a position facing the hammer 413B. The crushing liner is a tubular member having a central axis along the rotation axis direction of the crushing rotor 413, and even if the inner peripheral surface of the tubular member is provided with a triangular, corrugated, or wedge-shaped groove. Good.

The guide ring 414 is a cylindrical member for generating a swirling air flow in the casing 410 and guiding the powder processed in the casing 410 to the classification rotor 415. The guide ring 414 is arranged coaxially with the crushing rotor 413 above the crushing rotor 413 and fixed inside the casing 410. The fixing method of the guide ring 414 is not particularly limited, but it is necessary to fix the guide ring 414 without rotating inside the casing 410 during the operation of the powder processing apparatus 4A. This is to control the flow state of the powder to be processed to an appropriate state inside the casing 410. In the example of FIG. 2, the guide ring 414 whose inner diameter is continuously increased from the lower side to the upper side in the casing 410 is shown, but the guide ring whose inner diameter is continuously decreased toward the upper side is used. It may be a guide ring whose inner diameter does not change in the vertical direction.

The classification rotor 415 is a rotor provided with a plurality of classification blades 415A arranged radially, and is configured to rotate at a desired rotation speed by the power of the classification motor 415M (see FIG. 3). Here, the classification motor 415M is one of the driving units included in the powder processing system 1. The rotation speed of the classification motor 415M is measured by the rotation speed sensor S6 (see FIG. 3) at regular or periodic timings (for example, at 5-second intervals). The classification rotor 415 is provided above the crushing rotor 413, and only the powder having a particle size smaller than a predetermined particle size is passed through the powder processed in the casing 410 by centrifugal force due to high-speed rotation. Only the body is guided to the powder outlet 416.

Here, the particle size of the powder passing through the classification rotor 415 can be set by controlling the rotation speed of the classification rotor 415 and the like. That is, by controlling the rotation speed of the classification rotor 415, powder having a particle size smaller than a predetermined particle size can be taken out from the casing 410. On the other hand, the powder that cannot pass through the classification rotor 415 circulates in the casing 410 and is repeatedly processed.

As the material of the classification rotor 415, a known material conventionally used for the classification rotor of the powder processing apparatus may be used. For example, SS400, S25C, S45C, SUS304, SUS316, titanium, titanium alloy, aluminum alloy and the like can be used. In addition, the surface of the classification rotor 415 is specially subjected to heat hardening treatment such as carburizing and quenching, tungsten carbide spraying material treatment, plating treatment such as hard chrome plating, and heat hardening treatment after spraying in order to improve the durability of the device. Abrasion resistant treatment such as thermal spray material treatment, metal vapor deposition performed under vacuum, and carbon vapor deposition of a diamond structure may be performed.

Further, in order to prevent the powder from adhering or sticking to the classification rotor 415, the surface of the classification rotor 415 may be buffed, electrolytically polished, coated with PTFE, or plated with nickel or the like.

Although the powder processing apparatus 4A according to the present embodiment includes a crushing rotor 413 and a classification rotor 415, it may be used as a classification machine without using the crushing rotor 413. Further, since it is possible to introduce hot air from the hot air generator 3, the powder processing device 4A may be used as a dryer. Further, the powder processing apparatus 4A may be used as a particle design apparatus by providing a coarse powder recovery port in the casing 410 and collecting the coarse powder. Further, a plurality of types of raw materials may be supplied to the casing 410 and used as a continuous mixer in which the plurality of types of raw materials are mixed in the casing 410.

Further, in addition to the various sensors described above, a dust content concentration sensor for measuring the dust content concentration and a powder composition are added to at least one of the devices including the powder processing device 4A or at least one of the paths between the devices. A measurement sensor such as a NIR sensor for measuring, a moisture sensor for measuring the moisture content of powder, a pressure sensor for measuring pressure, and a sound pressure / frequency sensor for measuring sound pressure or frequency may be provided.

Hereinafter, the configuration of the control system in the powder processing system 1 will be described.
FIG. 3 is a block diagram showing an internal configuration of the control device 100 according to the first embodiment. The control device 100 is composed of a general-purpose or dedicated computer, and includes a control unit 101, a storage unit 102, an input unit 103, an output unit 104, a communication unit 105, an operation unit 106, and a display unit 107.

The control unit 101 includes, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. The ROM included in the control unit 101 stores a control program or the like that controls the operation of each hardware unit included in the control device 100. The CPU in the control unit 101 executes a control program stored in the ROM and various computer programs stored in the storage unit 102, which will be described later, to control the operation of each hardware unit, thereby as a control device according to the present invention. To realize the function of. The RAM included in the control unit 101 temporarily stores data and the like used during execution of the calculation.

The control unit 101 is configured to include a CPU, a ROM, and a RAM, but has a GPU (Graphics Processing Unit), an FPGA (Field Programmable Gate Array), a DSP (Digital Signal Processor), a quantum processor, and a volatile or non-volatile memory. It may be one or more arithmetic circuits or control circuits including the above. Further, the control unit 101 may have functions such as a clock for outputting date and time information, a timer for measuring the elapsed time from giving the measurement start instruction to giving the measurement end instruction, and a counter for counting the number.

The storage unit 102 includes a storage device that uses a hard disk, a flash memory, or the like. The storage unit 102 stores a computer program executed by the control unit 101, various data acquired from the outside, various data generated inside the device, and the like.

The computer program stored in the storage unit 102 includes a control program PG1 for controlling the operation of the device to be controlled, a particle size of the powder processed by the powder processing device 4A, and control parameters related to the powder processing device 4A. Includes a learning program PG2 and the like for learning the relationship with.

The computer program including the control program PG1 and the learning program PG2 may be provided by the non-temporary recording medium M1 in which the computer program is readablely recorded. The recording medium M1 is, for example, a portable memory such as a CD-ROM, a USB memory, a compact flash (registered trademark), an SD (Secure Digital) card, or a micro SD card. The control unit 101 reads various programs from the recording medium M1 using a reading device (not shown in the figure), and stores the read various programs in the storage unit 102.

In the present embodiment, when the control unit 101 executes the learning program PG2, the control device 100 shifts to the learning phase. The learning phase is carried out, for example, when the control device 100 is introduced. Further, the learning phase may be executed when a user's instruction is received through the operation unit 106 or at a regular timing.

In the learning phase, the control unit 101 includes the rotation speed of the crushing rotor 413 included in the powder processing apparatus 4A, the rotation speed of the classification rotor 415, and the discharge / suction flow rate when the powder is taken out from the casing 410 (processing chamber). The measurement data and the particle size data of the powder obtained from the powder processing apparatus 4A are acquired. These measurement data and particle size data may be collected in advance or may be collected after the transition to the learning phase. The control unit 101 uses the collected measurement data and particle size data as teacher data to learn the relationship between the particle size of the powder obtained from the powder processing device 4A and the control parameters related to the powder processing device 4A. , Generate a learning model 200. Here, the control parameters relating to the powder processing apparatus 4A include control parameters for directly controlling the operation of the powder processing apparatus 4A (for example, the rotation speed of the crushing rotor 413 and the rotation speed of the classification rotor 415). .. Further, the control parameters related to the powder processing device 4A may include control parameters for controlling the operations of the raw material supply machine 2, the hot air generator 3, the dust collector 6, and the blower 7 attached to the powder processing device 4A. In the present embodiment, the particle size obtained from the powder processing device 4A, the rotation speed of the crushing rotor 413 included in the powder processing device 4A, the rotation speed of the classification rotor 415, and the powder from the processing chamber of the powder processing device 4A. The configuration for learning the relationship with the control parameters that control the discharge / suction flow rate when taking out the body will be described.

The learning model 200 generated in the learning phase is stored in the storage unit 102. The learning model 200 is defined by its definition information. The definition information of the learning model 200 includes, for example, the structural information of the learning model 200, parameters such as weights and biases between nodes.

Further, in the present embodiment, the control device 100 shifts to the operation phase when the control unit 101 executes the control program PG1. The operation phase is carried out after the execution of the learning model 200.

In the operation phase, the control device 100 controls the operation of the device including the powder processing device 4A. At this time, the control unit 101 may accept the input of the particle diameter (desired particle diameter) desired by the user. The control unit 101 inputs the received particle size data into the learning model 200 and executes an operation using the learning model 200 to obtain the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the powder. Acquires the calculation result related to the control parameters that control the discharge / suction flow rate when taking out. The control unit 101 controls the operation of the device to be controlled based on the calculation result obtained from the learning model 200.

The input unit 103 includes a connection interface for connecting the device to be controlled. The device connected to the input unit 103 includes a raw material supply machine 2, a hot air generator 3, a powder processing device 4A, a cyclone 5, a dust collector 6, and a blower 7. The connection interface included in the input unit 103 may be a wired interface or a wireless interface. Data sent from the raw material supply machine 2, the hot air generator 3, the powder processing device 4A, the cyclone 5, the dust collector 6, and the blower 7 are input to the input unit 103. The data input to the input unit 103 is measured by the weight sensor S1, the particle size sensor S2, the flow rate sensor S3, the temperature sensor S4, the rotation speed sensor S5 for the crushing rotor 413, and the rotation speed sensor S6 for the classification rotor 415. The measurement data to be performed is included.

In the present embodiment, the device to be controlled is connected to the input unit 103, but at least a part of the sensors S1 to S6 may be directly connected to the input unit 103. In this case, the measurement data output from the sensors S1 to S6 is directly input to the input unit 103 without going through the device to be controlled. Further, when the sensors S1 to S6 have a communication interface, the control device 100 may acquire measurement data through the communication unit 105 described later.

The output unit 104 includes a connection interface for connecting the device to be controlled. The device connected to the output unit 104 includes a raw material supply machine 2, a hot air generator 3, a powder processing device 4A, a cyclone 5, a dust collector 6, and a blower 7. The connection interface included in the output unit 104 may be a wired interface or a wireless interface. The control unit 101 controls the operation of the device to be controlled by outputting a control command through the output unit 104. For example, when controlling the rotation speed of the crushing rotor 413, the control unit 101 generates a control command for the crushing motor 413M, which is a driving unit of the crushing rotor 413, and outputs the control command to the powder processing apparatus 4A through the output unit 104. , Control the rotation speed of the crushing rotor 413. The same applies to the case of controlling the rotation speed of the classification rotor 415, and the control unit 101 generates a control command for the classification motor 415M which is a drive unit of the classification rotor 415 and outputs the control command to the powder processing device 4A through the output unit 104. Thereby, the rotation speed of the classification rotor 415 is controlled. Further, when controlling the discharge / suction flow rate from the casing 410 included in the powder processing apparatus 4A, the control unit 101 generates a control command for the blower 7 and outputs the control command to the blower 7 through the output unit 104, whereby the casing 410. Controls the discharge / suction flow rate from.

The communication unit 105 includes a communication interface for transmitting and receiving various communication data. The communication interface included in the communication unit 105 is, for example, a communication interface conforming to the communication standard of LAN (Local Area Network) used in WiFi (registered trademark) and Ethernet (registered trademark). Alternatively, a communication interface conforming to a communication standard such as Bluetooth (registered trademark), ZigBee (registered trademark), 3G, 4G, 5G, LTE (Long Term Evolution) may be used.

The communication unit 105 communicates with, for example, the terminal device 500 used by the user of the powder processing system 1. The communication unit 105 may receive the operation data or the setting data transmitted from the terminal device 500 in order to accept the remote operation of the powder processing system 1. When the control unit 101 receives the operation data or the setting data transmitted from the terminal device 500 through the communication unit 105, the control unit 101 executes a process according to the received operation data or the setting data. For example, when the user receives the data of the desired particle size, the control unit 101 may execute the calculation using the learning model 200 and perform the process of acquiring the control parameters related to the powder processing apparatus 4A. Further, the control unit 101 may generate screen data of the user interface screen to be displayed on the terminal device 500, and may transmit the generated screen data to the terminal device 500 through the communication unit 105.

The operation unit 106 is provided with an input interface such as a keyboard and a mouse, and accepts various operations and various settings. The control unit 101 performs appropriate processing based on various operations and various settings received through the operation unit 106, and stores the setting information in the storage unit 102 as necessary. In the present embodiment, the control device 100 is provided with the operation unit 106, but the operation unit 106 is not indispensable, and the operation is received through a computer (for example, the terminal device 500) connected to the outside. You may.

The display unit 107 includes a display panel such as a liquid crystal panel or an organic EL (Electro-Luminescence) panel, and displays information to be notified to the user. The display unit 107 may display, for example, the measurement data of the various sensors S1 to S6 received through the communication unit 105, or may display information based on various operations and various settings received through the operation unit 106. Further, the display unit 107 may display the calculation result by the learning model 200. In the present embodiment, the control device 100 is configured to include the display unit 107, but the display unit 107 is not essential and outputs information to be notified to the user to an external computer (for example, the terminal device 500). , The information may be displayed on the output destination computer.

In the present embodiment, the control device 100 has been described as a single computer, but it does not have to be a single computer, and may be composed of a plurality of computers or a plurality of virtual computers.

FIG. 4 is a block diagram showing the internal configuration of the terminal device 500. The terminal device 500 is a computer such as a personal computer, a smartphone, or a tablet terminal, and includes a control unit 501, a storage unit 502, a communication unit 503, an operation unit 504, and a display unit 505.

The control unit 501 includes, for example, a CPU, a ROM, a RAM, and the like. The ROM included in the control unit 501 stores a control program or the like that controls the operation of each hardware unit included in the terminal device 500. The CPU in the control unit 501 executes a control program stored in the ROM and various computer programs stored in the storage unit 502 described later, and controls the operation of each hardware unit. Further, the control unit 501 may have functions such as a clock for outputting date and time information, a timer for measuring the elapsed time from giving the measurement start instruction to giving the measurement end instruction, and a counter for counting the number. Data and the like used during execution of the calculation are temporarily stored in the RAM included in the control unit 501.

The storage unit 502 includes a storage device that uses a hard disk, a flash memory, or the like. The storage unit 502 stores a computer program executed by the control unit 501, various data acquired from the outside, various data generated inside the device, and the like. The computer program stored in the storage unit 502 may include an application program for accessing the control device 100 from the terminal device 500.

The communication unit 503 includes a communication interface for transmitting and receiving various data. The communication interface included in the communication unit 503 is, for example, a communication interface conforming to the LAN communication standard used in WiFi (registered trademark) and Ethernet (registered trademark). Alternatively, a communication interface conforming to communication standards such as Bluetooth (registered trademark), ZigBee (registered trademark), 3G, 4G, 5G, and LTE may be used.

The communication unit 503 communicates with, for example, the control device 100 of the powder processing system 1. The data received by the terminal device 500 through the communication unit 503 includes screen data for displaying the interface screen of the control device 100 on the display unit 505, data indicating the setting state and control state of the device including the powder processing device 4A, and the like. Including. Here, the data indicating the set state includes a set value regarding the rotation speed of the crushing rotor 413, a set value regarding the rotation speed of the classification rotor 415, a set value regarding the discharge / suction flow rate by the blower 7, and the like. The data indicating the control state includes a measured value related to the rotation speed of the crushing rotor 413, a measured value related to the rotation speed of the classification rotor 415, a measured value related to the discharge / suction flow rate by the blower 7, and a powder processing device 4A. It includes alarm information and the like output from. The data received by the communication unit 503 is output to the control unit 501. The data transmitted by the terminal device 500 through the communication unit 503 includes operation data, setting data, and the like when the powder processing system 1 is remotely controlled.

The operation unit 504 is equipped with an input interface such as a keyboard and a mouse, and accepts various operations and various settings. The control unit 501 performs appropriate processing based on various operations and various settings received through the operation unit 504, and stores the setting information in the storage unit 502 as necessary.

The display unit 505 includes a display panel such as a liquid crystal panel or an organic EL panel, and displays information to be notified to the user. The display unit 505 displays the interface screen of the control device 100, for example, based on the screen data received by the communication unit 503. Further, the display unit 505 displays the setting state and the control state of the device including the powder processing device 4A based on the data indicating the setting state and the control state of the device including the powder processing device 4A received by the communication unit 503. You may.

Hereinafter, the learning model 200 used in the control device 100 will be described.
FIG. 5 is a schematic diagram showing a configuration example of the learning model 200 according to the first embodiment. The learning model 200 is, for example, a learning model for machine learning including deep learning, and is configured by a neural network. The learning model 200 includes an input layer 201,

intermediate layers

202A and 202B, and an output layer 203. In the example of FIG. 5, two

intermediate layers

202A and 202B are described, but the number of intermediate layers is not limited to two and may be three or more.

The input layer 201, the

intermediate layers

202A, 202B, and the output layer 203 have one or more nodes, and the nodes of each layer are connected to the nodes existing in the previous and next layers in one direction with a desired weight and bias. Has been done. The same number of data as the number of nodes included in the input layer 201 is input to the input layer 201 of the learning model 200. In the present embodiment, the data input to the node of the input layer 201 is the data of the particle diameter (desired particle diameter) desired by the user. Here, an example of the particle diameter data input to the node of the input layer 201 is the median diameter. The diameter is not limited to the median, and may be a mode diameter or various arithmetic mean values. Further, the particle diameter data given to the node of the input layer 201 is not limited to a single value such as the median diameter, the mode diameter, and various arithmetic mean values, but may be each value of D10, D50, and D90. It may be a range set for each of D10, D50, and D90 (that is, an upper limit value and a lower limit value of a range allowed by the user with respect to the particle size).

The particle size data input to the learning model 200 is output to the node included in the first intermediate layer 202A through the nodes constituting the input layer 201. The data input to the first intermediate layer 202A is output to the nodes included in the next intermediate layer 202B through the nodes constituting the intermediate layer 202A. At this time, the output is calculated using the activation function including the weights and biases set between the nodes. Hereinafter, in the same manner, the calculation using the activation function including the weight and the bias set between the nodes is executed, and the calculation is transmitted to the subsequent layers one after another until the calculation result by the output layer 203 is obtained. Parameters such as weights and biases that connect the nodes are learned by a predetermined learning algorithm. As a learning algorithm for learning various parameters, for example, a learning algorithm for deep learning is used. In the present embodiment, the particle size data regarding the particle size obtained from the powder processing device 4A (that is, the particle size data measured by the particle size sensor S2) and the rotation of the crushing rotor 413 included in the powder processing device 4A. Various parameters including weights and biases between nodes are learned by a predetermined learning algorithm using the measurement data including the speed, the rotation speed of the classification rotor 415, and the discharge / suction flow rate when taking out the powder from the casing 410 as training data. can do.

The output layer 203 outputs calculation results related to control parameters that control the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410. As the calculation result, for example, the probability indicating the quality of the combination of the plurality of control parameters described above may be output. Specifically, the output layer 203 is composed of n nodes from the first node to the nth node, and from the first node, the rotation speed of the crushing rotor 413 is G1, the rotation speed of the classification rotor 415 is C1, and the discharge. The probability P1 that the suction flow rate is V1 is output, and the rotation speed P2 of the crushing rotor 413 is G2, the rotation speed of the classification rotor 415 is C2, and the discharge / suction flow rate is V2 is output from the second node. ..., The probability Pn that the rotation speed of the crushing rotor 413 is Gn, the rotation speed of the classification rotor 415 is Cn, and the discharge / suction flow rate is Vn may be output from the nth node. The number of nodes constituting the output layer 203 and the calculation result assigned to each node are not limited to the above examples, and can be appropriately designed.

The control device 100 includes particle size data regarding the particle size obtained from the powder processing device 4A, the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410 of the powder processing device 4A. The measurement data including the above is collected, and these data are used as the teacher data to generate the learning model 200 as described above.

FIG. 6 is a conceptual diagram showing an example of data collected by the control device 100. The control unit 101 of the control device 100 acquires and acquires measurement data from the particle size sensor S2, the rotation speed sensor S5 of the crushing rotor 413, the rotation speed sensor S6 of the classification rotor 415, and the flow rate sensor S3 through the input unit 103. The data is stored in the storage unit 102 together with the time stamp, and the data used for the teacher data is collected. FIG. 6 shows an example of the data collected by the control device 100. Among the data shown in FIG. 6, the first record shows that the actual measurement value of the rotation speed of the crushing rotor 413 is 4749.6 rpm, the actual measurement value of the rotation speed of the classification rotor 415 is 3195.1 rpm, and the actual measurement value of the discharge / suction flow rate. When was 2281.2 m ³ / h, it indicates that the measured values of D10, D50, and D90 were 0.71 μm, 2.09 μm, and 9.06 μm, respectively. The same applies to the second and subsequent records, which are the actual measurement values of the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410 when the powder processing system 1 is operating. Is stored in the storage unit 102 in association with the data (D10, D50, D90) regarding the particle size of the powder obtained from the powder processing apparatus 4A. Although the example of FIG. 6 shows an example in which data is collected at intervals of 5 seconds, the time interval for collecting data is not limited to 5 seconds and may be set arbitrarily.

The control unit 101 uses the collected data as the teacher data to learn the relationship between the particle size of the powder obtained from the powder processing apparatus 4A and the control parameters related to the powder processing apparatus 4A, as described above. The learning model 200 is generated.

Hereinafter, the operation of the control device 100 in the learning phase will be described.
FIG. 7 is a flowchart illustrating a procedure for generating the learning model 200 by the control device 100. The control unit 101 of the control device 100 includes measurement data including the rotation speed of the crushing rotor 413 provided in the powder processing device 4A, the rotation speed of the classification rotor 415, and the discharge / suction flow rate when the powder is taken out from the casing 410. The particle size data of the powder obtained from the powder processing apparatus 4A is collected (step S101). The collected measurement data and particle size data are stored in the storage unit 102 together with the time stamp.

After collecting the measurement data and the particle size data, the control unit 101 selects a set of teacher data from the data stored in the storage unit 102 (step S102). That is, the control unit 101 has the measured values of the rotation speeds of the crushing rotor 413 and the classification rotor 415, the discharge / suction flow rate from the casing 410, and the powder particle diameter values obtained at that time (D10, D50, Select only one set with D90).

Next, the control unit 101 inputs the value of the particle diameter included in the selected teacher data into the learning model 200 (step S103), and executes the calculation by the learning model 200 (step S104). That is, the control unit 101 inputs the value of the particle diameter to the nodes constituting the input layer 201 of the learning model 200, performs the calculation using the weights and biases between the nodes in the

intermediate layers

202A and 202B, and outputs the calculation result. The process of outputting from the node of layer 203 is performed. In the initial stage before the start of learning, it is assumed that the definition information describing the learning model 200 is given an initial value.

Next, the control unit 101 evaluates the calculation result obtained in step S104 (step S105), and determines whether or not the learning is completed (step S106). Specifically, the control unit 101 can evaluate the calculation result by using an error function (also referred to as an objective function, a loss function, or a cost function) based on the calculation result obtained in step S104 and the teacher data. The control unit 101 is in the process of optimizing (minimizing or maximizing) the error function by a gradient descent method such as the steepest descent method, and when the error function is below the threshold value (or above the threshold value), the learning is completed. to decide. In order to avoid the problem of overfitting, techniques such as cross-validation and early stopping may be adopted to end learning at an appropriate timing.

When it is determined that the learning is not completed (S106: NO), the control unit 101 updates the weights and biases between the nodes of the learning model 200 (step S107), returns the process to step S102, and another teacher. Continue learning with the data. The control unit 101 may update the weights and biases between the nodes by using the error back propagation method in which the weights and biases between the nodes are sequentially updated from the output layer 203 to the input layer 201 of the learning model 200. it can.

When it is determined that the learning is completed (S106: YES), the control unit 101 stores the learned learning model 200 in the storage unit 102 (step S108), and ends the process according to this flowchart.

As described above, in the learning phase, the control device 100 according to the present embodiment includes particle size data regarding the particle size obtained from the powder processing device 4A, the rotation speed of the crushing rotor 413, and the rotation speed of the classification rotor 415. And the measurement data including the discharge / suction flow rate when the powder is taken out from the casing 410 is collected. By using the collected data as teacher data, the control device 100 uses the collected data as the teacher data, and according to the input of the particle size desired by the user, the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate to the casing 410. It is possible to generate a learning model 200 that outputs the calculation result related to the control parameter that controls.

In the present embodiment, the control device 100 is configured to generate the learning model 200, but even if an external server (not shown) for generating the learning model 200 is provided and the learning model 200 is generated by the external server. Good. In this case, the control device 100 may acquire the learning model 200 from the external server by communication or the like, and store the acquired learning model 200 in the storage unit 102.

Next, the operation of the control device 100 in the operation phase will be described. In the operation phase, it is assumed that the learning model 200 has already been learned.

FIG. 8 is a flowchart illustrating a control procedure by the control device 100. The control unit 101 of the control device 100 receives an input of a particle size (desired particle size) desired by the user through the operation unit 106 (step S121). Here, the control unit 101 may accept the value of D50 as the particle diameter desired by the user. Further, the control unit 101 may accept the respective values of D10, D50, and D90 instead of the value of D50, and may accept the upper limit value and the lower limit value for each of D10, D50, and D90. That is, the particle size data may be input according to the configuration of the input layer 201 included in the learning model 200.

Next, the control unit 101 inputs the received particle size data to the input layer 201 of the learning model 200, and executes the calculation by the learning model 200 (step S122). At this time, the control unit 101 gives the received particle size data to the node of the input layer 201. The data given to the node of the input layer 201 is output to the node of the adjacent intermediate layer 202A. In the intermediate layer 202A, an operation using an activation function including weights and biases between nodes is performed, and the operation result is output to the intermediate layer 202B in the subsequent stage. In the intermediate layer 202B, an operation using an activation function including weights and biases between nodes is further performed, and the operation result is output to each node of the output layer 203. Each node of the output layer 203 outputs the calculation result regarding the control parameter of the powder processing apparatus 4A. Specifically, each node of the output layer 203 outputs a calculation result regarding control parameters for controlling the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410.

Next, the control unit 101 acquires the calculation result from the learning model 200 (step S123) and determines the control parameters used for control (step S124). The output layer 203 of the learning model 200 outputs calculation results related to control parameters for controlling the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410. More specifically, in the output layer 203, the probability Pi (i = 1 to n) that the rotation speed of the crushing rotor 413 is Gi, the rotation speed of the classification rotor 415 is Ci, and the discharge / suction flow rate is Vi is obtained from each node. Output. The control unit 101 determines the control parameters used for control by specifying the combination of the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410, which have the highest probability.

Next, the control unit 101 executes control based on the control parameters determined in step S124 (step S125). That is, the control unit 101 generates a control command for the crushing motor 413M and the classification motor 415M so that the rotation speeds of the crushing rotor 413 and the classification rotor 415 become the values determined in step S124, and powder processing is performed through the output unit 104. Output to device 4A. Further, the control unit 101 generates a control command for the blower 7 so that the discharge / suction flow rate from the casing 410 becomes the value determined in step S124, and outputs the control command to the blower 7 through the output unit 104.

As described above, the control device 100 according to the present embodiment receives the input of the particle size desired by the user, thereby performing the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge from the casing 410. The control parameters that control the suction flow rate can be determined. That is, in the present embodiment, the control parameters related to the powder processing apparatus 4A can be determined without depending on the experience and intuition of the field engineer. The control device 100 executes control based on the determined control parameters so that a powder having a desired particle size can be obtained.

In the present embodiment, a configuration for acquiring calculation results related to control parameters using a machine learning learning model 200 configured by a neural network has been described, but the learning model 200 is limited to a model obtained by using a specific method. Not done.
For example, instead of the neural network by deep learning, a learning model by a perceptron, a convolutional neural network, a recurrent neural network, a residual network, a self-organizing map, or the like may be used.
In addition, instead of the above learning model by the neural network, a regression analysis method including linear regression, logistic regression, support vector machine, etc., and a method using a search tree such as a decision tree, a regression tree, a random forest, and a gradient boosting tree. , Bayesian estimation method including simple bays, AR (Auto Regressive), MA (Moving Average), ARIMA (Auto Regressive Integrated Moving Average), time series prediction method including state space model, clustering method including K neighborhood method, etc. With learning models learned by methods using ensemble learning including boosting, bagging, etc., hierarchical clustering, non-hierarchical clustering, clustering methods including topic models, association analysis, emphasis filtering, etc. There may be.
Further, a learning model may be constructed using multivariate analysis including PLS (Partial Least Squares) regression, multiple regression analysis, principal component analysis, factor analysis, cluster analysis and the like.

(Embodiment 2)
In the second embodiment, an example of application to a pin mill (for example, a fine impact mill manufactured by Hosokawa Micron Co., Ltd.) will be described. The powder processing system 1 in the second embodiment includes a pin mill (powder processing device 4B shown in FIG. 9) instead of the powder processing device 4A, but the other configurations are the same as those in the first embodiment. Therefore, the detailed description thereof will be omitted.

FIG. 9 is a schematic cross-sectional view showing the configuration of the powder processing apparatus 4B according to the second embodiment. The powder processing apparatus 4B includes a cylindrical casing 420 that performs powder processing inside the powder processing apparatus 4B. The casing 420 is provided with a raw material input port 421, a fixed crushing rotor 422, a rotary crushing rotor 423, a powder outlet 424, and the like.

As the material of the casing 420, it is preferable that the same material as the casing of the powder processing apparatus 4A described in the first embodiment is used. That is, iron-based steel materials such as SS400, S25C, S45C, and SPHC, stainless steel materials such as SUS304 and SUS316, iron casting materials such as FC20 and FC40, metals such as stainless casting materials such as SCS13 and FC14, ceramics, glass and the like. Should be used. Further, aluminum, other wood, or synthetic resin may be used as long as an abrasion resistant material is attached to the inner wall surface.

Further, on the inner surface of the casing 420, in order to improve the durability of the device, plating treatment such as hard chrome plating treatment, abrasion resistant thermal spraying material treatment such as tungsten carbide spraying, metal vapor deposition performed under vacuum, carbon of diamond structure Abrasion resistant treatment such as thermal spraying may be applied.

The casing 420 is provided with a raw material input port 421 for charging the powder raw material supplied from the raw material supply machine 2 into the casing 420. The raw material input port 421 is provided above, for example, the crushing

rotors

422 and 423.

The crushing rotor 422 is a fixed rotor, and includes a disk 422A fixed in the casing 420 and a plurality of pins 422B protruding toward the crushing rotor 423 side. On the other hand, the crushing rotor 423 is a rotary rotor, which includes a rotary disk 423A and a plurality of pins 423B protruding toward the crushing rotor 422 side, and is desired by the power of a crushing motor 413M (see FIG. 3). It is configured to rotate at a rotational speed. The pin 423B of the crushing rotor 423 is positioned so as not to collide with the pin 422B of the fixed crushing rotor 422 when the rotary disk 423A rotates. In the present embodiment, the shapes, dimensions, arrangement, number, and materials of the pins 422B and 423B included in the crushing

rotors

422 and 423 are appropriately designed according to the required particle size and circularity of the product powder. Will be done.

The crushing rotor 423 is rotated by the power of the crushing motor 413M to generate a swirling airflow in the casing 420, and the powder raw materials introduced into the casing 420 are impacted, compressed, and abraded by the action of the pins 422B and 423B. The powder raw material is crushed by giving mechanical energy such as crushing and shearing.

As the material of the crushing

rotors

422 and 423, known materials conventionally used for the crushing rotor of the powder processing apparatus may be used. For example, SS400, S25C, S45C, SUS304, SUS316, SUS630 and the like can be used. Further, for the pins 422B and 423B, a cemented carbide tip may be attached so as to withstand the impact force, or a composite of metal and ceramics such as ceramics having wear resistance and toughness or cermet may be used. Good.

Further, the surface of the crushing rotor 422,423 is subjected to plating treatment such as hard chrome plating, abrasion resistant thermal spraying material treatment such as tungsten carbide spraying, metal vapor deposition performed under vacuum, and diamond structure in order to improve the durability of the device. Abrasion resistant treatment such as carbon vapor deposition, quenching hardening treatment of SUS630, and the like may be performed. Further, the surfaces of the crushing

rotors

422 and 423 may be buffed, electrolytically polished, coated with PTFE, or plated with nickel or the like.

A cyclone 5, a dust collector 6, and a blower 7 are connected to the powder outlet 424 via a powder transport path TP3 or the like. The powder crushed by the crushing

rotors

422 and 423 is taken out from the powder outlet 424 together with the air flow.

The control device 100 generates a learning model that outputs as a calculation result related to the control parameters of the powder processing device 4B in response to the input of the desired particle size, so that the rotation speed of the crushing rotor 423 measured by the rotation speed sensor S5 is generated. , The discharge / suction flow rate measured by the flow rate sensor S3, and the particle size data measured by the particle size sensor S2 are collected as teacher data.

Using the collected teacher data, the control unit 101 of the control device 100 uses the collected teacher data to input the particle size desired by the user, the rotation speed of the crushing rotor 423, and the discharge when the powder is taken out from the powder outlet 424. A learning model 210 (see FIG. 10) that outputs calculation results related to control parameters that control the suction flow rate is generated. Since the procedure for generating the learning model 210 is the same as that in the first embodiment, the description thereof will be omitted. The learning model 210 may be generated by the control device 100 or by an external server device (not shown).

FIG. 10 is a schematic diagram showing a configuration example of the learning model 210 according to the second embodiment. Similar to the learning model 200 described in the first embodiment, the learning model 210 includes an input layer 211,

intermediate layers

212A, 212B, and an output layer 213, each of which has one or more nodes. The number of intermediate layers is not limited to two, and may be three or more.

The learning model 210 according to the second embodiment relates to a control parameter for controlling the rotation speed of the crushing rotor 423 and the discharge / suction flow rate from the casing 420 with respect to the input of the particle diameter (desired particle diameter) desired by the user. It is configured to output the calculation result.

When the control device 100 performs an calculation using the learning model 210, the control device 100 inputs the particle diameter (desired particle diameter) desired by the user into the learning model 210. The particle size data input to the learning model 210 is output to the node included in the first intermediate layer 212A through the nodes constituting the input layer 211. The data input to the first intermediate layer 212A is output to the nodes included in the next intermediate layer 212B through the nodes constituting the intermediate layer 212A. At this time, the output is calculated using the activation function including the weights and biases set between the nodes. Hereinafter, in the same manner, the calculation using the activation function including the weight and the bias set between the nodes is executed, and the calculation is transmitted to the subsequent layers one after another until the calculation result by the output layer 213 is obtained.

The output layer 213 outputs the calculation result regarding the rotation speed of the crushing rotor 423 and the control parameters for controlling the discharge / suction flow rate when the powder is taken out from the casing 420. As the calculation result, for example, the probability indicating the quality of the combination of the plurality of control parameters described above may be output. Specifically, the output layer 213 is composed of n nodes from the first node to the nth node, and from the first node, the rotation speed of the crushing rotor 423 is G1 and the discharge / suction flow rate is V1. Is output, and the probability P2 that the rotation speed of the crushing rotor 423 is G2 and the discharge / suction flow rate is V2 is output from the second node, ..., The rotation speed of the crushing rotor 423 is Gn, discharge / discharge from the nth node. The probability Pn that the suction flow rate is Vn may be output. The number of nodes constituting the output layer 213 and the calculation result assigned to each node are not limited to the above examples, and can be appropriately designed.

The control unit 101 of the control device 100 specifies the combination having the highest probability among the probabilities of being output from the output layer 213 (in the present embodiment, the combination of the rotation speed of the crushing rotor 423 and the discharge / suction flow rate). By doing so, the control parameters used for control are determined. The control unit 101 generates a control command for controlling the operation of the crushing motor 413M and the blower 7 based on the determined control parameters, and outputs the control command to each device through the output unit 104.

As described above, in the second embodiment, with respect to the pin mill (powder processing apparatus 4B), the powder is taken out from the rotation speed of the crushing rotor 423 and the powder outlet 424 according to the input of the particle size desired by the user. A learning model 210 that outputs calculation results related to control parameters for controlling the discharge / suction flow rate at the time is generated. Further, by using the learning model 210, calculation results regarding control parameters for controlling the rotation speed of the crushing rotor 423 and the discharge / suction flow rate from the casing 420 can be obtained. The control device 100 can control the operation of the powder processing system 1 based on the calculation result of the learning model 210 so that the powder having a desired particle size can be obtained.

(Embodiment 3)
In the third embodiment, an example of application to an impact crusher (for example, Grassis (registered trademark) manufactured by Hosokawa Micron Co., Ltd.) will be described. The powder processing system 1 in the third embodiment includes an impact type crusher (powder processing device 4C shown in FIG. 11) instead of the powder processing device 4A, but the other configurations are the embodiment. Since it is the same as 1, the detailed description thereof will be omitted.

FIG. 11 is a schematic cross-sectional view showing the configuration of the powder processing apparatus 4C according to the third embodiment. The powder processing apparatus 4C includes a cylindrical casing 430 that performs powder processing inside the powder processing apparatus 4C. The casing 430 is provided with a raw material input port 431, a crushing rotor 432, a liner 433, a powder outlet 434, and the like.

As the material of the casing 430, it is preferable to use the same material as the casing of the powder processing apparatus 4A described in the first embodiment. That is, iron-based steel materials such as SS400, S25C, S45C, and SPHC, stainless steel materials such as SUS304 and SUS316, iron casting materials such as FC20 and FC40, metals such as stainless casting materials such as SCS13 and FC14, ceramics, glass and the like. Should be used. Further, aluminum, other wood, or synthetic resin may be used as long as an abrasion resistant material is attached to the inner wall surface.

Further, on the inner surface of the casing 430, in order to improve the durability of the device, plating treatment such as hard chrome plating treatment, abrasion resistant thermal spraying material treatment such as tungsten carbide spraying, metal vapor deposition performed under vacuum, carbon of diamond structure Abrasion resistant treatment such as thermal spraying may be applied.

The casing 430 is provided with a raw material input port 431 for charging the powder raw material supplied from the raw material supply machine 2 into the casing 430. The raw material input port 431 is provided above, for example, the crushing rotor 432. The powder raw material supplied from the raw material supply machine 2 is conveyed by, for example, a screw feeder (not shown) in the raw material supply path TP1 and is charged into the casing 430 from the raw material input port 431. Further, a refrigerant may be introduced into the casing 430. The refrigerant may be introduced, for example, through a refrigerant introduction port provided on the peripheral surface of the casing 430 or a rotating shaft of the crushing rotor 432.

The crushing rotor 432 includes a plurality of

rotating disks

432A, 432A, ..., 432A arranged coaxially along the rotation axis, and is rotated at a desired rotation speed by the power of the crushing motor 413M (see FIG. 3). It is configured in. The peripheral surface of the rotating disk 432A is provided with triangular, corrugated, and wedge-shaped grooves. A liner 433 is arranged on the inner peripheral surface of the casing 430 at a position facing the peripheral surface of the rotating disk 432A. The liner 433 is a tubular member having a central axis along the rotation axis direction of the crushing rotor 432, and a triangular, corrugated, and wedge-shaped groove is provided on the inner peripheral surface of the tubular member. ..

As the material of the crushing rotor 432 and the liner 433, known materials that have been conventionally used may be used. For example, SS400, S25C, S45C, SUS304, SUS316, SUS630 and the like can be used. Further, a cemented carbide chip may be attached so as to withstand an impact force, or a composite of a metal and a ceramic such as ceramics or cermet having wear resistance and toughness may be used.

Further, the surfaces of the crushing rotor 432 and the liner 433 are subjected to plating treatment such as hard chrome plating, abrasion resistant thermal spraying material treatment such as tungsten carbide spraying, metal vapor deposition performed under vacuum, and diamond structure in order to improve the durability of the device. Abrasion resistant treatment such as carbon vapor deposition of SUS630, quenching hardening treatment of SUS630, and the like may be performed. Further, the surfaces of the crushing rotor 432 and the liner 433 may be buffed, electrolytically polished, coated with PTFE, or plated with nickel or the like.

The crushing rotor 432 is rotated by the power of the crushing motor 413M to generate a swirling airflow in the casing 430, and the powder raw material introduced into the casing 430 is impacted and compressed by the action of the crushing rotor 432 and the liner 433. , Grinding, shearing and other mechanical energy to crush the powder raw material.

A cyclone 5, a dust collector 6, and a blower 7 are connected to the powder outlet 434 via a powder transport path TP3 or the like. The powder crushed by the crushing rotor 432 and the liner 433 is taken out from the powder outlet 434 together with the air flow.

The control device 100 generates a learning model that outputs as a calculation result related to the control parameters of the powder processing device 4C in response to the input of the desired particle size, so that the rotation speed of the crushing rotor 432 measured by the rotation speed sensor S5 is generated. , The discharge / suction flow rate measured by the flow rate sensor S3, and the particle size data measured by the particle size sensor S2 are collected as teacher data.

The control unit 101 of the control device 100 uses the collected teacher data to input the particle size desired by the user, the rotation speed of the crushing rotor 432, and the discharge when the powder is taken out from the powder outlet 434. Generate a learning model that outputs the calculation results related to the control parameters that control the suction flow rate. The learning model may be generated by the control device 100 or by an external server device (not shown).
Since the configuration and generation procedure of the learning model generated in the third embodiment are the same as those in the second embodiment, the description thereof will be omitted.

As described above, in the third embodiment, with respect to the powder processing apparatus 4C, the rotation speed of the crushing rotor 432 and the discharge when the powder is taken out from the powder outlet 434 according to the input of the particle size desired by the user. -Generate a learning model that outputs the calculation results related to the control parameters for controlling the suction flow rate. Further, by using the learning model, calculation results regarding control parameters for controlling the rotation speed of the crushing rotor 432 and the discharge / suction flow rate from the casing 430 can be obtained. The control device 100 can control the operation of the powder processing system 1 based on the calculation result of the learning model so that the powder having a desired particle size can be obtained.

(Embodiment 4)
In the fourth embodiment, an example of application to a medium stirring type powder processing apparatus (for example, Purvis (registered trademark) manufactured by Hosokawa Micron Co., Ltd.) will be described. The powder processing system 1 according to the fourth embodiment includes a medium stirring type powder processing device (powder processing device 4D shown in FIG. 12) instead of the powder processing device 4A, but has other configurations. Is the same as that of the first embodiment, and therefore detailed description thereof will be omitted.

FIG. 12 is a schematic cross-sectional view showing the configuration of the powder processing apparatus 4D according to the fourth embodiment. The powder processing apparatus 4D includes a cylindrical casing 440 that performs powder processing inside the powder processing apparatus 4D. The casing 440 is provided with a raw material input port 441, a gas introduction port 442, a crushing unit 443, a classification rotor 444, a powder outlet 445, an agitator 446, and the like.

As the material of the casing 440, it is preferable that the same material as the casing of the powder processing apparatus 4D described in the first embodiment is used. That is, iron-based steel materials such as SS400, S25C, S45C, and SPHC, stainless steel materials such as SUS304 and SUS316, iron casting materials such as FC20 and FC40, metals such as stainless casting materials such as SCS13 and FC14, ceramics, glass and the like. Should be used. Further, aluminum, other wood, or synthetic resin may be used as long as an abrasion resistant material is attached to the inner wall surface.

Further, on the inner surface of the casing 440, in order to improve the durability of the device, plating treatment such as hard chrome plating treatment, abrasion resistant thermal spraying material treatment such as tungsten carbide spraying, metal vapor deposition performed under vacuum, carbon of diamond structure Abrasion resistant treatment such as thermal spraying may be applied.

The casing 440 is provided with a raw material input port 441 for charging the powder raw material supplied from the raw material supply machine 2 into the casing 440. The raw material input port 441 is provided above, for example, the crushing section 443. The powder raw material supplied from the raw material supply machine 2 is conveyed by, for example, a screw feeder (not shown) in the raw material supply path TP1 and is charged into the casing 440 from the raw material input port 441.

In the powder processing apparatus 4D according to the fourth embodiment, an air flow is generated inside the casing 440 by introducing a gas such as compressed air into the casing 440 through the gas introduction port 442. Further, the powder processing apparatus 4D includes a crushing unit 443 having an agitator 446 that agitates balls (or beads) as a medium. The agitator 446 is configured to rotate at a desired rotation speed by the power of the stirring motor 446M. Here, the stirring motor 446M is one of the driving units included in the powder processing system 1. The powder raw material charged into the casing 440 is in a state of covering the surface of the medium. In the crushing section 443, the medium is agitated by rotating the agitator 446, and the powder raw material covering the surface of the medium is given impact, compression, shearing, and grinding to crush the powder raw material. it can. The crushed powder is conveyed together with the air flow and is classified by the classification rotor 444 provided on the upper part of the apparatus.

The configuration of the classification rotor 444 is the same as that of the first embodiment. That is, the classification rotor 444 is a rotor provided with a plurality of classification blades 444A arranged radially, and is configured to rotate at a desired rotation speed by the power of the classification motor 415M. The rotation speed of the classification motor is measured by the rotation speed sensor S6 (see FIG. 3) at regular or periodic timings (for example, at 5-second intervals). The classification rotor 444 passes only powder having a particle size smaller than a predetermined particle size among the powders processed in the casing 440 by centrifugal force due to high-speed rotation, and guides only the passed powder to the powder outlet 445. On the other hand, the powder that cannot pass through the classification rotor 444 circulates in the casing 440 and is repeatedly processed.

Here, the particle size of the powder passing through the powder outlet 445 can be set by controlling the rotation speed of the agitator 446, the rotation speed of the classification rotor 444, and the discharge / suction flow rate.

The control device 100 generates a learning model 220 that outputs control parameters related to the powder processing device 4D as a calculation result in response to an input of a desired particle size, so that the rotation speed of the agitator 446 measured by the rotation speed sensor S5 is generated. , The rotation speed of the classification rotor 444 measured by the rotation speed sensor S6, the discharge / suction flow rate measured by the flow rate sensor S3, and the particle size data measured by the particle size sensor S2 are collected as teacher data.

Using the collected teacher data, the control unit 101 of the control device 100 uses the collected teacher data to input the particle size desired by the user, the rotation speed of the agitator 446, the rotation speed of the classification rotor 444, and the powder from the powder outlet 445. Generates a learning model 220 (see FIG. 13) that outputs calculation results related to control parameters that control the discharge / suction flow rate when taking out. Since the procedure for generating the learning model 220 is the same as that in the first embodiment, the description thereof will be omitted. The learning model 220 may be generated by the control device 100 or may be generated by an external server device (not shown).

FIG. 13 is a schematic diagram showing a configuration example of the learning model 220 in the fourth embodiment. Similar to the learning model 200 described in the first embodiment, the learning model 220 includes an input layer 221 having one or a plurality of nodes,

intermediate layers

222A and 222B, and an output layer 223, respectively. The number of intermediate layers is not limited to two, and may be three or more.

The learning model 220 according to the fourth embodiment has the rotation speed of the agitator 446, the rotation speed of the classification rotor 444, and the discharge / suction flow rate from the casing 440 in response to the input of the particle diameter (desired particle diameter) desired by the user. It is configured to output the calculation result related to the control parameters that control.

When the control device 100 performs an operation using the learning model 220, the control device 100 inputs the particle diameter (desired particle diameter) desired by the user into the learning model 220. The particle size data input to the learning model 220 is output to the node included in the first intermediate layer 222A through the nodes constituting the input layer 221. The data input to the first intermediate layer 222A is output to the node included in the next intermediate layer 222B through the nodes constituting the intermediate layer 222A. At this time, the output is calculated using the activation function including the weights and biases set between the nodes. Hereinafter, in the same manner, the calculation using the activation function including the weight and the bias set between the nodes is executed, and the calculation is transmitted to the subsequent layers one after another until the calculation result by the output layer 223 is obtained.

The output layer 223 outputs calculation results related to control parameters that control the rotation speed of the agitator 446, the rotation speed of the classification rotor 444, and the discharge / suction flow rate when the powder is taken out from the casing 440. As the calculation result, for example, the probability indicating the quality of the combination of the plurality of control parameters described above may be output. Specifically, the output layer 223 is composed of n nodes from the first node to the nth node, and from the first node, the rotation speed of the agitator 446 is A1, the rotation speed of the classification rotor 444 is C1, and discharge. The probability P1 that the suction flow rate is V1 is output, and the rotation speed of the agitator 446 is A2, the rotation speed of the classification rotor 444 is C2, and the probability P2 that the discharge / suction flow rate is V2 is output from the second node. From the nth node, the probability Pn that the rotation speed of the agitator 446 is An, the rotation speed of the classification rotor 444 is Cn, and the discharge / suction flow rate is Vn may be output. The number of nodes constituting the output layer 223 and the calculation result assigned to each node are not limited to the above examples, and can be appropriately designed.

The control unit 101 of the control device 100 has the highest probability of being output from the output layer 223 (in this embodiment, the rotation speed of the agitator 446, the rotation speed of the classification rotor 444, and the discharge / suction flow rate). The control parameters used for control are determined by specifying the combination of). The control unit 101 generates a control command for controlling the operation of the stirring motor 446M, the classification motor 415M, and the blower 7 based on the determined control parameters, and outputs the control command to each device through the output unit 104.

As described above, in the fourth embodiment, regarding the powder processing apparatus 4D, the rotation speed of the agitator 446, the rotation speed of the classification rotor 444, and the powder from the powder outlet 445 are obtained according to the input of the particle size desired by the user. A learning model 220 that outputs calculation results related to control parameters for controlling the discharge / suction flow rate when taking out the body is generated. Further, by using the learning model 220, calculation results regarding control parameters for controlling the rotation speed of the agitator 446, the rotation speed of the classification rotor 444, and the discharge / suction flow rate from the casing 440 can be obtained. The control device 100 can control the operation of the powder processing system 1 based on the calculation result of the learning model 220 so that the powder having a desired particle size can be obtained.

(Embodiment 5)
In the first embodiment, the powder processing apparatus 4A for crushing the powder raw material by applying mechanical energy such as impact, compression, grinding, and shearing to the powder raw material introduced into the casing 410 has been described. In place of the powder processing apparatus 4A of the above, the present invention is applied to an airflow type powder processing apparatus (for example, Micron Jet (registered trademark) T type manufactured by Hosokawa Micron Co., Ltd.) that performs pulverization treatment by the action of a jet airflow introduced into a casing. May be applied.
In the fifth embodiment, an example of application to an airflow type powder processing apparatus will be described. The powder processing system 1 according to the fifth embodiment includes a flow-type powder processing device 4E (see FIG. 14) instead of the mechanical powder processing device 4A, but other configurations are implemented. Since it is the same as the first form of the above, the detailed description thereof will be omitted.

FIG. 14 is a schematic cross-sectional view showing the configuration of the powder processing apparatus 4E according to the fifth embodiment. The powder processing apparatus 4E includes a cylindrical casing 450 that performs powder processing inside the powder processing apparatus 4E. The casing 450 is provided with a raw material input port 451, an ejector gas introduction port 452, a crushed gas introduction port 453, a collision plate 454, a classification rotor 455, a powder outlet 456, and the like.

As the material of the casing 450, it is preferable that the same material as the casing of the powder processing apparatus 4A described in the first embodiment is used. That is, iron-based steel materials such as SS400, S25C, S45C, and SPHC, stainless steel materials such as SUS304 and SUS316, iron casting materials such as FC20 and FC40, metals such as stainless casting materials such as SCS13 and FC14, ceramics, glass and the like. Should be used. Further, aluminum, other wood, or synthetic resin may be used as long as an abrasion resistant material is attached to the inner wall surface.

Further, on the inner surface of the casing 450, in order to improve the durability of the device, plating treatment such as hard chrome plating treatment, abrasion resistant thermal spraying material treatment such as tungsten carbide spraying, metal vapor deposition performed under vacuum, carbon of diamond structure Abrasion resistant treatment such as thermal spraying may be applied.

The casing 450 is provided with a raw material input port 451 for charging the powder raw material supplied from the raw material supply machine 2 into the casing 450. The raw material input port 451 is provided above, for example, the ejector gas introduction port 452. The powder raw material supplied from the raw material supply machine 2 is conveyed by, for example, a screw feeder (not shown) in the raw material supply path TP1 and is charged into the casing 450 from the raw material input port 451.

In the powder processing apparatus 4E according to the fifth embodiment, a gas such as compressed air is introduced into the casing 450 through the ejector gas introduction port 452 and the crushed gas introduction port 453, thereby causing a jet stream inside the casing 450. generate. The powder raw material charged from the raw material charging port 451 is suction-accelerated by the jet stream in the casing 450 and is crushed by colliding with the collision plate 454. Alternatively, the powders sucked and accelerated by the jet stream collide with each other to be crushed. Inside the casing 450, a pressure sensor (not shown) for measuring the crushing pressure at a constant or regular timing (for example, at 5-second intervals) is provided.

The powder crushed by the action of the jet stream is conveyed together with the stream and classified by the classification rotor 455 provided on the upper part of the device. A cylindrical guide ring 457 may be provided around the collision plate 454 to guide the crushed powder to the classification rotor 455.

The configuration of the classification rotor 455 is the same as that of the first embodiment. That is, the classification rotor 455 is a rotor provided with a plurality of classification blades 455A arranged radially, and is configured to rotate at a desired rotation speed by the power of the classification motor 415M. The rotation speed of the classification motor is measured by the rotation speed sensor S6 (see FIG. 3) at regular or periodic timings (for example, at 5-second intervals). The classification rotor 455 passes only powder having a particle size smaller than a predetermined particle size among the powder processed in the casing 450 by centrifugal force due to high-speed rotation, and guides only the passed powder to the powder outlet 456.

Here, the particle size of the powder passing through the classification rotor 455 can be set by controlling the rotation speed of the classification rotor 455 and the like. That is, by controlling the rotation speed of the classification rotor 455, powder having a particle size smaller than a predetermined particle size can be taken out from the casing 450. On the other hand, the powder that cannot pass through the classification rotor 455 circulates in the casing 450 and is repeatedly processed.

The control device 100 is a pressure sensor (non-standard) provided in the crushed gas introduction port 453 in order to generate a learning model 230 that outputs a calculation result regarding the control parameters of the powder processing device 4E in response to the input of the desired particle size. The crushing pressure measured by (shown), the rotation speed of the classification rotor 455 measured by the rotation speed sensor S6, the discharge / suction flow rate measured by the flow rate sensor S3, and the particle size data measured by the particle size sensor S2 are trained. Collect as data.

Using the collected teacher data, the control unit 101 of the control device 100 uses the collected teacher data to input the particle size desired by the user, the crushing pressure in the casing 450, the rotation speed of the classification rotor 455, and the powder from the powder outlet 456. A learning model 230 (see FIG. 15) that outputs calculation results related to control parameters that control the discharge / suction flow rate when the body is taken out is generated. Since the procedure for generating the learning model 230 is the same as that in the first embodiment, the description thereof will be omitted. The learning model 230 may be generated by the control device 100 or may be generated by an external server device (not shown).

FIG. 15 is a schematic diagram showing a configuration example of the learning model 230 according to the fifth embodiment. Similar to the learning model 200 described in the first embodiment, the learning model 230 includes an input layer 231 having one or a plurality of nodes,

intermediate layers

232A and 232B, and an output layer 233, respectively. The number of intermediate layers is not limited to two, and may be three or more.

In the learning model 230 according to the fifth embodiment, the crushing pressure in the casing 450, the rotation speed of the classification rotor 455, and the discharge / suction from the casing 450 are obtained in response to the input of the particle diameter (desired particle diameter) desired by the user. It is configured to output the calculation result related to the control parameters that control the flow rate.

When the control device 100 performs an calculation using the learning model 230, the control device 100 inputs the particle diameter (desired particle diameter) desired by the user into the learning model 230. The particle size data input to the learning model 230 is output to the node included in the first intermediate layer 232A through the nodes constituting the input layer 231. The data input to the first intermediate layer 232A is output to the node included in the next intermediate layer 232B through the nodes constituting the intermediate layer 232A. At this time, the output is calculated using the activation function including the weights and biases set between the nodes. Hereinafter, in the same manner, the calculation using the activation function including the weight and the bias set between the nodes is executed, and the calculation is transmitted to the subsequent layers one after another until the calculation result by the output layer 233 is obtained.

The output layer 233 outputs calculation results related to control parameters that control the crushing pressure of the crushing gas introduction port 453, the rotation speed of the classification rotor 455, and the discharge / suction flow rate when the powder is taken out from the casing 450. As the calculation result, for example, the probability indicating the quality of the combination of the plurality of control parameters described above may be output. Specifically, the output layer 233 is composed of n nodes from the first node to the nth node, and from the first node, the crushing pressure of the crushing gas introduction port 453 is B1, and the rotation speed of the classification rotor 455 is C1. , The probability P1 that the discharge / suction flow rate is V1 is output, and the probability P2 that the crushing pressure of the crushing gas introduction port 453 is B2, the rotation speed of the classification rotor 455 is C2, and the discharge / suction flow rate is V2 from the second node. , ..., The probability Pn that the crushing pressure of the crushing gas introduction port 453 is Bn, the rotation speed of the classification rotor 455 is Cn, and the discharge / suction flow rate is Vn may be output from the nth node. The number of nodes constituting the output layer 233 and the calculation result assigned to each node are not limited to the above examples, and can be appropriately designed.

The control unit 101 of the control device 100 has the highest probability of being output from the output layer 233 (in the present embodiment, the crushing pressure of the crushing gas introduction port 453, the rotation speed of the classification rotor 455, and the discharge). -By specifying the combination of suction flow rates), the control parameters used for control are determined. Based on the determined control parameters, the control unit 101 generates control commands for controlling the crushing pressure of compressed air or the like that generates a jet stream in the casing 450, the operation of the classification motor 415M, and the blower 7, and each of them is generated through the output unit 104. Output to the device.

As described above, in the fifth embodiment, regarding the air flow type powder processing apparatus 4E, the crushing pressure of the crushing gas introduction port 453, the rotation speed of the classification rotor 455, and the powder are obtained according to the input of the particle size desired by the user. A learning model 230 is generated that outputs calculation results related to control parameters for controlling the discharge / suction flow rate when the powder is taken out from the body take-out port 456. Further, by using the learning model 230, calculation results regarding the crushing pressure of the crushing gas introduction port 453, the rotation speed of the classification rotor 455, and the control parameters for controlling the discharge / suction flow rate from the casing 450 can be obtained. The control device 100 can control the operation of the powder processing system 1 based on the calculation result of the learning model 230 so that the powder having a desired particle size can be obtained.

(Embodiment 6)
In the sixth embodiment, regarding the powder processing apparatus 4A, the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, the discharge / suction flow rate from the casing 410, and the raw material supply machine with respect to the input of the particle size desired by the user. The configuration of the learning model 240 that outputs the calculation results regarding the control parameters that control the supply amount of the powder raw material and the temperature in the casing 410 according to No. 2 will be described.
Since the overall configuration of the powder processing system 1 and the configuration of each device in the powder processing system 1 are the same as those in the first embodiment, the description thereof will be omitted.

FIG. 16 is a schematic diagram showing a configuration example of the learning model 240 in the sixth embodiment. Similar to the learning model 200 described in the first embodiment, the learning model 240 includes an input layer 241 having one or a plurality of nodes,

intermediate layers

242A and 242B, and an output layer 243, respectively. The number of intermediate layers is not limited to two, and may be three or more.

In the learning model 240 according to the sixth embodiment, the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410 are obtained in response to the input of the particle diameter (desired particle diameter) desired by the user. , The calculation result regarding the control parameter for controlling the supply amount of the powder raw material by the raw material supply machine 2 and the temperature in the casing 410 is output. The supply amount of the powder raw material may be the supply amount per unit time (that is, the supply rate).

In such a learning model 240, the particle size data (particle size data measured by the particle size sensor S2) regarding the particle size obtained from the powder processing device 4A and the rotation of the crushing rotor 413 included in the powder processing device 4A Includes speed, rotational speed of classifying rotor 415, discharge / suction flow rate when taking out powder from casing 410, supply amount of powder raw material obtained by weight sensor S1, and temperature in casing 410 obtained by temperature sensor S4. It is generated by learning the relationship between the particle size and the control parameters using the measured data as the teacher data. The supply amount of the powder raw material may be the supply amount per unit time (that is, the supply rate). Further, only one of the supply amount of the powder raw material and the temperature inside the casing 410 may be adopted as the control parameter. The learning model 240 may be generated by the control device 100 or by an external server device (not shown).

When the control device 100 performs an operation using the learning model 240, the control device 100 inputs the particle diameter (desired particle diameter) desired by the user into the learning model 240. The particle size data input to the learning model 240 is output to the node included in the first intermediate layer 242A through the nodes constituting the input layer 241. The data input to the first intermediate layer 242A is output to the node included in the next intermediate layer 242B through the nodes constituting the intermediate layer 242A. At this time, the output is calculated using the activation function including the weights and biases set between the nodes. Hereinafter, in the same manner, the calculation using the activation function including the weight and the bias set between the nodes is executed, and the calculation is transmitted to the subsequent layers one after another until the calculation result by the output layer 243 is obtained.

The output layer 243 controls the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, the discharge / suction flow rate when taking out powder from the casing 410, the supply amount of the powder raw material, and the temperature inside the casing 410. Outputs the calculation result related to the parameter. As the calculation result, for example, the probability indicating the quality of the combination of the plurality of control parameters described above may be output. Specifically, the output layer 243 is composed of n nodes from the first node to the nth node, and the rotation speed of the crushing rotor 413 is G1, the rotation speed of the classification rotor 415 is C1, and the discharge is performed from the first node. The probability P1 that the suction flow rate is V1, the supply amount of the powder raw material is F1, and the temperature inside the casing 410 is T1 is output, and the rotation speed of the crushing rotor 413 is G2 and the rotation speed of the classification rotor 415 is from the second node. Is C2, the discharge / suction flow rate is V2, the supply amount of the powder raw material is F2, the probability P2 that the temperature inside the casing 410 is T2 is output, and ..., The rotation speed of the crushing rotor 413 is Gn from the nth node. You may output the probability Pn that the rotation speed of the classification rotor 415 is Cn, the discharge / suction flow rate is Vn, the supply amount of the powder raw material is Fn, and the temperature in the casing 410 is Tn. The number of nodes constituting the output layer 243 and the calculation result assigned to each node are not limited to the above examples, and can be appropriately designed.

The control unit 101 of the control device 100 has the highest probability of being output from the output layer 243 (in the present embodiment, the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate). , The supply amount of the powder raw material, and the combination of the temperature in the casing 410), the control parameters used for the control are determined. Based on the determined control parameters, the control unit 101 generates control commands for controlling the operations of the crushing motor 413M, the classification motor 415M, the blower 7, the raw material supply machine 2, and the hot air generator 3, and each of them is generated through the output unit 104. Output to the device.

As described above, when the learning model 240 according to the sixth embodiment is used, in addition to the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410, the supply amount of the powder raw material. And the calculation result regarding the control parameter for controlling the temperature in the casing 410 is obtained. The control device 100 can control the operation of the powder processing system 1 based on the calculation result of the learning model 240 so that the powder having a desired particle size can be obtained.

In the sixth embodiment, an application example to the mechanical powder processing apparatus 4A has been described, but it goes without saying that the same method can be applied to the powder processing apparatus 4B to 4E.

(Embodiment 7)
In the seventh embodiment, regarding the powder processing apparatus 4A, the particle size desired by the user, the circularity of the powder, the water content of the powder raw material, the driving power or driving current of the powder processing apparatus 4A, and the heat medium temperature. , The learning model 250 that outputs the calculation result with respect to the input of at least one of the medium weight, the environmental temperature, and the environmental humidity will be described.
Since the overall configuration of the powder processing system 1 and the configuration of each device in the powder processing system 1 are the same as those in the first embodiment, the description thereof will be omitted.

FIG. 17 is a schematic diagram showing a configuration example of the learning model 250 according to the seventh embodiment. Similar to the learning model 200 described in the first embodiment, the learning model 250 includes an input layer 251 having one or a plurality of nodes,

intermediate layers

252A and 252B, and an output layer 253, respectively. The number of intermediate layers is not limited to two, and may be three or more.

In the learning model 250 according to the seventh embodiment, the particle size (desired particle size) desired by the user, the circularity of the powder, the water content of the powder raw material, the driving power or the driving current of the powder processing apparatus 4A, For input of at least one of the temperature of the fluid (heat medium temperature) to be introduced into the powder processing apparatus 4A, the weight of the medium (medium weight) in the medium stirring type powder processing apparatus, the environmental temperature, and the environmental humidity. It is configured to output calculation results related to control parameters that control the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410.

In such a learning model 250, the particle size data regarding the particle size obtained from the powder processing device 4A, the circularity of the powder, the water content of the powder raw material, the driving power or the driving current of the powder processing device 4A, Data on at least one of heat medium temperature, medium weight, environmental temperature, and environmental humidity, and the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the casing 410 included in the powder processing apparatus 4A. It is generated by learning the relationship between input and output using the measurement data including the discharge / suction flow rate at the time as the teacher data. Here, the circularity of the powder is a value represented by, for example, (perimeter of a circle having a projected area equal to the projected area of the particles) / (perimeter of the particles), and is collected using a known measuring device. can do. The water content of the powder raw material is a value represented by, for example,% by mass, and can be collected using a known measuring device such as an infrared moisture meter. The drive power and drive current of the powder processing device 4A may be obtained by acquiring values measured inside the powder processing device 4A. The heat medium temperature may be a value measured inside the powder processing apparatus 4A, and the medium weight may be a value measured in advance. The environmental temperature and humidity are the temperature and humidity of the ambient environment in which the powder processing apparatus 4A is set, and can be collected by using a known temperature sensor and humidity sensor. The learning model 250 may be generated by the control device 100 or by an external server device (not shown).

When the control device 100 performs the calculation using the learning model 250, the particle size (desired particle size) of the powder desired by the user, the circularity of the powder desired by the user, and the moisture content of the powder desired by the user At least one of the content, the driving power or driving current of the powder processing apparatus 4A set by the user, the heat medium temperature, the medium weight, the measured value of the environmental temperature, and the measured value of the environmental humidity is input to the learning model 250. The data input to the learning model 250 is output to the nodes included in the first intermediate layer 252A through the nodes constituting the input layer 251. The data input to the first intermediate layer 252A is output to the node included in the next intermediate layer 252B through the nodes constituting the intermediate layer 252A. At this time, the output is calculated using the activation function including the weights and biases set between the nodes. Hereinafter, in the same manner, the calculation using the activation function including the weight and the bias set between the nodes is executed, and the calculation is transmitted to the subsequent layers one after another until the calculation result by the output layer 253 is obtained.

The output layer 253 controls the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, the discharge / suction flow rate when powder is taken out from the casing 410, the supply amount of the powder raw material, and the temperature inside the casing 410. Outputs the calculation result related to the control parameter. As the calculation result, for example, the probability indicating the quality of the combination of the plurality of control parameters described above may be output. Specifically, the output layer 253 is composed of n nodes from the first node to the nth node, and the rotation speed of the crushing rotor 413 is G1, the rotation speed of the classification rotor 415 is C1, and the discharge is performed from the first node. The probability P1 that the suction flow rate is V1 is output, and the rotation speed P2 of the crushing rotor 413 is G2, the rotation speed of the classification rotor 415 is C2, and the discharge / suction flow rate is V2 is output from the second node. ..., The probability Pn that the rotation speed of the crushing rotor 413 is Gn, the rotation speed of the classification rotor 415 is Cn, and the discharge / suction flow rate is Vn may be output from the nth node. The number of nodes constituting the output layer 253 and the calculation result assigned to each node are not limited to the above examples, and can be appropriately designed.

The control unit 101 of the control device 100 has the highest probability of being output from the output layer 253 (in this embodiment, the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction). By specifying the combination of flow rates), the control parameters used for control are determined. Based on the determined control parameters, the control unit 101 generates control commands for controlling the operations of the crushing motor 413M, the classification motor 415M, and the blower 7, and outputs the control commands to each device through the output unit 104.

As described above, when the learning model 250 according to the seventh embodiment is used, the particle size of the powder desired by the user, the circularity of the powder desired by the user, the water content of the powder desired by the user, and the like. The powder can be generated on condition of the drive power or drive current of the powder processing apparatus 4A set by the user.

In the learning model 250 according to the seventh embodiment, the particle size desired by the user, the circularity of the powder, the water content of the powder raw material, the driving power or driving current of the powder processing apparatus 4A, and the heat medium temperature , The composition is such that the calculation result is output for the input of at least one of the medium weight, the environmental temperature, and the environmental humidity, but the dust content concentration, the powder composition, the powder moisture, the pressure, the sound pressure, or the like. At least one of the frequencies may be further input.

Further, the learning model 250 according to the seventh embodiment is configured to output calculation results regarding control parameters for controlling the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410. However, in addition to these control parameters, the configuration may be such that the calculation result regarding the control parameter that controls at least one of the supply amount of the powder raw material and the temperature in the casing 410 is output.

Further, when at least one of the control parameters input to the learning model 250 is changed in time series, the learning model 250 may adopt a model constructed by the recurrent neural network.

In the seventh embodiment, an application example to the mechanical powder processing apparatus 4A has been described, but it goes without saying that the same method can be applied to the powder processing apparatus 4B to 4E.

(Embodiment 8)
In the eighth embodiment, the configuration for generating the learning model 200 according to the type of the powder raw material will be described with respect to the powder processing apparatus 4A.
Since the overall configuration of the powder processing system 1 and the configuration of each device in the powder processing system 1 are the same as those in the first embodiment, the description thereof will be omitted.

The control device 100 according to the eighth embodiment has the powder particle size data obtained from the powder processing device 4A, the rotation speed of the crushing rotor 413 included in the powder processing device 4A, and the classification for each type of powder raw material. The rotation speed of the rotor 415 and the measurement data including the discharge / suction flow rate when the powder is taken out from the casing 410 are collected. FIG. 18 is a conceptual diagram showing an example of collecting data in the eighth embodiment. The control unit 101 of the control device 100 acquires measurement data from the particle size sensor S2, the rotation speed sensor S5 of the crushing rotor 413, the rotation speed sensor S6 of the classification rotor 415, and the flow rate sensor S3 through the input unit 103. The data is stored in the storage unit 102 together with the time stamp and information on the type of the powder raw material. Here, the information regarding the type of the powder raw material may be character information indicating the type name of the powder raw material, or may be an arbitrary identifier that can specify the type of the powder raw material. Information on the type of powder raw material may be received through the operation unit 106 before data collection or after data collection. The example of FIG. 18 shows a state in which data is collected for each type of powder raw material such as toner raw material, graphite, and manganese dioxide (MnO ₂ ) and stored in the storage unit 102.

The control unit 101 uses the data collected for each type of powder raw material as the teacher data to generate the learning model 200.

FIG. 19 is a flowchart illustrating a procedure for generating the learning model 200 according to the eighth embodiment. The control unit 101 of the control device 100 receives information regarding the type of the powder raw material prior to the generation of the learning model 200 (step S401). For example, the control unit 101 displays a screen for inquiring the user about the type of the powder raw material on the display unit 107, and can receive information on the type of the powder raw material through the displayed screen. Further, the control unit 101 may receive information on the type of the powder raw material by transmitting an inquiry about the type of the powder raw material from the communication unit 105 to the terminal device 500 and receiving a reply from the terminal device 500. ..

Next, the control unit 101 reads the data stored in association with the type received in step S401 from the storage unit 102 (step S402). The control unit 101 selects the data to be used for the teacher data from the read data (step S403). That is, the control unit 101 uses the read data to measure the rotational speed of the crushing rotor 413, the rotational speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410, and the powder obtained at that time. Only one set of particle size values (D10, D50, D90) is selected.

Next, the control unit 101 inputs the value of the particle diameter included in the selected teacher data into the learning model 200 (step S404), and executes the calculation by the learning model 200 (step S405). That is, the control unit 101 inputs the value of the particle diameter to the nodes constituting the input layer 201 of the learning model 200, performs the calculation using the weights and biases between the nodes in the

intermediate layers

Next, the control unit 101 evaluates the calculation result obtained in step S405 (step S406), and determines whether or not the learning is completed (step S407). Specifically, the control unit 101 can evaluate the calculation result by using an error function (also referred to as an objective function, a loss function, or a cost function) based on the calculation result obtained in step S405 and the teacher data. .. The control unit 101 is in the process of optimizing (minimizing or maximizing) the error function by a gradient descent method such as the steepest descent method, and when the error function is below the threshold value (or above the threshold value), the learning is completed. to decide. In order to avoid the problem of overfitting, techniques such as cross-validation and early stopping may be adopted to end learning at an appropriate timing.

When it is determined that the learning is not completed (S407: NO), the control unit 101 updates the weights and biases between the nodes of the learning model 200 (step S408), returns the process to step S403, and another teacher. Continue learning with the data. The control unit 101 may update the weights and biases between the nodes by using the error back propagation method in which the weights and biases between the nodes are sequentially updated from the output layer 203 to the input layer 201 of the learning model 200. it can.

When it is determined that the learning is completed (S407: YES), the control unit 101 stores the learned learning model 200 in the storage unit 102 in association with the information regarding the type of the powder raw material (step S409), and this flowchart. Ends the processing by.

As described above, the control device 100 according to the eighth embodiment can generate the learning model 200 according to the type of the powder raw material. The learning model 200 in the eighth embodiment is the same as that in the first embodiment, but instead of the learning model 200, the learning models 210 to 250 described in the second to seventh embodiments are generated for each type of powder raw material. You may.

FIG. 20 is a flowchart illustrating a control procedure by the control device 100. The control unit 101 of the control device 100 receives information regarding the type of the powder raw material and input of the particle size desired by the user through the operation unit 106 (step S421).

Next, the control unit 101 reads out the learning model 200 according to the type of the received powder raw material from the storage unit 102 (step S422). The control unit 101 inputs the received particle size data to the input layer 201 of the learning model 200 read out in step S422, and executes the calculation by the learning model 200 (step S423). At this time, the control unit 101 gives the received particle size data to the node of the input layer 201, and executes the calculation by the

intermediate layers

202A and 202B. The output layer 203 of the learning model 200 outputs calculation results related to control parameters for controlling the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410.

Next, the control unit 101 acquires the calculation result from the learning model 200 (step S424) and determines the control parameters used for control (step S425). The control unit 101 may determine a combination of the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate based on the probability of being output as the calculation result.

Next, the control unit 101 executes control based on the control parameters determined in step S425 (step S426). That is, the control unit 101 generates control commands for the crushing motor 413M and the classification motor 415M so that the rotation speeds of the crushing rotor 413 and the classification rotor 415 become the values determined in step S425, and powder processing is performed through the output unit 104. Output to device 4A. Further, the control unit 101 generates a control command for the blower 7 so that the discharge / suction flow rate from the casing 410 becomes the value determined in step S425, and outputs the control command to the blower 7 through the output unit 104.

As described above, the control device 100 according to the present embodiment can determine the control parameters by using the learning model 200 according to the type of the powder raw material. The control device 100 can perform control based on the determined control parameters so that a powder having a desired particle size can be obtained.

In the present embodiment, the control parameters are determined using the learning model 200 generated for each type of powder raw material. However, the learning models 210 to 250 described in the second to seventh embodiments are used as powder. The control parameters may be determined by generating for each type of body material and using the generated learning models 210 to 250.

In the eighth embodiment, an application example to the mechanical powder processing apparatus 4A has been described, but it goes without saying that the same method can be applied to the powder processing apparatus 4B to 4E.

(Embodiment 9)
In the ninth embodiment, the re-learning procedure of the learning model 200 will be described.
Since the overall configuration of the powder processing system 1 and the configuration of each device in the powder processing system 1 are the same as those in the first embodiment, the description thereof will be omitted.

FIG. 21 is a flowchart illustrating a re-learning procedure of the learning model 200. As described in the first embodiment, in the operation phase, the control unit 101 of the control device 100 accepts the input of the particle diameter (desired particle diameter) desired by the user, and transfers the received particle diameter data to the learning model 200. By inputting, the calculation result related to the control parameter is acquired. The control unit 101 controls the operation of each device constituting the powder processing system 1 based on the calculation result acquired from the learning model 200. In the operation phase, the control unit 101 includes measurement data including the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410, and the powder particles obtained from the powder processing apparatus 4A. Diameter data and may be collected. The collected measurement data and particle size data are stored in the storage unit 102 together with the time stamp.

The control unit 101 compares the particle size indicated by the particle size data (that is, the actually measured value of the particle size) with the desired particle size of the user at an appropriate timing after the start of the operation phase (step S501). When the input to the learning model 200 is only the value of D50, the control unit 101 may compare the value of D50 obtained as the actually measured value with the value of D50 input as the desired particle size. When the input to the learning model 200 is the value of D10, D50, D90 (or the upper limit value and the lower limit value thereof), the control unit 101 sets the value of D10, D50, D90 obtained as the measured value. , The values of D10, D50, and D90 (or their upper and lower limit values) input as the desired particle size may be compared. The control unit 101 determines whether or not to execute re-learning based on the comparison result (step S502).

When the measured value of the particle size is close to the desired particle size (for example, when the difference between the two is less than 10%), the control unit 101 determines that re-learning is not executed (S502: NO), and processes according to this flowchart. To finish.

On the other hand, when the measured value of the particle size is not close to the desired particle size (for example, when the difference between the two is 10% or more), the control unit 101 determines that the re-learning is executed (S502: YES).

When it is determined to execute re-learning, the control unit 101 reads the data collected after the start of operation from the storage unit 102 (step S503) and selects the teacher data (step S504). The re-learning procedure after step S503 may be executed at a timing when the powder processing system 1 is not operating.

The control unit 101 inputs the value of the particle diameter included in the selected teacher data into the learning model 200 (step S505), and executes the calculation by the learning model 200 (step S506). That is, the control unit 101 inputs the value of the particle diameter to the nodes constituting the input layer 201 of the learning model 200, performs the calculation using the weights and biases between the nodes in the

intermediate layers

202A and 202B, and outputs the calculation result. The process of outputting from the node of layer 203 is performed.

Next, the control unit 101 evaluates the calculation result obtained by the calculation in step S506 (step S507), and determines whether or not the learning is completed (step S508). Specifically, the control unit 101 can evaluate the calculation result by using an error function (also referred to as an objective function, a loss function, or a cost function) based on the calculation result obtained in step S506 and the teacher data.

When it is determined that the learning is not completed (S508: NO), the control unit 101 updates the weights and biases between the nodes of the learning model 200 (step S509), returns the process to step S504, and another teacher. Continue learning with the data. The control unit 101 may update the weights and biases between the nodes by using the error back propagation method in which the weights and biases between the nodes are sequentially updated from the output layer 203 to the input layer 201 of the learning model 200. it can.

When it is determined that the learning is completed (S508: YES), the control unit 101 stores the learned learning model 200 in the storage unit 102 (step S510), and ends the process according to this flowchart.

As described above, since the control device 100 according to the present embodiment relearns the learning model 200 as necessary, the accuracy of powder processing by the powder processing system 1 even after the start of the operation phase. Can be enhanced.

Although the re-learning procedure of the learning model 200 has been described in the ninth embodiment, the learning models 210 to 250 described in the second to seventh embodiments and the learning for each type of the powder raw material described in the eighth embodiment have been described. The model 200 can also be relearned by the same procedure.

In the ninth embodiment, an application example to the mechanical powder processing apparatus 4A has been described, but it goes without saying that the same method can be applied to the powder processing apparatus 4B to 4E.

(Embodiment 10)
In the tenth embodiment, the procedure for adjusting the control parameters in the powder processing apparatus 4A will be described.
Since the overall configuration of the powder processing system 1 and the configuration of each device in the powder processing system 1 are the same as those in the first embodiment, the description thereof will be omitted.

FIG. 22 is a flowchart illustrating a control parameter adjustment procedure. As described in the first embodiment, in the operation phase, the control unit 101 of the control device 100 accepts the input of the particle diameter (desired particle diameter) desired by the user, and transfers the received particle diameter data to the learning model 200. By inputting, the calculation result related to the control parameter is acquired. The control unit 101 controls the operation of each device constituting the powder processing system 1 based on the calculation result acquired from the learning model 200. In the operation phase, the control unit 101 includes measurement data including the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the discharge / suction flow rate from the casing 410, and the powder particles obtained from the powder processing apparatus 4A. Diameter data and may be collected. The collected measurement data and particle size data are stored in the storage unit 102 together with the time stamp.

The control unit 101 compares the particle diameter indicated by the particle diameter data (that is, the actually measured value of the particle diameter) with the desired particle diameter of the user at an appropriate timing after the start of the operation phase (step S601). When the input to the learning model 200 is only the value of D50, the control unit 101 may compare the value of D50 obtained as the actually measured value with the value of D50 input as the desired particle size. When the input to the learning model 200 is the value of D10, D50, D90 (or the upper limit value and the lower limit value thereof), the control unit 101 sets the value of D10, D50, D90 obtained as the measured value. , The values of D10, D50, and D90 (or their upper and lower limit values) input as the desired particle size may be compared. The control unit 101 determines whether or not to adjust the control parameter based on the comparison result (step S602).

When the measured value of the particle size is close to the desired particle size (for example, when the difference between the two is less than 10%), the control unit 101 determines that the control parameter is not adjusted (S602: NO), and processes according to this flowchart. To finish.

On the other hand, when the measured value of the particle size is not close to the desired particle size (for example, when the difference between the two is 10% or more), the control unit 101 determines that the control parameter is adjusted (S602: YES).

When it is determined that the control parameter is to be adjusted, the control unit 101 adjusts the control parameter according to the comparison result in step S602 (step S603), and operates the device including the powder processing device 4A based on the adjusted control parameter. Control (step S604). For example, when the measured value of the particle size is smaller than the desired particle size, the control unit 101 makes the rotation speeds of the crushing rotor 413 and the classification rotor 415 lower than the current values, and the discharge / suction flow rate is the current value. The control parameters are adjusted so as to be larger, and the operation of the device including the powder processing device 4A is controlled based on the adjusted control parameters. When the measured value of the particle size is larger than the desired particle size, the control unit 101 sets the rotation speeds of the crushing rotor 413 and the classification rotor 415 to be higher than the current values, and the discharge / suction flow rate is the current value. The control parameters are adjusted so as to be smaller, and the operation of the device including the powder processing device 4A is controlled based on the adjusted control parameters.

As described above, in the present embodiment, when a deviation of a certain amount or more occurs between the measured value of the particle size and the particle size desired by the user during the operation of the powder processing system 1, the deviation is reduced. Control parameters can be adjusted to.

In the tenth embodiment, an application example to the mechanical powder processing apparatus 4A has been described, but it goes without saying that the same method can be applied to the powder processing apparatus 4B to 4E.

(Embodiment 11)
In the eleventh embodiment, a configuration for controlling the operation of the device including the powder processing device 4A from the terminal device 500 will be described.
Since the overall configuration of the powder processing system 1 and the configuration of each device in the powder processing system 1 are the same as those in the first embodiment, the description thereof will be omitted.

FIG. 23 is a flowchart illustrating a procedure of processing executed by the terminal device 500 and the control device 100. The control unit 501 of the terminal device 500 accesses the control device 100 through the communication unit 503 (step S701).

When the control unit 101 of the control device 100 receives access from the terminal device 500 when the powder processing device 4A is not operating, for example, a screen related to an input screen for receiving an input of a particle size desired by the user. Data is generated (step S702), and the generated screen data is transmitted from the communication unit 105 to the terminal device 500 (step S703). When an access is received from the terminal device 500 while the powder processing device 4A is operating, the control unit 101 may execute the processes after step S703 described below.

The control unit 501 of the terminal device 500 receives the screen data of the input screen transmitted from the control device 100 from the communication unit 503 (step S704). The control unit 501 displays the input screen on the display unit 505 based on the received screen data (step S705), and receives the input of the particle size desired by the user (step S706).

FIG. 24 is a schematic diagram showing an example of an input screen. This input screen shows a screen for accepting a lower limit value and an upper limit value for each of D10, D50, and D90. Instead of the configuration in which the lower and upper limits of D10, D50, and D90 are accepted, the values of D10, D50, and D90 desired by the user may be accepted, or only the value of D50 (median diameter) may be accepted. .. The control unit 501 transmits the particle size data received in step S706 to the control device 100 (step S707).

The control unit 101 of the control device 100 receives the particle size data transmitted from the terminal device 500 from the communication unit 105 (step S708). The control unit 101 inputs the received particle size data into, for example, the learning model 200, and executes the calculation by the learning model 200 (step S709). Next, the control unit 101 acquires the calculation result from the learning model 200 (step S710), and determines the control parameter based on the calculation result (step S711). The control unit 101 controls the operation of the device including the powder processing device 4A based on the determined control parameter (step S712).

During the operation of the powder processing system 1, the control unit 101 determines the particle size of the powder obtained from the powder processing device 4A, the rotation speeds of the crushing rotor 413 and the classification rotor 415, and the discharge / suction flow rate from the casing 410. The measurement data of is acquired from the input unit 103 at any time. The control unit 101 generates screen data of the monitoring screen including at least one of the above measurement data and an overall view of the powder processing system 1 at an appropriate timing (step S713). The control unit 101 transmits the generated screen data from the communication unit 105 to the terminal device 500 (step S714).

The control unit 501 of the terminal device 500 receives the screen data transmitted from the control device 100 from the communication unit 503 (step S715). The control unit 501 displays the monitoring screen on the display unit 505 based on the received screen data (step S716). FIG. 25 is a schematic view showing an example of the monitoring screen. In the example shown in FIG. 25, the measurement data of the particle size obtained from the powder processing device 4A (that is, the actually measured value of the particle size measured by the particle size sensor S2) and the overall view of the powder processing system 1 are shown. The state in which the including monitoring screen is displayed on the display unit 505 is shown. Instead of the particle size measurement data, or together with the particle size measurement data, the rotation speed of the crushing rotor 413, the rotation speed of the classification rotor 415, and the measurement data of the discharge / suction flow rate from the casing 410 are displayed on the monitoring screen. You may. Further, the monitoring screen monitors the desired particle size input by the user, the rotation speed of the crushing rotor 413 set based on the calculation result of the learning model 200, the rotation speed of the classification rotor 415, the value of the discharge / suction flow rate from the casing 410, and the like. It may be displayed in.

As described above, in the present embodiment, the powder processing system 1 can be remotely controlled by using the terminal device 500, and the operating status of the powder processing system 1 can be monitored by the terminal device 500.

In the eleventh embodiment, an application example to the mechanical powder processing apparatus 4A has been described, but it goes without saying that the same method can be applied to the powder processing apparatus 4B to 4E.

The embodiments disclosed this time should be considered to be exemplary in all respects and not restrictive. The scope of the present invention is indicated by the scope of claims, not the above-mentioned meaning, and is intended to include all modifications within the meaning and scope equivalent to the claims.

For example, the present invention is not limited to the powder processing devices 4A to 4E described in the first to eleventh embodiments, but the impact type crusher, the shear type crusher, the air flow type crusher, the grinding type crusher, and the medium stirring type. It can be applied to powder processing devices classified into crushers, screen mills, bead mills, ball mills, pin mills, and jet mills (that is, powder processing devices having at least a crushing function).

Further, in the first to eleventh embodiments, the gas is introduced into the powder processing devices 4A to 4E, but a liquid may be introduced instead of the gas. In this case, as the flow rate sensor S3, a sensor that measures the flow rate of the liquid may be used. Further, a pump may be used instead of the blower 7.

1 ... Powder processing system, 2 ... Raw material supply machine, 3 ... Hot air generator, 4A-4E ... Powder processing device, 5 ... Cyclone, 6 ... Dust collector, 7 ... Blower, 410 ... Casing, 411 ... Raw material input port, 412 ... Gas inlet, 413 ... Crushing rotor, 414 ... Guide ring, 415 ... Classification rotor, 416 ... Powder outlet, S1 ... Weight sensor, S2 ... Particle size sensor, S3 ... Flow sensor, S4 ... Temperature sensor, S5 ... Rotation speed sensor, S6 ... Rotation speed sensor, 100 ... Control device, 101 ... Control unit, 102 ... Storage unit, 103 ... Input unit, 104 ... Output unit, 105 ... Communication unit, 106 ... Operation unit, 107 ... Display unit , 500 ... Terminal device, 501 ... Control unit, 502 ... Storage unit, 503 ... Communication unit, 504 ... Operation unit, 505 ... Display unit

Claims

Using a computer
With respect to the powder processing apparatus having at least a function of crushing the powder raw material, measurement data indicating the operating state of the powder processing apparatus and particle size data of the powder obtained from the powder processing apparatus are acquired.
A learning model that uses the acquired measurement data and particle size data as training data and outputs calculation results for control parameters that control the operating state of the powder processing device in response to the input of the particle size desired by the user. How to generate a training model.
The powder processing apparatus includes a crushing rotor for crushing the powder raw material and a classification rotor for classifying the powder.
The measurement data includes the rotation speed of the crushing rotor, the rotation speed of the classification rotor, and the discharge / suction flow rate when the powder is taken out from the processing chamber of the powder processing apparatus.
A claim for generating a learning model that outputs calculation results for control parameters related to the rotation speed of the crushing rotor, the rotation speed of the classification rotor, and the discharge / suction flow rate from the processing chamber in response to the input of the particle size. The method for generating a learning model according to 1.
The powder processing apparatus includes a crushing rotor for crushing the powder raw material.
The measurement data includes the rotation speed of the crushing rotor and the discharge / suction flow rate when the powder is taken out from the processing chamber of the powder processing apparatus.
The generation of the learning model according to claim 1, which generates a learning model that outputs calculation results for control parameters related to the rotation speed of the crushing rotor and the discharge / suction flow rate from the processing chamber in response to the input of the particle size. Method.
The powder processing apparatus includes an agitator for stirring the medium and a classification rotor for classifying the powder.
The measurement data includes the rotation speed of the agitator, the rotation speed of the classification rotor, and the discharge / suction flow rate when the powder is taken out from the processing chamber of the powder processing apparatus.
The first aspect of claim 1 is to generate a learning model that outputs calculation results of control parameters related to the rotation speed of the agitator, the rotation speed of the classification rotor, and the discharge / suction flow rate from the processing chamber in response to the input of the particle size. How to generate the described training model.
The powder processing apparatus includes a classification rotor for classifying powder.
The measurement data includes the rotation speed of the classification rotor, the crushing pressure of the fluid introduced into the processing chamber of the powder processing apparatus, and the discharge / suction flow rate when the powder is taken out from the processing chamber.
Claim 1 to generate a learning model that outputs calculation results for control parameters related to the rotation speed of the classification rotor, the crushing pressure of the fluid, and the discharge / suction flow rate from the processing chamber in response to the input of the particle size. How to generate the training model described in.
Using the measurement data including at least one of the supply amount of the powder raw material and the temperature in the processing chamber to be supplied to the processing chamber of the powder processing apparatus as the teacher data, the supply amount and at least one of the temperatures are used. The method for generating a learning model according to any one of claims 1 to 5, wherein a learning model for outputting an operation result relating to a control parameter including the above is generated.
The teacher data includes the circularity of the powder, the water content of the powder raw material, the driving power or driving current of the powder processing apparatus, the heat medium temperature, the medium weight, the environmental temperature, and at least 1 of the environmental humidity. Including one more
For input of a particle size desired by the user and at least one of the circularity, the water content, the driving power or the driving current, the heat medium temperature, the medium weight, the environmental temperature, and the environmental humidity. The method for generating a learning model according to any one of claims 1 to 6, wherein a learning model that outputs a calculation result related to the control parameter is generated.
Re-acquire the measurement data and particle size data used for the teacher data,
The method for generating a learning model according to any one of claims 1 to 7, wherein the re-acquired measurement data and the particle size data are used as teacher data to relearn the learning model.
The method according to any one of claims 1 to 8, wherein a learning model that outputs a calculation result related to the control parameter according to the input of the particle size desired by the user is generated according to the type of the powder raw material. How to generate a learning model for.
The method for generating a learning model according to any one of claims 1 to 9, wherein the fluid flowing in the system including the powder processing apparatus is a gas or a liquid.
On the computer
With respect to the powder processing apparatus having at least a function of crushing the powder raw material, measurement data indicating the operating state of the powder processing apparatus and particle size data of the powder obtained from the powder processing apparatus are acquired.
A learning model that uses the acquired measurement data and particle size data as training data and outputs calculation results for control parameters that control the operating state of the powder processing device in response to the input of the particle size desired by the user. A computer program to perform the process of generating.
An input layer in which the particle size desired by the user is input,
An output layer that outputs calculation results for control parameters that control the operating state of the powder processing apparatus, measurement data indicating the operating state of the powder processing apparatus, and a particle size of powder obtained from the powder processing apparatus. It is provided with an intermediate layer in which the relationship between the particle size and the control parameter is learned by using the data as the teacher data.
When the particle size desired by the user is input to the input layer, the intermediate layer calculates and makes the computer function to output the calculation result for the control parameter that controls the operating state of the powder processing apparatus. model.
A reception unit that accepts input of the particle size desired by the user,
Using the particle size data of the powder obtained from the powder processing device and the measurement data indicating the operating state of the powder processing device as the teacher data, the particle size of the powder and the operation of the powder processing device are used. A learning model that has learned the relationship between the control parameters that control the state,
An arithmetic processing unit that inputs the particle size data received by the reception unit into the learning model and executes an operation by the learning model.
A control device including a control unit that controls the operation of the device including the powder processing device based on the calculation result of the learning model.
The arithmetic processing unit includes the rotation speed of the crushing rotor for crushing the powder raw material, the rotation speed of the classification rotor for classifying the powder, the discharge / suction flow rate when taking out the powder from the processing chamber of the powder processing apparatus, and the medium. Particles received by the reception unit in a learning model in which the relationship has been learned using at least one of the rotation speed of the agitator for stirring and the crushing pressure of the fluid introduced into the processing chamber of the powder processing apparatus. By inputting the diameter, the calculation by the learning model is executed,
The control device according to claim 13, wherein the control unit controls the operation of the device including the powder processing device based on the calculation result of the learning model.
The arithmetic processing unit has learned the relationship by using measurement data including at least one of the supply amount of the powder raw material to be supplied to the processing chamber of the powder processing apparatus and the temperature in the processing chamber. By inputting the particle size received by the reception unit into the model, the calculation by the learning model is executed.
The control device according to claim 13 or 14, wherein the control unit controls the operation of the device including the powder processing device based on the calculation result of the learning model.
The receiving unit includes the particle size desired by the user, the circularity of the powder, the water content of the powder raw material, the driving power or driving current of the powder processing apparatus, the heat medium temperature, the medium weight, and the environmental temperature. And accept at least one of the environmental humidity,
The arithmetic processing unit further adds at least one of the particle size, the circularity, the water content, the driving power or the driving current, the heat medium temperature, the medium weight, the environmental temperature, and the environmental humidity. By inputting the value received by the reception unit into the learning model in which the relationship is learned using the included teacher data, the calculation by the learning model is executed.
The control device according to any one of claims 13 to 15, wherein the control unit controls the operation of the device including the powder processing device based on the calculation result of the learning model.
The reception section accepts the type of powder raw material,
The arithmetic processing unit selects a learning model corresponding to the type received by the reception unit from a plurality of learning models in which the relationship is learned for each type of the powder raw material, and uses the selected learning model as the user. By inputting a value including the desired particle size, the calculation by the learning model is executed.
The control device according to any one of claims 13 to 16, wherein the control unit controls the operation of the device including the powder processing device based on the calculation result of the learning model.
It is provided with an adjusting unit that compares the particle size of the powder desired by the user with the particle size data of the powder obtained from the powder processing apparatus and adjusts the control parameter according to the comparison result.
The control device according to any one of claims 13 to 17, wherein the control unit controls the operation of the device including the powder processing device based on the adjusted control parameters.
A screen generator that generates at least one of the measurement data, particle size data of the powder obtained from the powder processing apparatus, and screen data including an overall view of the powder processing apparatus.
The control device according to any one of claims 13 to 18, further comprising a transmission unit that transmits the generated screen data to an external terminal.
Accepts input of particle size desired by the user,
The received particle size data is used as the teacher data of the powder particle size data obtained from the powder processing device and the measurement data indicating the operating state of the powder processing device to obtain the powder particle size. , Input to the learning model that has learned the relationship with the control parameters that control the operating state of the powder processing apparatus.
Execute the calculation by the learning model
A control method for controlling the operation of an apparatus including the powder processing apparatus based on a calculation result obtained from the learning model.
Based on the received particle size data, the rotation speed of the crushing rotor that crushes the powder raw material, the rotation speed of the classification rotor that classifies the powder, and the discharge / suction flow rate when the powder is taken out from the processing chamber of the powder processing apparatus. , The rotation speed of the agitator that agitates the medium, and at least one of the crushing pressures of the fluid introduced into the processing chamber of the powder processing apparatus are used to input the relationship into a learning model that has been trained.
The control method according to claim 20, wherein the operation of the device including the powder processing device is controlled based on the calculation result obtained from the learning model.
The relationship is learned by using the measurement data including at least one of the supply amount of the powder raw material to be supplied to the processing chamber of the powder processing apparatus and the temperature of the processing chamber from the received particle size data. Enter into the training model and
The control method according to claim 20 or 21, wherein the operation of the device including the powder processing device is controlled based on the calculation result obtained from the learning model.
The particle size desired by the user, the circularity of the powder, the water content of the powder raw material, the driving power or driving current of the powder processing apparatus, the heat medium temperature, the medium weight, the environmental temperature, and the environmental humidity are low. Also accept one,
The received data further includes at least one of the particle size, the circularity, the water content, the driving power or the driving current, the heat medium temperature, the medium weight, the environmental temperature, and the environmental humidity. Input the above relationship into a learning model that has been trained using the teacher data,
The control method according to any one of claims 20 to 22, which controls the operation of the device including the powder processing device based on the calculation result obtained from the learning model.
Accepts the type of powder raw material,
A learning model corresponding to the received type is selected from a plurality of learning models in which the above relationship is learned for each type of powder raw material.
By inputting the data into the selected learning model, the calculation by the learning model is executed.
The control method according to any one of claims 20 to 23, which controls the operation of the device including the powder processing device based on the calculation result obtained from the learning model.
The particle size of the powder desired by the user is compared with the particle size data of the powder obtained from the powder processing apparatus.
Adjust the control parameters according to the comparison result,
The control method according to any one of claims 20 to 24, which controls the operation of the device including the powder processing device based on the adjusted control parameters.
At least one of the measurement data in the powder processing apparatus, the particle size data of the powder obtained from the powder processing apparatus, and the screen data including the overall view of the powder processing apparatus are generated.
The control method according to any one of claims 20 to 25, wherein the generated screen data is transmitted to an external terminal.
A device including the powder processing device includes the powder processing device and at least one of a raw material feeder, a hot air generator, a cyclone, a dust collector, and a blower or a pump connected to the powder processing device. The control method according to any one of 20 to 26.
On the computer
Accepts input of particle size desired by the user,
Using the particle size data of the powder obtained from the powder processing device and the measurement data indicating the operating state of the powder processing device as training data, the particle size of the powder and the powder received are used. By inputting the relationship between the control parameters that control the operating state of the body processing device into the learned learning model, the calculation by the learning model is executed.
A computer program for executing a process of controlling the operation of the device including the powder processing device based on the calculation result of the learning model.