CN118043584A

CN118043584A - Data processing system, data processing method, and computer readable medium

Info

Publication number: CN118043584A
Application number: CN202280056955.1A
Authority: CN
Inventors: 冲野史岳
Original assignee: Kaneko Sangyo Co Ltd
Current assignee: Kaneko Sangyo Co Ltd
Priority date: 2022-01-17
Filing date: 2022-10-26
Publication date: 2024-05-14
Also published as: WO2023135892A1; JP2023104150A; KR20240032108A

Abstract

The invention provides a data processing system which can estimate the acting torque acting on a valve rod without a torque sensor and can judge whether a fluid pressure driven valve is abnormal or not according to the estimated acting torque. A data processing system (300) is provided with at least a main valve (11), a driving device (12) for driving a valve rod (13) connected to the main valve (11), and a fluid pressure driving valve (10) for controlling a solenoid valve (1) for driving the supply and discharge of a fluid, and is provided with: an input data acquisition unit (301) that acquires input data including at least time series data of a valve opening degree of a main valve (11) and time series data of a solenoid valve output side pressure of a driving fluid supplied and discharged from a solenoid valve (1) to a driving device (12); an inference unit (301) for inputting the input data acquired by the input data acquisition unit (301) to a learning model in which a correlation between the input data and output data including time series data of an acting torque acting on the valve rod (13) corresponding to the input data is learned by machine learning, and inferring the time series data of the acting torque corresponding to the input data; an abnormality determination unit (305) that performs predetermined processing on the time-series data of the applied torque inferred by the inference unit (302), and determines whether or not there is an abnormality in at least one of the main valve (11) and the drive device (12) based on the result of the processing.

Description

Data processing system, data processing method, and computer readable medium

Technical Field

The present invention relates to a data processing system and a data processing method.

Background

Conventionally, a fluid pressure driven valve is known in which a driving fluid is controlled by a solenoid valve to open and close a main valve via a valve stem driven by a driving device. For example, patent document 1 discloses an emergency stop valve device as a fluid pressure driven valve used in a piping of a plant, which controls a driving fluid by an electromagnetic valve and closes a ball valve (main valve) via a driving device to shut off the fluid flowing in the piping when an emergency such as an abnormality occurs in the plant.

Prior art literature

Patent literature

Patent document 1: japanese patent laid-open No. 2009-97539

Disclosure of Invention

Problems to be solved by the invention

In the fluid pressure driven valve such as the emergency stop valve for the plant disclosed in patent document 1, it is desirable to grasp the operation state of the fluid pressure driven valve in order to improve the operation rate and reliability of the whole plant. For example, when the main valve is opened and closed via the valve stem driven by the driving device, if the magnitude and the amount of change of the acting torque acting on the valve stem can be measured, the operating state of the fluid pressure driven valve can be accurately grasped.

On the other hand, in order to measure the torque applied to the valve stem, a torque sensor needs to be provided in the valve stem, but the torque sensor is expensive, and therefore the cost of the fluid pressure driven valve increases. Further, depending on the environment in which the fluid pressure driven valve is used, additional measures for protecting the torque sensor from contamination or the like that may cause a failure may be required.

The present invention has been made in view of the above-described problems, and an object thereof is to provide a data processing system and a data processing method capable of estimating an operating torque acting on a valve rod without a torque sensor and determining the presence or absence of an abnormality of a fluid pressure actuated valve based on the estimated operating torque.

Solution for solving the problem

In order to achieve the above object, a data processing system according to one aspect of the present invention is a data processing system for a fluid pressure driven valve including at least a main valve, a driving device that drives a valve stem connected to the main valve, and a solenoid valve that controls supply and discharge of driving fluid to and from the driving device, the data processing system including:

An input data acquisition unit that acquires input data including at least time series data of a valve opening degree of the main valve in a predetermined period and time series data of a solenoid valve output side pressure of the driving fluid supplied from the solenoid valve to the driving device in the predetermined period;

An inference unit that infers time series data of the acting torque acting on the valve stem in the predetermined period corresponding to the input data by inputting the input data acquired by the input data acquisition unit to a learning model in which a correlation between the input data and output data including time series data of the acting torque acting on the valve stem in the predetermined period corresponding to the input data is learned by machine learning; and

And an abnormality determination unit configured to perform predetermined processing on the time-series data of the acting torque inferred by the inference unit, and determine whether or not at least one of the main valve and the driving device is abnormal based on a result of the processing.

Effects of the invention

The data processing system according to one aspect of the present invention can provide a learned model that can estimate the applied torque applied to the valve stem by causing the learned model to learn the correlation between the time series data of the valve opening degree and the time series data of the solenoid valve output side pressure and the time series data of the applied torque applied to the valve stem. Therefore, by using the learned model, the torque applied to the valve rod can be estimated without using a torque sensor. Further, it is possible to easily determine whether or not the fluid pressure driven valve is abnormal based on the estimated operating torque.

Problems, structures, and effects other than those described above will be clarified in the following embodiments.

Drawings

Fig. 1 is a schematic diagram showing an example of a fluid pressure driven valve to which a machine learning device or the like according to an embodiment of the present invention is applied.

Fig. 2 is a schematic diagram showing an example of a driving device such as a machine learning device to which an embodiment of the present invention is applied.

Fig. 3 is a schematic diagram showing an example of a solenoid valve to which the machine learning device and the like according to an embodiment of the present invention are applied.

Fig. 4 is a block diagram showing an example of a solenoid valve to which the machine learning device and the like according to an embodiment of the present invention are applied.

Fig. 5 is a schematic block diagram of a machine learning device of an embodiment of the present invention.

Fig. 6 is a diagram showing a configuration example (supervised learning) of data used in the machine learning apparatus and the like according to an embodiment of the present invention.

Fig. 7 is a diagram showing an example of a neural network model for supervised learning implemented in a machine learning apparatus of an embodiment of the present invention.

Fig. 8 is a flowchart showing an example of a machine learning method of an embodiment of the present invention.

FIG. 9 is a schematic block diagram illustrating a data processing system in accordance with an embodiment of the present invention.

Fig. 10 is a diagram illustrating an example of abnormality determination by the abnormality determination unit in the data processing system according to the embodiment of the present invention.

Fig. 11 is a diagram illustrating another example of abnormality determination by the abnormality determination unit in the data processing system according to the embodiment of the present invention.

Fig. 12 is a flowchart showing an example of a data processing method of the data processing system of an embodiment of the present invention.

Detailed Description

Embodiments for carrying out the present invention will be described below with reference to the accompanying drawings. The range necessary for the description of the present invention to achieve the object is schematically shown below, and the range necessary for the description of the corresponding portions of the present invention is mainly described, and the portions omitted from the description are known in the art.

Before describing a data processing system and a data processing method according to an embodiment of the present invention, a fluid pressure driven valve applied to a data processing system and the like will be described below.

(Fluid pressure actuated valve)

Fig. 1 is a schematic diagram showing an example of a fluid pressure actuated valve 10 according to an embodiment of the present invention. The fluid pressure driven valve 10 according to the present embodiment can be used as an emergency shut-off valve which is provided in a plant in a pipe 100 through which various gases, oil, etc. flow, and which shuts off the flow of the pipe 100 when an emergency stop such as an abnormality occurs in the plant. The place where the fluid pressure actuated valve 10 is installed and the application are not limited to the above examples.

The fluid pressure driven valve 10 shown in fig. 1 includes: a main valve 11 disposed in the middle of the pipe 100; a fluid pressure type driving device 12 for opening and closing the main valve 11 by driving a valve stem 13 connected to the main valve 11 according to the fluid pressure of the driving fluid; and an electromagnetic valve 1 having a function of controlling supply and discharge of driving fluid to and from the driving device 12.

The driving fluid for the fluid pressure driving valve 10 employs a pneumatic gauge compressed air (hereinafter simply referred to as "air") a. The air a is supplied from the air supply source 14 to the solenoid valve 1 via the first air pipe 140 and further to the driving device 12 via the second air pipe 141. Further, a communication cable 150 for transmitting and receiving various data between the external device 15 and the solenoid valve 1 and a power cable 160 for supplying power from the external power source 16 to the solenoid valve 1 are connected to the fluid pressure driven valve 10. The driving fluid is not limited to the air a described above, and may be other gas or liquid (e.g., oil).

The external device 15 is a device for transmitting and receiving various information to and from the fluid pressure driven valve 10, and is constituted by an external storage unit such as a computer for plant management (including a local server and a cloud server), a diagnostic computer used by an operator (maintenance inspector), a USB (universal serial bus) memory, and an external HDD (hard disk drive), for example. The external device 15 may be connected to the machine learning device 200 described later to transmit various data constituting the learning data set. The external device 15 includes a notification means, such as a GUI (GRAPHICAL USER INTERFACE: graphical user interface), for notifying the worker of the occurrence of an abnormality or the content thereof when the abnormality occurs in the fluid pressure actuated valve 10. The communication between the external device 15 and the solenoid valve 1 may be wireless communication.

The fluid pressure actuated valve 10 of the present embodiment is driven by an airless closure method. Therefore, during steady operation, the main valve 11 is opened by supplying (supplying) air a from the air supply source 14 to the driving device 12 via the solenoid valve 1, and during emergency stop or test operation, the main valve 11 is closed by discharging (exhausting) air a from the driving device 12 via the solenoid valve 1. In this case, the main valve 11 is closed by supplying air a to the driving device 12, and the main valve 11 is closed by discharging air a from the driving device 12.

The main valve 11 is a ball valve. The main valve 11 has a specific structure including: a valve body 110 disposed in the middle of the pipe 100; and a spherical valve element 111 rotatably provided in the valve body 110, and a valve stem 13 is connected to an upper portion of the valve element 111. As the valve stem 13 rotates to 0 to 90 degrees, the valve element 111 rotates in the valve body 110, and is switched between a fully open state (state shown in fig. 1) and a fully closed state of the main valve 11. The valve used as the main valve 11 is not limited to a ball valve, and may be, for example, a butterfly valve or another on-off valve.

The driving device 12 employs a single-acting cylinder mechanism disposed between the main valve 11 and the solenoid valve 1. The specific configuration of the driving device 12 includes: a cylindrical cylinder 120 having a spring chamber 127 and a cylinder chamber 128; a pair of pistons 122A, 122B provided in the cylinder 120 so as to be capable of reciprocating linear movement, and coupled together via a piston rod 121; a coil spring 123 provided in a spring chamber 127 formed on the first piston 122A side; an air supply/discharge port 124 connected to a cylinder chamber 128 formed on the second piston 122B side; and a transmission mechanism 125 provided at a portion where the valve rod 13 disposed so as to penetrate the cylinder 120 in the radial direction is orthogonal to the piston rod 121. The driving device 12 is not limited to the single-acting type, and may be configured in other forms such as a double-acting type.

The first piston 122A is biased by a coil spring 123 provided in the spring chamber 127 to actuate the main valve 11 in the closing direction. The second piston 122B is pressed by the air a (supplied air) supplied from the air supply/discharge port 124 to the cylinder chamber 128, so that the main valve 11 is operated in the opening direction (against the biasing force of the coil spring 123). The transmission mechanism 125 is constituted by a rack and pinion mechanism, a scotch yoke mechanism, a link mechanism, a cam mechanism, and the like, and converts the reciprocating linear motion of the piston rod 121 into a rotational motion and transmits the rotational motion to the valve stem 13.

The valve stem 13 is formed in a rotatable shaft shape and includes a first shaft 13a, a second shaft 13b, and a third shaft 13 c. Both ends of the first shaft 13a are coupled to the second shaft 13b and the third shaft 13c via a coupling, a connector, or the like, respectively, and the first shaft 13a, the second shaft 13b, and the third shaft 13c are coaxially arranged. The first shaft 13a is disposed to penetrate the driving device 12 and is driven by the driving device 12. The second shaft 13b is coupled to the first shaft 13a and to the main valve 11. The third shaft 13c is coupled to the first shaft 13a, and is inserted into the housing 6 of the solenoid valve 1 to be axially supported. The valve stem 13 performs rotational movement as a whole in synchronization with driving the first shaft 13 a.

Fig. 2 is a schematic diagram showing an example of the driving device 12 to which the machine learning device or the like of the embodiment of the present invention is applied. Fig. 2 (a) shows an example in which the transmission mechanism 125 is a rack-and-pinion mechanism, and fig. 2 (b) shows an example in which the transmission mechanism 125 is a scotch yoke mechanism.

For example, as shown in fig. 2 (a), when the transmission mechanism 125 is a rack-and-pinion mechanism, air a is supplied to the cylinder 120 or discharged from the cylinder 120, whereby the piston 122 and the piston rod 121 reciprocate linearly. Then, the rack 125a provided to the piston rod 121 performs a reciprocating linear motion. Next, the pinion 125b, which is in contact with and engaged with the rack 125a, rotates. Then, the first shaft 13a, which rotates similarly to the pinion 125b, performs a rotational movement.

In addition, as shown in fig. 2 (B), when the transmission mechanism 125 is a scotch yoke mechanism, air is supplied to the cylinder 120 or discharged from the cylinder 120 through the air a, whereby the pistons 122A, 122B and the piston rod 121 perform reciprocating linear motion. Then, the needle roller 125c, which moves in the same manner as the piston rod 121, reciprocates linearly. Next, the yoke 125d, in which the shaft eccentrically contacts the needle roller 125c and is assembled, is rotated by 90 °. Then, the first shaft 13a, which is rotated by 90 ° in the same manner as the yoke 125d, performs a rotational movement.

The drive device 12 further includes stoppers 126A, 126B and a drive state sensor 49, the stoppers 126A, 126B being capable of changing positions at which the movement of the pistons 122A, 122B is restricted, respectively, and the drive state sensor 49 acquiring the state of each part of the drive device 12. In the example shown in fig. 2 (a) and 2 (B), the stoppers 126A, 126B are constituted by bolts provided on the shafts of the piston rods 121 of the cylinder housings 120A, 120B of the cylinders 120, respectively.

The driving state sensor 49 is, for example, a position sensor 491, an acceleration sensor 492, a temperature and humidity sensor 493, or the like. In the example shown in fig. 2 (a) and 2 (B), as the driving state sensor 49, two position sensors 491, two acceleration sensors 492, and two (three in fig. 2 (B)) temperature and humidity sensors 493 are mounted on the stoppers 126A, 126B or the cylinder 120.

The position sensor 491 detects the position of the piston 122A, 122B or the piston rod 121 with respect to the cylinder housing 120A, 120B of the cylinder 120. The position sensor 491 is, for example, mounted on the first stopper 126A and the second stopper 126B, respectively, and measures the position (distance) of the piston 122A, 122B or the piston rod 121 with respect to the cylinder 120. The position sensor 491 is constituted by an ultrasonic sensor, an infrared sensor, a hall sensor, a magnetic reed sensor, or the like.

The acceleration sensor 492 measures acceleration generated in the cylinder 120. Specifically, the acceleration sensor 492 measures, as the acceleration of the driving device 12, the vibration of the driving device 12 generated when the pistons 122A and 122B reciprocate in the cylinder 120 and the vibration (impact) of the driving device 12 generated when the piston 122 collides with the stoppers 126A and 126B.

The temperature and humidity sensor 493 is installed in, for example, a screw hole formed in the cylinder 120, and detects the temperature and humidity inside the cylinder 120. The temperature/humidity sensor 493 is configured by combining a temperature sensor for detecting temperature and a humidity sensor for detecting relative humidity.

As shown in fig. 2, in the case where the driving device 12 is of a single-action type, since the air supply/discharge port 124a is connected to the solenoid valve 1 and the air supply/discharge ports 124b, 124c are open to the outside environment, the temperature/humidity sensor 493 detects the temperature and humidity of the air a (driving fluid) supplied from the solenoid valve 1 into the cylinder 120 via the air supply/discharge port 124a and the temperature and humidity of the outside air (outside air) supplied from the outside environment into the cylinder 120 via the air supply/discharge ports 124b, 124 c.

When the driving device 12 is double-acting, the air supply/discharge ports 124a and 124B are connected to the solenoid valve 1, and air a is alternately supplied or discharged from the solenoid valve 1 to the left and right air chambers formed in the cylinder 120 via the air supply/discharge ports 124a and 124B by way of the second piston 122B, and the air supply/discharge port 124c is opened to the outside environment, so that the temperature/humidity sensor 493 detects the temperature and humidity of the air a (driving fluid) supplied from the solenoid valve 1 into the cylinder 120 via the air supply/discharge ports 124a and 124B, and detects the temperature and humidity of the outside air (outside air) supplied from the outside environment into the cylinder 120 via the air supply/discharge port 124 c.

The positions and the number of the driving state sensors 49 mounted on the driving device 12 are not limited to the example shown in fig. 2, and may be changed as appropriate. For example, the position sensor 491 may be mounted to the cylinder housings 120A, 120B. The acceleration sensor 492 may be attached to the cylinder cases 120A and 120B, or may be attached to the distal end portion of the stopper 126A instead of the proximal end portion of the stopper 126A (see fig. 2B). The temperature/humidity sensor 493 may be installed inside the air supply/discharge port 124a, the first air pipe 140, the second air pipe 141, or the solenoid valve 1 connected to the solenoid valve 1. The temperature/humidity sensor 493 may be attached to either one of the air supply/discharge ports 124b and 124c opened to the outside environment, or may be attached to either one of a long pipe joint and a T-pipe connected to either one of the air supply/discharge ports 124b and 12 c.

The solenoid valve 1 has a function of controlling the supply and discharge of air a to and from the driving device 12, and is configured as a two-position normally-closed three-way solenoid valve (i.e., open when energized and closed when not energized). The solenoid valve 1 includes a spool valve portion 2 that switches a flow passage through which air a flows, and a solenoid portion 3 that displaces the spool valve portion 2 in accordance with an energized state (when energized or when non-energized), in an interior of a housing portion 6 that functions as a casing of the indoor type or explosion-proof type solenoid valve 1. The solenoid valve 1 is not limited to the three-way solenoid valve of the above type, and may be a three-position normal four-way solenoid valve or the like, or may be formed in various manners based on any combination of these. In the present embodiment, the solenoid valve 1 is used as a pilot valve in the fluid pressure driven valve 10, but the use of the solenoid valve 1 is not limited thereto.

The spool 2 has an input port 20 connected to the air supply source 14 via a first air pipe 140, an output port 21 connected to the drive device 12 via a second air pipe 141, and an exhaust port 22 for exhausting exhaust gas from the drive device 12.

The solenoid portion 3 displaces the spool portion 2 so as to communicate between the input port 20 and the output port 21 when energized, and displaces the spool portion 2 so as to communicate between the output port 21 and the exhaust port 22 when de-energized.

According to the above-described series of configurations, when the solenoid valve 1 is in the energized state, air a (supplied air) from the air supply source 14 flows through the first air pipe 140, the input port 20, the output port 21, and the second air pipe 141 in this order, and is supplied to the air supply/discharge port 124, thereby pressing the second piston 122B and compressing the coil spring 123. When the first shaft 13a (valve stem 13) rotates the piston rod 121 via the piston rod 121 and the transmission mechanism 125 by an amount corresponding to the compression of the coil spring 123, the valve body 110 rotates the valve element 111, and the main valve 11 is operated in the fully open state.

On the other hand, when the solenoid valve 1 is in the non-energized state, the air a (exhaust gas) in the cylinder 120 flows through the second air pipe 141, the output port 21, and the exhaust port 22 in this order from the air supply/exhaust port 124, and is discharged to the outside air, whereby the pressing force of the second piston 122B is reduced, and the coil spring 123 returns from the compressed state. When the first shaft 13a (valve stem 13) rotates the piston rod 121 via the transmission mechanism 125 by an amount corresponding to the restoration of the coil spring 123, the valve body 111 rotates within the valve body 110, and the main valve 11 is operated in the fully closed state.

Fig. 3 is a cross-sectional view showing an example of the solenoid valve 1 according to an embodiment of the present invention. As shown in fig. 3, the solenoid valve 1 of the present embodiment includes, in addition to the spool valve portion 2 and the solenoid portion 3 described above, a plurality of sensors 4 for obtaining states of respective portions of the solenoid valve 1, a substrate 5 on which at least one of the plurality of sensors 4 is mounted, the spool valve portion 2, the solenoid portion 3, and a housing portion 6 for housing the plurality of sensors 4 and the substrate 5.

The housing section 6 includes: a first housing portion 60 that houses the spool 2; a second housing portion 61 adjacent to the first housing portion 60 and housing the solenoid portion 3, the plurality of sensors 4, and the substrate 5; junction box 62 connects communication cable 150 and power cable 160. The first and second accommodation portions 60 and 61 are made of a metal material such as aluminum, for example.

The first housing portion 60 has openings (not shown) that function as the input port 20, the output port 21, and the exhaust port 22, respectively.

The second housing portion 61 includes: a cylindrical housing 610 having both ends (a first housing end 610a and a second housing end 610 b) open; a main body 611 disposed inside the housing 610; a solenoid cover 612 that covers the solenoid portion 3 fixed to the first housing end 610a from the outside; and a terminal block cover 613 that covers the terminal block 62 fixed to the second housing end 610b from the outside.

The housing 610 has: a shaft insertion port 610c formed at a lower portion of the housing 610 for inserting the third shaft 13c; a main body insertion port 610d formed at an upper portion of the housing 610 for insertion of the main body 611; a cable insertion port 610e formed on the second housing end 610b side, into which the communication cable 150 and the power cable 160 are inserted.

The first housing portion 60 and the second housing portion 61 are formed with a first flow passage 63, a second flow passage 64, and a spool flow passage 65 penetrating the main body 611, the first flow passage 63 branching from the input side flow passage 26, the input side flow passage 26 communicating with the first pressure sensor 40, the second flow passage 64 branching from the output side flow passage 27, the output side flow passage 27 communicating with the second pressure sensor 41, and the spool flow passage 65 for flowing the air a for interlocking the spool valve portion 2 with the solenoid portion 3.

The spool valve section 2 includes: a spool hole 23 formed in a second housing portion 61 functioning as a spool housing; a spool 24 movably disposed in the spool hole 23; a spool spring 25 that biases the spool 24; an input-side flow passage 26 that communicates between the input port 20 and the spool hole 23; an output-side flow passage 27 that communicates the output port 21 with the spool hole 23; and an exhaust runner 28 that communicates between the exhaust port 22 and the spool hole 23.

The solenoid section 3 includes: the solenoid housing 30, a solenoid coil 31 accommodated in the solenoid housing 30, a movable iron core 32 movably arranged in the solenoid coil 31, a fixed iron core 33 arranged in the solenoid coil 31 in a fixed state, and a solenoid spring 34 urging the movable iron core 32.

When the solenoid valve 1 is switched from the non-energized state to the energized state, a coil current flows to the solenoid coil 31 in the solenoid portion 3, and the solenoid coil 31 generates an electromagnetic force, by which the movable iron core 32 is attracted by the fixed iron core 33 against the urging force of the electromagnetic spring 34, and the circulation state of the air a flowing through the spool flow passage 65 is switched. In the spool portion 2, the spool 24 is moved against the urging force of the spool spring 25 by switching the flow state of the air a flowing through the spool flow passage 65, whereby the state of communication between the input port 20 and the exhaust port 22 is switched to the state of communication between the input port 20 and the output port 21.

The substrate 5 includes: a first substrate 50 having substrate surfaces 500A and 500B arranged along a third axis 13c inserted from a shaft insertion port 610 c; a second substrate 51 disposed adjacent to the junction box 62; and a third substrate 52 disposed adjacent to the solenoid portion 3.

The main body 611, the solenoid portion 3, and the third substrate 52 are disposed on the first substrate surface 500A side of the substrate surfaces 500A, 500B of the first substrate 50. On the second substrate surface 500B side opposite to the first substrate surface 500A side, the second substrate 51 and the junction box 62 are arranged.

The sensor 4 is arranged at a suitable position of the first substrate 50, the second substrate 51 and the third substrate 52. As the sensor 4, for example, there are: a first pressure sensor 40 that measures the fluid pressure of the air a flowing in the input-side flow passage 26 and the first flow passage 63; a second pressure sensor 41 that measures the fluid pressure of the air a flowing in the output-side flow passage 27 and the second flow passage 64; and a main valve opening sensor 42 that measures a rotation angle at which the third shaft 13c (valve stem 13) rotates, and obtains valve opening information of the main valve 11 based on the rotation angle.

The main valve opening sensor 42 is constituted by, for example, a magnetic sensor, measures the intensity of magnetic force generated by the permanent magnet 131 attached to the third shaft 13c, and obtains valve opening information of the main valve 11 based on the intensity of magnetic force. The main valve opening sensor 42 is preferably placed at a position facing the outer circumference of the third shaft 13c on the first substrate surface 500A of the first substrate 5 disposed along the third shaft 13c inserted from the shaft insertion port 610 c. Accordingly, the main valve opening sensor 42 and the third shaft 13c mounted on the first substrate 50 can be disposed close to each other in the housing portion 6 without wasting the disposing space, and the valve opening information can be accurately acquired.

Fig. 4 is a block diagram showing an example of the solenoid valve 1 and the fluid pressure driven valve 10 according to the embodiment of the present invention. As shown in fig. 4, the solenoid valve 1 includes, as an electrical configuration example, a control unit 7 for controlling the solenoid valve 1, a communication unit 8 for communicating with an external device 15, and a power supply circuit unit 9 connected to an external power supply 16, in addition to the solenoid unit 3 and the sensor 4.

The plurality of sensors 4 are provided as a sensor group for measuring physical quantities of the respective parts, and include, in addition to the first pressure sensor 40, the second pressure sensor 41, and the main valve opening sensor 42 described above, a voltage sensor 43 for measuring a supply voltage to the solenoid unit 3, a current resistance sensor 44 for measuring a current value at the time of energization and a resistance value at the time of non-energization in the solenoid unit 3, a temperature sensor 45 for measuring an internal temperature of the housing unit 6, and a magnetic sensor 46 for measuring a magnetic force intensity generated by the solenoid unit 3.

The plurality of sensors 4 are provided as a sensor group for acquiring information on the operation history of each part, and include: an operation timer (timer) 47 that measures at least one of a total of the energization time to the solenoid portion 3 and a current energization linked time as an operation time of the fluid pressure driven valve 10; and an operation counter (counter) 48 that counts the number of operations of each of the solenoid valve 1, the driving device 12, and the main valve 11.

The plurality of sensors 4 (hereinafter, "plurality of sensors 4" are defined as terms indicating the sensors indicated by reference numerals 40 to 49 including the driving state sensor 49) are not limited to being individually provided as described above, and may be provided with the function of other sensors by a specific sensor, and thus the other sensors are not individually provided. For example, the magnetic sensor 46 may measure the intensity of magnetic force generated by the solenoid unit 3, and determine the current value at the time of energization in the solenoid unit 3 from the intensity of magnetic force, so that the current resistance sensor 44 may not be separately provided. Further, the microcontroller 70 may have a function of a sensor built therein or may have a function of a sensor partially implemented therein, and for example, the microcontroller 70 may have the operation timer 47 and the operation counter 48 built therein so that the operation timer 47 and the operation counter 48 are not separately provided.

The control unit 7 includes: a microcontroller 70 that processes information indicating the states of the respective parts of the fluid pressure-driven valve 10 including the solenoid valves 1 and the driving devices 12 acquired by the plurality of sensors 4, and controls the respective parts of the fluid pressure-driven valve 10; and a valve test switch 71 for controlling the state of energization of the solenoid portion 3 and performing an opening/closing operation of the main valve 11 during test operation.

The microcontroller 70 includes a processor (not shown) such as a CPU (Central Processing Unit ) and a Memory including a ROM (Read Only Memory), a RAM (Random Access Memory ), and the like. Microcontroller 70 can include functionality to implement data processing system 300 described below in this embodiment.

The valve test switch 71 is configured to receive a command from the microcontroller 70 when a predetermined test operation condition is satisfied, and to perform a stroke test of the fluid pressure driven valve 10 as a test operation.

For example, the stroke test is performed by any one of the full stroke test and the partial stroke test. The full stroke test is executed by switching the main valve 11 from the energized state to the non-energized state in the fully open state to the fully closed state and from the non-energized state to the energized state to return to the fully open state in the fully closed state. The main valve 11 is not operated in the fully closed state (i.e., the plant is not stopped), and the main valve 11 is partially closed to the predetermined opening degree by being switched from the energized state to the non-energized state in the fully open state, and the partial stroke test is performed by being switched from the non-energized state to the energized state and returned to the fully open state in the partially closed state.

For example, when the execution timing based on the execution frequency (for example, 1 year) specified as the set value by the manager, the specific specified date arrives, the execution command from the external device 15 is received, and the manager operates a test button (not shown) provided in the solenoid valve 1, the test operation condition is satisfied, and the test operation (trip test) may be executed.

(Machine learning device)

In the fluid pressure driven valve 10 having the above-described series of configurations, by providing the plurality of sensors 4, various information of the fluid pressure driven valve 10 can be acquired at the time of steady operation and at the time of unsteady operation (for example, at the time of test operation including opening and closing operation and at the time of emergency stop). Therefore, the machine learning device 200 that learns an inference model (learned model) capable of estimating the state of the fluid pressure driven valve 10 based on information (state variable) that can be obtained from the fluid pressure driven valve 10 will be described below. The machine learning device 200 is provided not only as a device that works alone, but also in the form of a non-transitory computer-readable medium storing a program for causing an arbitrary processor to execute actions described below or one or more instructions for executing the actions.

Fig. 5 is a schematic block diagram of a machine learning device 200 of an embodiment of the present invention. As shown in fig. 5, the machine learning device 200 of the present embodiment includes a learning data set acquisition unit 201, a learning data set storage unit 202, a learning unit 203, and a learning model storage unit 204.

The learning data set acquisition unit 201 is, for example, an interface unit for acquiring a plurality of data constituting a learning (training) data set from various devices connected via a wired or wireless communication line.

Here, examples of the various devices connected to the learning data set acquisition unit 201 include the external device 15, a worker computer PC1 used by a worker of the fluid pressure driven valve 10, and the like. In fig. 5, the external device 15 and the worker computer PC1 are shown as separate devices, but the external device 15 and the worker computer PC1 may be configured by the same computer.

The test device 17 for performing various tests of the fluid pressure actuated valve 10 is connected to the external device 15. The test device 17 includes the fluid pressure driven valve 10 as a test target, and a first torque sensor 170A and a second torque sensor 170B for measuring the acting torque acting on the valve stem 13 when a test involving an opening/closing operation is performed. The test device 17 may be provided with at least one of the first torque sensor 170A and the second torque sensor 170B.

The first torque sensor 170A measures a first acting torque acting between the first shaft 13a and the second shaft 13 b. The second torque sensor 170B measures a second acting torque acting between the first shaft 13a and the third shaft 13 c.

The learning data set acquisition unit 201 acquires, for example, measurement results measured by the plurality of sensors 4 of the fluid pressure-driven valve 10 as input data from the test device 17 via the external device 15. The learning data set acquisition unit 201 acquires, for example, a measurement result of the applied torque measured by at least one of the first torque sensor 170A and the second torque sensor 170B as output data from the test device 17 via the external device 15. Then, by associating these input data and output data with each other, one learning data set described later is constituted.

Fig. 6 is a diagram showing a configuration example (supervised learning) of data used in the machine learning apparatus 200 according to an embodiment of the present invention. Note that fig. 6 is also appropriately referred to in the description of the data processing system 300 and the inference means.

As shown in fig. 6, the learning data set is a data set for use in machine learning described later, and includes, as input data, at least time series data of a valve opening degree of the main valve 11 and time series data of a solenoid valve output side pressure of the air a in a predetermined period, and also includes, as output data, time series data of an applied torque in the predetermined period. Examples of these various data will be described below in detail, but the present invention is not limited to these.

The valve opening of the main valve 11 included in the input data is a value of the open/closed state of the main valve 11, and can be obtained from the main valve opening sensor 42.

The solenoid valve output side pressure of the air a included in the input data is the pressure of the air a supplied from the solenoid valve 1 to the driving device 12, and includes the pressure of the air a (supply air) when the air a is supplied from the solenoid valve 1 to the driving device 12 and the pressure of the air a (exhaust air) when the air a is discharged from the driving device 12 to the outside via the solenoid valve 1. The solenoid valve output side pressure of the air a can be obtained by the second pressure sensor 41.

The operating torque included in the output data is constituted by at least one of the torque acting on the valve stem 13 by the pressure generated by the gas, oil, or the like flowing through the pipe 100 and the torque acting on the valve stem 13 when the valve stem 13 is rotated by the electromagnetic valve 1 and the driving device 12, and is acquired by the first torque sensor 170A or the second torque sensor 170B. Alternatively, both the applied torque measured by the first torque sensor 170A and the applied torque measured by the second torque sensor 170B may be obtained as output data.

The time series data is constituted by a plurality of data acquired at a plurality of different times within a predetermined period, for example, acquired at a predetermined sampling period. In the present embodiment, the time series data of the valve opening degree of the main valve 11, the time series data of the solenoid valve output side pressure of the air a, and the time series data of the applied torque are acquired at a plurality of times with the same sampling period and the same phase (the state without the phase difference), but at least one of the sampling period and the phase may be different.

The predetermined period is a period during which the opening and closing operation of the main valve 11 is performed at the time of steady operation and the time of unsteady operation (including the time of test operation and the time of emergency stop) of the working environment of the fluid pressure driven valve 10.

The predetermined period during the unstable operation is constituted by, for example, an execution period during which the stroke test in the fluid pressure driven valve 10 is executed. The predetermined period may be the entire period from the start of the test to the end of the test in the execution period of the stroke test, or may be a part of the period. Therefore, the predetermined period may be, for example, the entire period of the full stroke test (full open state→full closed state→full open state), the entire period of the partial stroke test (full open state→partially closed state), the partial period of the full stroke test (full open state→full closed state, or fully closed state→full open state, etc.), or the partial period of the partial stroke test (full open state→partially closed state, partially closed state→fully open state, etc.), and is not limited thereto.

The predetermined period may be the entire period from the start to the end of the period in which the opening/closing operation of the main valve 11 is performed, or may be a partial period thereof. Therefore, the predetermined period may be, for example, the entire period of the opening operation from the start of the opening operation of the main valve 11 to the end of the opening operation, or may be a partial period thereof. The entire period of the closing operation from the start of the closing operation of the main valve 11 to the end of the closing operation may be the entire period or a part of the period. Further, the entire period including the opening operation period and the closing operation period may be the opening operation period or the closing operation period, or a part thereof, and the present invention is not limited thereto.

Alternatively, the input data in the learning data set may include time series data of the pressure of the air a (excluding the solenoid valve output side pressure of the air a), time series data of the control parameter of the solenoid portion 3, time series data of the temperature of the solenoid valve 1, time series data of the operation state of the driving device 12, the total operation time of the fluid pressure driven valve 10, the operation time since the last power supply to the fluid pressure driven valve 10, the number of operations of the main valve 11, the number of operations of the driving device 12, the number of operations of the solenoid portion 3, and the opening/closing time of the main valve 11.

The pressure of the air a (except the solenoid valve output side pressure of the air a) is preferably the pressure of the air a flowing in each portion inside the fluid pressure driven valve 10, and specifically, preferably includes at least one of the solenoid valve input side pressure of the air a supplied from the air supply source 14 to the solenoid valve 1 and the differential pressure between the solenoid valve input side pressure of the air a and the solenoid valve output side pressure of the air a.

The control parameter of the solenoid unit 3 is various parameter information for controlling the solenoid unit 3, and specifically preferably includes at least one of a supply voltage to be supplied to the solenoid coil 31, a current value when the solenoid coil 31 is energized, a resistance value when the solenoid coil 31 is not energized, an operation time of the solenoid unit 3, and a magnetic force intensity generated in the solenoid unit 3. The supply voltage to be supplied to the solenoid coil 31 can be obtained by the voltage sensor 43, the current value when the solenoid coil 31 is energized and the resistance value when the solenoid coil 31 is not energized can be obtained by the current resistance sensor 44, the operation time of the solenoid unit 3 can be obtained by the operation timer 47, and the intensity of the magnetic force generated in the solenoid unit 3 can be obtained by the magnetic sensor 46.

The temperature of the solenoid valve 1 is a value of the solenoid valve 1, particularly the internal temperature, and can be obtained by the temperature sensor 45 described above.

The operational state of the drive device 12 preferably includes at least one of a position of the first piston 122A relative to the cylinder 120, a position of the second piston 122B relative to the cylinder 120, an acceleration of the drive device 12 based on vibrations generated between the cylinder 120 and the pistons 122A, 122B, and a temperature and humidity within the cylinder 120.

The position of the pistons 122A, 122B relative to the cylinder 120 may be obtained by the position sensor 491 described above. The acceleration of the driving device 12 includes acceleration of the cylinder 120 based on vibration generated when the pistons 122A, 122B reciprocate linearly in the cylinder 120 and acceleration of the cylinder 120 based on vibration (shock) generated when the pistons 122A, 122B collide with the stoppers 126A, 126B. The acceleration of the driving device 12 can be obtained by the acceleration sensor 492. The temperature and humidity in the cylinder 120 preferably include at least one of the temperature and humidity of the air a (driving fluid) supplied from the solenoid valve 1 to the driving device 12 via the air supply/discharge port 124 (124 a in the case of the single action type shown in fig. 2, 124a, 124b in the case of the double action type), and the temperature and humidity of the outside air (outside air) supplied from the outside environment to the driving device 12 via the air supply/discharge port 124 (124 b, 124c in the case of the single action type shown in fig. 2, 124c in the case of the double action type). The temperature and humidity in the cylinder 120 may be obtained by the temperature and humidity sensor 493 described above.

The total operation time of the fluid pressure driven valve 10 and the operation time since the last power on of the fluid pressure driven valve 10 can be obtained by the operation timer 47, the number of operations of the main valve 11, the number of operations of the driving device 12, and the number of operations of the solenoid portion 3 can be obtained by the operation counter 48, and the opening and closing time of the main valve 11 can be obtained by using a timer or the like, not shown.

As described above, increasing the type of input data contributes to substantially improving the estimation accuracy of the learned model obtained after machine learning, but using input data having a low degree of correlation with output data may rather prevent improvement of the estimation accuracy of the learned model. Therefore, the number and type of data to be used as input data should be appropriately selected in consideration of the state of the fluid pressure-driven valve 10 to which the learned model is applied, and the like.

Therefore, in the present embodiment, the estimation of the applied torque acting on the valve rod 13 is performed based on the time series data of the valve opening degree of the main valve 11 and the time series data of the solenoid valve output side pressure of the air a. That is, the piston 122 moves relative to the cylinder 120 due to the change in the solenoid valve output side pressure of the air a supplied from the solenoid valve 1 to the driving device 12, and the valve opening degree of the main valve 11 changes, but it is possible to estimate time series data of the operating torque acting on the valve rod 13 according to the change in the valve opening degree and the solenoid valve output side pressure.

The learning data set storage unit 202 is a database for storing a plurality of data for constituting the learning data set acquired by the learning data set acquisition unit 201 as one learning data set by associating the plurality of data with associated input data and output data. The specific structure of the database constituting the learning data set storage unit can be appropriately adjusted. For example, in fig. 5, the learning data set storage unit 202 and the learning model storage unit 204 described later are shown as separate storage units for convenience of explanation, but they may be constituted by one storage medium (database).

The learning unit 203 performs machine learning by using the plurality of learning data sets stored in the learning data set storage unit 202, thereby generating a learning-completed model in which the correlation between the input data and the output data included in the plurality of learning data sets is learned. In the present embodiment, as a specific method of machine learning, as will be described in detail later, supervised learning using a neural network is employed. However, the specific method of machine learning is not limited to this, and other learning methods may be employed as long as the correlation between input and output can be learned from the learning data set. For example, ensemble learning (random forest, boosting method, etc.) may also be used.

The learned model storage unit 204 is a database for storing the learning model generated by the learning unit 203. The learned model stored in the learned model storage unit 204 is applied to an actual system via a communication line including the internet, a storage medium, and the like according to a request. The specific manner in which the learning model is applied to the actual system (data processing system 300) will be described in detail later.

Next, a learning method in the learning unit 203 using the plurality of learning data sets obtained as described above will be described centering on supervised learning. Fig. 7 is a diagram showing an example of a neural network model for supervised learning implemented in the machine learning apparatus according to an embodiment of the present invention.

The neural network in the neural network model shown in fig. 7 is composed of l neurons (x 1 to xl) in the input layer, m neurons (y 11 to y1 m) in the first intermediate layer, n neurons (y 21 to y2 n) in the second intermediate layer, and o neurons (z 1 to zo) in the output layer. The first intermediate layer and the second intermediate layer are also called hidden layers, and may have a plurality of hidden layers in addition to the first intermediate layer and the second intermediate layer as a neural network, or may use only the first intermediate layer as a hidden layer. In fig. 7, a neural network model in which a plurality of (o) output layers are set is illustrated.

Further, between the input layer and the first intermediate layer, between the first intermediate layer and the second intermediate layer, and between the second intermediate layer and the output layer, synapses (ends) connecting the inter-layer neurons (nodes) are laid, and weights wij (i, j denote natural numbers, i denotes the ends of which layer, and j denotes the numbers of the ends of the neurons connecting the same layer) are provided corresponding to each end.

The neural network in the neural network model of the present embodiment learns, using the learning data set, the correlation between the time series data of the valve opening degree of the main valve 11 and the time series data of the solenoid valve output side pressure of the air a in the predetermined period and the time series data of the acting torque acting on the valve rod 13.

Specifically, the time series data of the valve opening degree of the main valve 11 and the time series data of the solenoid valve output side pressure of the air a in a predetermined period as state variables are associated with neurons of the input layer, and the value of the neurons in the output layer is calculated by using a general method of calculating the output value of the neural network, that is, a method of calculating the value of the neurons on the output side for all neurons other than the neurons in the input layer as the sum of the value of the neurons on the input side connected to the neurons and the number of the multiplication value series corresponding to the end connecting the neurons on the output side and the neurons on the input side.

The form of the information acquired as the state variable when the state variable is input to the neuron of the input layer can be appropriately set in consideration of the accuracy of the generated learning model. Specifically, preprocessing can be performed on specific input data to adjust the number of neurons corresponding to the input data, respectively, or to a value that can correspond to the neurons. For example, as the preprocessing, the range of values may be converted into values within a certain range by normalization.

Then, the following operations are repeatedly performed: the values of the o neurons z1 to zo in the calculated output layer, that is, the training data t1 to constituted by time series data of the acting torque acting on the valve stem 13, which constitute a part of the learning data set in the present embodiment, are compared to calculate errors, and the weights wij (back propagation) corresponding to the respective ends are adjusted so that the calculated errors become smaller.

When the series of steps is repeated a predetermined number of times or when the error is smaller than a predetermined condition such as an allowable value, learning is completed, and the neural network model (ownership wij corresponding to each end) is stored in the learned model storage unit 204 as a learned model.

(Machine learning method)

In summary, the present invention provides a machine learning method. Next, the machine learning method according to the present invention will be described with reference to fig. 6 (learning stage), fig. 7, and fig. 8. Fig. 8 is a flowchart showing an example of a machine learning method according to an embodiment of the present invention. In the machine learning method shown below, the description is based on the machine learning device 200 described above, but the structure to be provided is not limited to the machine learning device 200 described above. The machine learning method is implemented by using a computer, but various computers can be applied as the computer, and examples thereof include a computer device constituting the external device 15, the personal computer PC1 or the microcontroller 70, a server device disposed on a network, and the like. The specific configuration of the computer may be, for example, a configuration including an arithmetic device including at least a CPU, a GPU, and the like, a storage device including a volatile or nonvolatile memory, a communication device for communicating with a network or other devices, and a bus connecting these devices.

As the supervised learning of the machine learning method of the present embodiment, as a preparation in advance for starting the machine learning, first, a desired number of learning data sets are prepared (see fig. 6), and the prepared plurality of learning data sets are stored in the learning data set storage unit 202 (step S11). The number of learning data sets prepared here can be set in consideration of the inference accuracy required for the finally obtained learning model.

As a method for preparing a learning data set used in supervised learning, time series data acquired with the same sampling period and the same phase are acquired in a predetermined period, time series data of a valve opening degree of the main valve 11 and time series data of a solenoid valve output side pressure of the air a are used as input data, time series data of an acting torque of the valve stem 13 is used as output data, and detection of the valve opening degree by the valve opening degree sensor 42, detection of a solenoid valve output side pressure of the air a by the second pressure sensor 41, and detection of the acting torque by the first torque sensor 170A are used. The predetermined period for acquiring the time series data may be the entire period from the start of the test to the end of the test in the execution period of the stroke test, or may be a partial period thereof, but the time series data obtained by sampling the same period is used as the learning data set. Since the valve opening fluctuates greatly for a predetermined time period from the start of the stroke test and a predetermined time period before the end of the stroke test, the sample data in this period may not be placed in the learning data set. By performing the multi-pass test, a desired number of learning data sets can be prepared.

However, since various information on the fluid pressure-driven valves 10 often tends to be unique to each of the fluid pressure-driven valves 10, it is preferable that the data constituting the learning data set be collected only from a predetermined one of the fluid pressure-driven valves 10 to which a learning model obtained by machine learning described later is applied.

When step S11 is completed, the neural network model before learning is next prepared to start learning in the learning unit 203 (S12). The neural network model before learning prepared here has, for example, a structure shown in fig. 7 as its structure, and the weights wij of the respective ends are set to initial values. Then, one learning data set is selected at random from the plurality of learning data sets stored in the learning data set storage unit 202, for example (step S13), and input data in the one learning data set is input to the input layer of the neural network model before learning (see fig. 7) that is ready (step S14).

Here, since the value of the output layer (refer to fig. 7) generated as a result of the above-described step S14 is generated by the neural network model before learning, it is a value different from the desired result, that is, a value different from the time series data of the acting torque of the valve stem 13 in many cases. Therefore, next, machine learning is performed using the time series data of the acting torque of the valve stem 13 as training data in one learning data set acquired in step S13 and the value of the output layer generated in step S13 (step S15). The machine learning performed here may be, for example, a process (back propagation) of comparing time series data of the acting torque of the valve stem 13 constituting the training data with the value of the output layer and adjusting weights wij corresponding to the respective ends in the neural network model before learning to obtain a preferable value of the output layer. The number and form of the values output to the output layer of the neural network model before learning are the same as those of the training data in the learning data set to be learned.

Specifically exemplified are machine learning as referred to herein. The time series data of the acting torque of the valve stem 13 constituting the training data may be compared with the value output from the output layer one by one, but the time series data of the acting torque may be used as an acting torque vector, and the data output from the output layer may be used as an output vector, and for example, when the scalar of the acting torque vector matches the scalar of the output data, and the inner product of the normalized acting torque vector and the vector of the normalized output data is "1", it may be determined that the time series data of the acting torque matches the output data. Therefore, in step S15, the weights wij corresponding to the respective ends of the neural network model in the learning process are adjusted so that the time series data of the applied torque of the valve stem 13 matches the output data.

When machine learning is performed in step S15, it is further determined whether or not machine learning needs to be continued, for example, based on the remaining number of learning data sets that are not learned and stored in the learning data set storage unit 202 (step S16). Then, when the machine learning is continued (no in step S16), the routine returns to step S13, and when the machine learning is completed (yes in step S16), the routine proceeds to step S17. In the case of continuing the machine learning, the steps S13 to S15 are performed a plurality of times using the learning data set that has not been learned, with respect to the neural network model during the learning. The accuracy of the finally generated learning model generally increases in proportion to the number of times.

When the machine learning is completed (yes in step S16), the weights wij associated with the respective ends are stored in the learning model storage unit 204 as the learning model of the neural network generated by the series of process adjustment (step S17), and a series of learning processes are completed. The learned model stored here can be applied to and used by the data processing system 300 described later.

In the learning process and the machine learning method of the machine learning apparatus described above, a method of obtaining a learned model sufficient to be applied to the data processing system 300 by repeating the machine learning process on one (before-learning) neural network model a plurality of times to improve the accuracy thereof in order to generate one learned model is described. However, the present invention is not limited to this acquisition method. For example, the learned model subjected to the predetermined number of machine learning may be stored in advance in the plurality of learned model storage sections 204 as one candidate, the data set for validity determination may be input to the plurality of learned model groups, the (neuron values of the) output layer may be generated, the accuracy of the values determined at the output layer may be compared, and the one optimal learning model to be applied to the data processing system 300 may be selected. The validity judgment data set may be composed of the same data set as the learning data set used for learning, and may not be used for learning.

As described above, by applying the machine learning device and the machine learning method according to the present embodiment, a learned model can be obtained in which the time series data of the operating torque of the valve rod 13 can be reliably derived from the time series data of the valve opening degree of the main valve 11 and the time series data of the solenoid valve output side pressure of the air a.

In the above description, the case where the time-series data of the acting torque detected by the first torque sensor 170A is used for the learning data set is described, but the time-series data of the acting torque detected by the second torque sensor 170B may be changed to the time-series data of the acting torque detected by the first torque sensor 170A as the learning data set, and a learning model of the neural network that derives the time-series data of the valve shaft 13 may be obtained.

Alternatively, the number of nodes in the output layer of the neural network model may be increased, and a learning model of the neural network may be prepared in which both of the time series data detected by the first torque sensor 170A and the second torque sensor 170B are used as a learning data set of training data, and the time series data of the acting torques of the two types of valve rods 13 are derived.

In the above, the case where the time series data of the valve opening degree of the main valve 11 and the time series data of the solenoid valve output side pressure of the air a are used as the input data has been described, but the time series data of the solenoid valve input side pressure of the air a detected by the first pressure sensor 40 may be added to the input data as a learning data set.

In addition, when the input data is added, a node for inputting the time series data of the solenoid valve input side pressure of the air a may be added to a node of the input layer of the neural network model, and a learning model of the neural network may be obtained in which the time series data of the valve opening degree of the main valve 11, the time series data of the solenoid valve output side pressure of the air a, and the time series data of the solenoid valve input side pressure of the air a are used as input data to derive the time series data of the operating torque of the valve rod 13 (or the time series data of the operating torques of the 2 kinds of valve rods 13). Alternatively, a learning dataset using time series data of differential pressure between the solenoid valve input side pressure of air a and the solenoid valve output side pressure of air a instead of time series data of the solenoid valve input side pressure of air a may be used.

Further, as the learning data set, data obtained by adding the operation time detected by the operation timer (timer) 47 to the input data may be used, and in addition to the time series data of the valve opening degree of the main valve 11 and the time series data of the solenoid valve output side pressure of the air a (or the time series data of the valve opening degree of the main valve 11, the time series data of the solenoid valve output side pressure of the air a, and the time series data of the solenoid valve input side pressure of the air a), a node for inputting the operation time may be added to the input data to obtain a learning model of the neural network from which the time series data of the operating torque of the valve stem 13 (or the time series data of the operating torques of the 2 kinds of valve stems 13) is derived. Since the operating torque of the valve stem 13 varies according to the operating time, the operating time is added to the input data, whereby the operating torque of the valve stem 13 can be estimated more accurately.

Alternatively, instead of the operation time, the count of the counter 48 that counts the number of operations of the solenoid valve 1, the driving device 12, and the main valve 11 may be added to the learning data set in the input data. Since the operating torque of the valve rod 13 changes every time the solenoid valve 1, the driving device 12, and the main valve 11 are operated, the change in the operating torque of the valve rod 13 can be estimated more accurately by adding the value of the counter to the input data.

The running time or the count is preferably adjusted to a value within an appropriate range in the preprocessing in consideration of the balance with the numerical value of the time series data.

Further, the learning data set may be data to which any one of the data acquired from the position sensor 491, the acceleration sensor 492, and the temperature/humidity sensor 493 provided as the driving state sensor 49 is added. By learning the data indicating the driving state as input data, it is possible to estimate more accurate time series data of the operating torque of the valve stem 13 from the driving state. For example, by adding data obtained from the temperature/humidity sensor 493 as input data, it is possible to estimate data in which a change in the operating torque of the valve stem 13 according to the air temperature and humidity is considered. Alternatively, for example, by adding the position of the piston 122A or 122B or the piston rod 121 with respect to the cylinder 120 acquired by the position sensor 491 and the data of the vibration of the driving device 12 acquired by the acceleration sensor 492 as input data, it is possible to estimate data in which a change in the operating torque of the valve rod 13 corresponding to the state of the driving device 12 is taken into consideration.

In the above, the case where the sampling period and the phase of the time series data included in the learning data set are identical has been described, but the learning data set in which any one of the sampling period and the phase is not identical may be used for learning as long as the time series data in the same period is used.

(Data processing System)

Next, an application example of the learned model generated by the machine learning device 200 and the machine learning method described above will be described with reference to fig. 9. FIG. 9 is a schematic block diagram illustrating a data processing system 300 in accordance with an embodiment of the present invention.

As the data processing system 300 of the present embodiment, a system mounted in the microcontroller 70 of the fluid pressure driven valve 10 is exemplified. It should be noted that at least a portion of data processing system 300 can also be applied to other devices, such as external device 15, other devices connected to fluid pressure actuated valve 10.

The data processing system 300 includes at least an input data acquisition unit 301, an inference unit 302, a learned model storage unit 303, an abnormality determination unit 305, and a notification unit 304.

The input data acquisition unit 301 is connected to the plurality of sensors 4 included in the fluid pressure driven valve 10, and is an interface unit for acquiring data output from each sensor 4. The input data acquisition unit 301 acquires at least time series data of the valve opening degree of the main valve 11 and time series data of the solenoid valve output side pressure of the air a in a predetermined period.

In the example shown in fig. 9, all of the input data that can be used for reasoning described later are connected to all of the sensors 4 included in the fluid pressure-driven valve 10, but it is possible to appropriately select which sensor 4 is connected to the input data acquisition means 301 according to a learned model or the like used in the reasoning means 302 described later. The inference result of the inference unit 302 is preferably stored in a storage unit, not shown, and the stored past inference result can be used, for example, as a learning data set for online learning, which is used to further improve the inference accuracy of the learned model in the learned model storage unit 303.

The inference unit 302 is configured to infer time series data of the acting torque of the valve stem 13 from various data of the fluid pressure driven valve 10 acquired by the input data acquisition unit 301. For example, a learned model learned by the above-described machine learning device 200 and machine learning method is used for this reasoning, and the learned model is stored in a learned model storage unit 303 made up of an arbitrary storage medium. The inference unit 302 includes not only a function of performing inference processing using a learned model, but also a preprocessing function of inputting the learned model by adjusting the input data acquired by the input data acquisition unit 301 to a desired form or the like, as preprocessing of the inference processing, and a function of converting an output value outputted by the learned model to a predetermined form, and storing the converted output value.

As described above, the learned model storage unit 303 is a storage medium for storing the learned model used in the inference unit 302. A plurality of learned models may be stored in the learned model storage unit 303 and selectively used by the inference unit 302. The plurality of learned models are generated, for example, when the number of input data, the learning method, the use environment of the fluid pressure-driven valve 10, or the specification of the fluid pressure-driven valve 10 are different each time. The environment in which the fluid pressure-driven valve 10 is used is, for example, the type of driving fluid, the supply pressure, the type of fluid flowing through the pipe 100, the pressure, and the like. The specifications of the fluid pressure driven valve 10 are, for example, the specifications of the solenoid valve 1 (flow rate, response speed, etc.), the specifications of the main valve 11 (size, kind of seal, etc.), the specifications of the driving device 12 (pressure receiving areas of the pistons 122A, 122B, spring constant of the coil spring 123, etc.), and the like.

The abnormality determination unit 305 performs predetermined processing on the time series data of the acting torque of the valve stem 13 inferred by the inference unit 302, and determines whether or not there is an abnormality in at least one of the main valve 11 and the driving device 12 based on the result of the processing. Examples of the abnormality of the main valve 11 include a valve leakage, a valve seat leakage, a ground leakage, and biting of foreign matter in the valve. Examples of the abnormality of the drive device 12 include wear and tear of the parts of the rack and pinion mechanism, more specifically wear and tear of the parts of the rack 125a and pinion 125b, wear and tear of the parts of the scotch yoke mechanism, more specifically wear and tear of the parts of the needle roller 125c and yoke 125d that contact the needle roller 125c, and deterioration of the biasing force of the coil spring 123. The abnormality determination unit 305 can determine not only the presence or absence of occurrence of an abnormality, but also the presence or absence of occurrence of a sign of an abnormality. That is, the abnormality in the present invention includes at least one of an abnormality that has occurred and a sign of an abnormality that will occur in the future.

The predetermined process by the abnormality determination unit 305 includes a statistical process of obtaining a statistical value of time series data of the acting torque inferred by the inference unit 302. Examples of the statistical value of the time series data of the applied torque include, but are not limited to, a maximum value, a minimum value, and a difference between the maximum value and the minimum value of the time series data of the applied torque.

Fig. 10 is a diagram illustrating an example of abnormality determination by the abnormality determination unit 305 in the data processing system 300 according to the embodiment of the present invention. Fig. 10 (a) is a diagram showing an example of a statistical value which is a result of the inference of the time series data of the action torque acting on the valve stem 13 in the predetermined period by the inference unit 302 and the result of the statistical processing of the time series data by the abnormality determination unit 305, and fig. 10 (b) is a diagram illustrating a method of the abnormality determination unit 305 for performing the abnormality determination based on the statistical value.

Fig. 10 (a) shows a maximum value 501, a minimum value 502, and a difference 503 between the maximum value 501 and the minimum value 502 as statistical values that are the results of statistical processing performed on the time-series data of the acting torque by the abnormality determination unit 305 and the time-series data 500 of the acting torque acting on the valve stem 13 in the predetermined period inferred by the inference unit 302. In addition to the abnormality determination unit 305, a median, a frequency-most value, an average value, and the like, which are not shown, may be obtained as statistical values.

The abnormality determination unit 305 determines whether or not at least one of the main valve 11 and the driving device 12 is abnormal with respect to the statistical value obtained by the statistical processing, for example, using a threshold value. Fig. 10 (b) shows an example in which abnormality determination section 305 performs abnormality determination based on whether maximum value 501 is equal to or greater than predetermined threshold 504. The abnormality determination unit 305 may determine different statistical values by performing different statistical processing for each type of abnormality of the main valve 11 and the driving device 12, and may determine an abnormality including at least one of an abnormality occurring and a sign of an abnormality to be occurring in the future by using a corresponding threshold value for each type of abnormality and each different statistical value.

The input data acquisition unit 301 may acquire a plurality of input data having different time periods within a predetermined period. In this case, the inference unit 302 inputs the plurality of input data acquired by the input data acquisition unit 301 to the learned model, and infers time series data of a plurality of applied torques corresponding to the plurality of input data, respectively. Then, abnormality determination section 305 performs statistical processing for obtaining statistical values of the applied torque from time series data of the applied torques, and determines whether or not there is an abnormality from time transitions of the plurality of statistical values obtained by the statistical processing.

Fig. 11 is a diagram illustrating another example of abnormality determination by the abnormality determination unit 305, that is, abnormality determination by statistical processing of time-lapse of applied torque in the data processing system 300 according to an embodiment of the present invention. Fig. 11 (a) is a diagram showing a time course 600 in which the abnormality determination unit 305 performs statistical processing on each of the time series data of the plurality of operating torques acting on the valve stem 13 in the predetermined period inferred by the inference unit 302, and as a result thereof, obtains a plurality of maximum values corresponding to the plurality of time series data, and indicates the statistical values thereof along the time axis, and fig. 11 (b) is a diagram illustrating a method in which the abnormality determination unit 305 performs abnormality determination based on the time course of the statistical values.

In fig. 11 (a), the traces represent statistics obtained by performing statistical processing on time series data of the applied torque corresponding to one input data, and here, as an example, the statistics use the maximum value. Also, the temporal variation of the plurality of statistics is illustrated as a time lapse 600. The horizontal axis of the graph (a) of fig. 11 shows time, but this is an example, and an elapsed time variable such as the number of samples may be used. It is to be noted that the statistical value of the applied torque, which is the result of the statistical processing by the abnormality determination unit 305, is not limited to the maximum value, and may be the minimum value, or the difference between the maximum value and the minimum value.

The abnormality determination unit 305 determines whether or not at least one of the main valve 11 and the driving device 12 is abnormal using, for example, differential values with respect to the time lapse of the statistical value obtained by the statistical processing. Fig. 11 (b) shows an abnormality determination method in which the abnormality determination unit 305 performs abnormality determination based on whether or not the differential value of the statistic value in the predetermined specified period 601 is equal to or greater than the predetermined variation threshold 604, and here, the maximum value is used as an example of the statistic value. The differential value of the statistic value in the specified period 601 is obtained, for example, from the difference 602 between the statistic value at the predetermined time and the statistic value at the time when the specified period 601 has elapsed from the predetermined time. The abnormality determination unit 305 may calculate different statistical values by performing different statistical processing for each abnormality type of the main valve 11 and the driving device 12, or may use a variation threshold 604 corresponding to each abnormality type and each different statistical value. The amount of change in the specified period 601 is not limited to the differential value, and may be an amount of change such as an average change rate.

The notification unit 304 is configured to notify an operator or the like of the inference result of the inference unit 302 and the abnormality determination result of the abnormality determination unit 305. Various specific notifying means can be employed, and for example, the inference result can be transmitted to the external device 15 via the communication unit 8, and displayed graphically on the GUI of the external device 15. The inference result of the inference unit 302 may be displayed graphically in real time or afterwards, or may be stored in any data form in a storage device provided in the data processing system 300 or the external device 15. Similarly, the abnormality determination result of the abnormality determination unit 305 may be displayed in real time or afterwards, or may be stored in any data form in a storage device provided in the data processing system 300 or the external device 15.

(Data processing method)

A data processing method of the data processing system having the above configuration will be described with reference to fig. 6 (reasoning stage) and fig. 12. Fig. 12 is a flowchart showing an example of a data processing method of the data processing system 300 of an embodiment of the present invention.

When the supply of electric power from the external power source 16 to the solenoid valve 1 of the fluid pressure driven valve 10 is started and the data processing process of the data processing system 300 is started, the input data acquisition unit 301 acquires various data indicating the state of each part of the fluid pressure driven valve 10, including at least the time series data of the valve opening degree of the main valve 11 and the time series data of the solenoid valve output side pressure of the air a in the predetermined period acquired by the plurality of sensors 4 (step S21).

At the time when the input data acquisition unit 301 can acquire desired input data (time series data of the valve opening degree of the main valve 11 in a predetermined period and time series data of the solenoid valve output side pressure of the air a (see fig. 6)), the inference unit 302 performs inference based on the input data (step S22). In this case, a learning model for reasoning is preferably predetermined. In addition, when the learning model determined in advance requires predetermined time series data as input data, for example, the input data acquisition unit 301 acquires the necessary data amount, and then performs reasoning in step S22.

Specifically, when the input data is preprocessed and input to the learned model, the inference unit 302 outputs time series data of the acting torque acting on the valve stem 13 as the inference result.

In step S22, the abnormality determination unit 305 performs predetermined processing on the result of the inference performed by the inference unit 302, and determines abnormality of the main valve 11 and the driving device 12 based on the result of the processing. (step S23).

In step S23, the result of the abnormality determination by the abnormality determination unit 305 is notified to the operator or the like by means of mail or the like by the notification unit 304, and the operator or the like can confirm whether or not at least one of the main valve 11 and the driving device 12 is abnormal (step S24).

(Inference means)

The present invention is provided not only in the manner of the data processing system 300 described above, but also in the manner of an inference means for performing inferences. In this case, the inference means comprises a memory and at least one processor, wherein the processor is capable of executing a series of processes. The series of processes includes: a process of acquiring time series data of the valve opening degree of the main valve 11 and time series data of the solenoid valve output side pressure of the air a in a predetermined period, and a process of deducing time series data of the acting torque acting on the valve rod 13 in the predetermined period when the input data is inputted.

By providing the present invention in the manner of the above-described inference means, it can be simply applied to various fluid pressure driven valves 10, as compared with the case where the data processing system 300 is installed. At this time, it should be understood by those skilled in the art that the inference method implemented by the inference unit of the data processing system using the learned model learned by the machine learning device and the machine learning method in the present invention as described hereinabove in the present specification may be applied when the inference device performs the process of inferring the time-series data of the applied torque.

In the above description, the case where the neural network model is used to estimate the time series data of the acting torque of the valve stem 13 has been described, but the time series data of the acting torque of the valve stem 13 may be estimated based on the correlation between the time series data of the valve opening degree of the main valve 11 and the time series data of the solenoid valve output side pressure of the air a stored in the learning data set storage unit 202 and the time series data of the acting torque of the valve stem 13, by taking the valve opening degree of the main valve 11 and the solenoid valve output side pressure of the air a as functions of variables. Alternatively, if the learning data set storage unit 202 stores time series data of the valve opening degree of the main valve 11, time series data of the solenoid valve output side pressure of the air a, and time series data of the solenoid valve input side pressure of the air a, time series data of the acting torque of the valve rod 13 may be estimated by using the correlation between these time series data and time series data of the acting torque of the valve rod 13 as a function of variables, the valve opening degree of the main valve 11, the solenoid valve output side pressure of the air a, and the solenoid valve input side pressure of the air a.

Further, although the case where the computers constituting the machine learning device 200, the data processing system 300, the microcontroller 70, and the like have CPUs and GPUs has been described, a dedicated logic circuit such as a programmable logic device (programmable logic device: PLD) such as an FPGA (Field-programmable GATE ARRAY) or an ASIC (Application SPECIFIC IC) may be provided as the processor. Similarly, the inference device may be configured to provide a programmable logic device such as an FPGA or a dedicated logic circuit such as an ASIC as a processor.

The present invention is not limited to the above-described embodiments, and can be variously modified and implemented within a scope not departing from the gist of the present invention. Moreover, they are all included in the technical idea of the present invention.

Description of the reference numerals

A solenoid valve 1, a solenoid portion 3, a main valve 4, a sensor 4, a fluid pressure driving valve 10, a main valve 11, a driving device 12 (fluid pressure), a valve stem 13, a first shaft 13a, a second shaft 13B, a third shaft 13c, a supply source 14 of air (driving fluid supply source), an external device 15, a test device 17, a flow path 26 input side, a flow path 27 output side, a flow path 28 exhaust side, a solenoid case 30, a solenoid coil 31, a movable iron core 32, a first pressure sensor 40, a second pressure sensor 41, a main valve opening sensor 42, a voltage sensor 43, a current resistance sensor 44, a temperature sensor 45, a magnetic sensor 46, an operation timer 47 (timer), an operation counter 48 (counter), a driving state sensor 49, a microcontroller 70, a pipe 100, a first torque sensor 170A, a second torque sensor 170B, a machine learning device 200, a learning data set acquisition unit 201, a learning data set storage unit 202, a learning model storage unit 203, 204 learning model storage unit 300 data processing system 300, input data acquisition unit 301, 304, learning storage unit 305, calculation model determination unit 305, and an abnormality calculation unit 1 (PC).

Claims

1.A data processing system including at least a main valve, a driving device for driving a valve stem connected to the main valve, and a fluid pressure driving valve for controlling a solenoid valve for supplying and discharging a driving fluid to and from the driving device, the data processing system comprising:

an input data acquisition unit that acquires input data including at least time series data of a valve opening degree of the main valve in the predetermined period and time series data of a solenoid valve output side pressure of the driving fluid supplied from the solenoid valve to the driving device in the predetermined period;

2. The data processing system of claim 1, wherein the data processing system further comprises a data processing system,

The abnormality determination means performs a statistical process of obtaining a statistical value of the acting torque based on the time series data of the acting torque inferred by the inference means, and determines whether or not the abnormality is present based on the statistical value obtained by the statistical process as the process.

3. The data processing system of claim 1, wherein the data processing system further comprises a data processing system,

The input data acquisition means acquires a plurality of input data having different time periods in the predetermined period,

The inference unit inputs the plurality of input data acquired by the input data acquisition unit to the learned model, infers time series data of the plurality of acting torques corresponding to the plurality of input data,

The abnormality determination means performs statistical processing for obtaining statistical values of the acting torque based on the time series data of the plurality of acting torques inferred by the inference means, and determines whether or not the abnormality is present based on time transitions of the plurality of statistical values obtained by the statistical processing.

4. A data processing system according to any one of claims 1 to 3, characterized in that,

The predetermined period is any one of an opening operation period from a start of an opening operation to an end of the opening operation of the main valve, a closing operation period from a start of a closing operation to an end of the closing operation of the main valve, and an opening and closing operation period including the opening operation period and the closing operation period, among periods in which the opening and closing operation of the main valve is performed.

5. A data processing system according to any one of claims 1 to 4, characterized in that,

The applied torque timing data is at least one of timing data of a first applied torque and timing data of a second applied torque,

The first acting torque acts between a first shaft forming part of the valve stem and driven by the driving means and a second shaft forming part of the valve stem and coupled to the first shaft and the main valve,

The second acting torque acts between the first shaft and a third shaft that forms a part of the valve stem, is coupled to the first shaft, and is inserted into a housing portion of the solenoid valve.

6. A data processing system according to any one of claims 1 to 5, characterized in that,

The fluid pressure driven valve further includes a timer for detecting an operation time of the fluid pressure driven valve,

The input data also includes the run time.

7. The data processing system of any one of claims 1 to 6, wherein,

The fluid pressure driven valve further includes a counter for detecting the number of times the main valve, the driving device, or the solenoid valve is operated,

The input data also includes the number of jobs.

8. The data processing system according to any one of claims 1 to 7, wherein,

The fluid pressure driven valve is connected to a driving fluid supply source for supplying the driving fluid to the solenoid valve,

The input data further includes time series data of a solenoid valve input side pressure of the driving fluid supplied from the driving fluid supply source to the solenoid valve during the predetermined period.

9. A data processing method for a fluid pressure driven valve including at least a main valve, a driving device that drives a valve stem connected to the main valve, and a solenoid valve that controls supply and discharge of a driving fluid to and from the driving device, the method comprising:

an input data acquisition step of acquiring input data including at least time series data of a valve opening degree of the main valve in the predetermined period and time series data of a solenoid valve output side pressure of the driving fluid supplied from the solenoid valve to the driving device in the predetermined period;

An inference step of inferring, in the learning model in which a correlation between the input data and output data including time series data of the acting torque acting on the valve stem in the predetermined period corresponding to the input data, the time series data of the acting torque acting on the valve stem in the predetermined period corresponding to the input data, by inputting the input data acquired in the input data acquisition step, the learning model having learned the correlation between the input data and the output data by machine learning; and

And an abnormality determination step of performing predetermined processing on the time-series data of the applied torque inferred by the inference step, and determining whether or not at least one of the main valve and the driving device is abnormal based on a result of the processing.