Machine learning device, pressure tank state estimation device, inference device, machine learning method, machine learning program, pressure tank state estimation method, pressure tank state estimation program, inference method, and inference program

The machine learning device uses pump and motor data to estimate pressure tank condition, addressing the limitations of existing systems by improving accuracy and reducing sensor reliance.

JP7588522B2Active Publication Date: 2024-11-22EBARA CORP
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
JP2021021740
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2021-02-15
Publication Date
2024-11-22
Estimated Expiration
2041-02-15

AI Technical Summary

Technical Problem

Existing pressure tank diagnostics in water supply systems rely on expensive sensors and operator experience, lacking accuracy and automation, and cannot handle complex events or multiple physical quantities effectively.

Method used

A machine learning device that estimates pressure tank state using pump pressure data, motor current, and vibration data without additional sensors, employing a learning model trained on historical data to infer tank condition accurately.

Benefits of technology

Accurately estimates pressure tank condition without relying on expensive sensors or operator experience, providing high-accuracy diagnostics and reducing installation costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007588522000001
    Figure 0007588522000001
  • Figure 0007588522000002
    Figure 0007588522000002
  • Figure 0007588522000003
    Figure 0007588522000003
Patent Text Reader

Abstract

To provide a machine learning device highly accurately estimating a state of a pressure tank of a fluid supply device without depending on experience of a worker.SOLUTION: A machine learning device 4 generates a learning model for estimating sealing pressure of a pressure tank in a fluid supply device 1 having a pump for transferring fluid and the pressure tank provided on a discharge side of the pump. The machine learning device includes: a learning data storage section 401 storing a plurality of sets of learning data that include pump pressure data related to the pump for a predetermined period and data on an operating state of the pump for predetermined period as input data and data on sealing pressure of the pressure tank as output data, respectively and in which the input data and the output data are associated with each other; a machine learning section 402 causing the learning model to learn correlation between the input data and the output data by inputting the plurality of sets of learning data to the learning model; and a learned model storage section 403 storing the learning model learned by the machine learning section.SELECTED DRAWING: Figure 3
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a machine learning device, a pressure tank state estimation device, an inference device, a machine learning method, and a mechanical learning program, pressure tank state estimation method, pressure tank state estimation program, inference method, and and inference programs. [Background technology]

[0002] A device for supplying water (tap water) to water supply targets such as office buildings or apartment buildings. Water supply devices are widely used for water supply. Water supply devices generally consist of a pump for pumping water. a motor for driving the pump, a control unit for controlling the operation of the pump, and a and a pressure tank connected thereto.

[0003] Generally, a pressure tank has a diaphragm inside. The diaphragm separates the inside of the container into an air chamber and a water chamber. Air is sealed in the air chamber beforehand, and water that flows into the water chamber from the pump passes through the diaphragm. The air is compressed through a check valve between the pressure tank and the pump. The stop valve and pressure tank ensure that the water pressure in the water supply pipes is maintained properly even after the pump is stopped. When water is sent to the destination, the diaphragm moves from the air chamber side to the water chamber side. Protrude to.

[0004] The diaphragm of such a water supply device expands and contracts over the years of pump operation or water supply to the destination. Repeated expansion and contraction of the diaphragm gradually deteriorates it. Also, the fixation of the diaphragm to the container The fixed part will gradually deteriorate due to repeated bending. However, the deterioration of the diaphragm will occur as the Since the time required for replacement varies depending on the usage situation, it was difficult to determine the optimal timing for replacement. This is a system that can accurately diagnose the deterioration state of the pressure tank membrane and determine the appropriate replacement time. A membrane inspection device has been proposed (see, for example, Patent Documents 1 and 2). [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent No. 3596909 [Patent Document 2] Japanese Patent Application Publication No. 8-86278 Summary of the Invention [Problem to be solved by the invention]

[0006] In the devices disclosed in Patent Document 1 and Patent Document 2, individual physical quantities (pressure, motor The condition of the pressure tank is diagnosed based on the following: Therefore, the change in the state of the pressure tank may be expressed in physical quantities other than those mentioned above, or may be compounded into multiple physical quantities. In cases where the image is expressed in a negative manner, the devices disclosed in Patent Documents 1 and 2 are In addition, diagnosis of complex phenomena based on multiple physical quantities was not possible. This depends on the experience (including tacit knowledge) of the worker and on the individual differences of the worker. There is a strong demand for a device that can automate such diagnostics.

[0007] Furthermore, the devices disclosed in Patent Documents 1 and 2 each reduce the pressure in the pressure tank. However, the water supply device has a pressure sensor to measure the pressure. The pressure in the pressure tank is not used in the maintenance of the water supply system. In order to use this method, it was necessary to install an expensive pressure sensor.

[0008] In view of the above-mentioned problems, the present invention is directed to a method for detecting a temperature rise without relying on the experience of an operator and without installing expensive sensors. Therefore, it is possible to estimate the state of the pressure tank in the fluid supply device with high accuracy. Machine learning device, pressure tank state estimation device, inference device, machine learning method, machine learning program , pressure tank state estimation method, pressure tank state estimation program, inference method, and inference program The aim is to provide grams. [Means for solving the problem]

[0009] In order to achieve the above object, a machine learning device according to an aspect of the present invention comprises: A fluid supply system including a pump for transporting a fluid and a pressure tank provided on the discharge side of the pump. A machine learning method for generating a learning model for estimating the charged pressure of the pressure tank in the supply system. A learning device comprising: Pump pressure data for the pump for a predetermined period of time, and The data on the operation status of the pump is used as input data, and the data on the charged pressure of the pressure tank is used as input data. and a learning method in which the input data and the output data are associated with each other. A learning data storage unit that stores a plurality of sets of learning data; By inputting a plurality of sets of the learning data to the learning model, the input data and the output data are a machine learning unit that causes the learning model to learn a correlation with force data; A trained model storage unit that stores the learning model trained by the machine learning unit; Equipped with. Effect of the Invention

[0010] According to the machine learning device of the present invention, data on motor current values, data on contact electrical resistance, and This is a learning model that can infer (estimate) information about pressure tanks with high accuracy from vibration data. In addition, the pressure tank state estimating device of the present invention can provide a The state of the pressure tank can be estimated with high accuracy without relying on previous experience.

[0011] Other objects, configurations and effects will become apparent from the detailed description of the invention described below. can be. [Brief description of the drawings]

[0012] [Figure 1] 1 is a diagram showing an example of a water supply device 1 to which a pressure tank state estimation system 2 according to an embodiment of the present invention is applied. [Diagram 2] A hardware configuration diagram showing an example of a computer 200 that constitutes the machine learning device 4 and the pressure tank state estimation device 5. [Diagram 3] FIG. 2 is a block diagram showing an example of a machine learning device 4 according to the present embodiment. [Figure 4] FIG. 2 is a data configuration diagram showing an example of data (supervised learning) used in the machine learning device 4 according to the present embodiment. [Diagram 5] FIG. 2 is a schematic diagram showing an example of a neural network model used in the machine learning device 4 according to the present embodiment. [Figure 6] 4 is a flowchart showing an example of a machine learning method performed by the machine learning device 4 according to the present embodiment. [Figure 7] FIG. 2 is a block diagram showing an example of a pressure tank state estimating device 5 according to the present embodiment. [Figure 8] 4 is a flowchart showing an example of a pressure tank state estimation method performed by the pressure tank state estimation device 5 according to the present embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The scope necessary for the description of the present invention to achieve the object of the present invention is shown in a schematic manner, and the relevant part of the present invention is explained. The following description will focus on the scope necessary for clarity, and the parts that are omitted will be based on the prior art. It shall be so.

[0014] FIG. 1 is a diagram showing a fluid supply system to which a pressure tank state estimation system 2 according to an embodiment of the present invention is applied. The fluid supply device 1 is mainly used in apartment buildings, offices, commercial facilities, factories, schools, etc. It is a device for supplying water to buildings, houses, parks, and other facilities (water supply targets). The water supply device 1 introduces water from a fluid supply source 41 through an inlet port via an introduction pipe 42. The fluid supply source 41 may be a water tank type using a water tank. The water supply pipe 43 is connected to a fluid supply destination 44 such as a faucet in each building. The fluid supply device 1 pressurizes water from a fluid supply source 41 and supplies it to each fluid supply destination 44 such as a building. Provide water.

[0015] In addition, the fluid supply device 1 of the present embodiment is illustrated as a water supply device that supplies water. The supply device 1 supplies liquids other than water, such as fuel and chemicals, as well as gases, such as hydrogen and oxygen. The fluid supply device may be one that supplies a fluid containing the fluid.

[0016] The fluid supply device 1 includes a pump 12 and a motor 1 as a drive unit for driving the pump 12. The inverter 14 controls the motor 13. The motor 14 is connected to a commercial power source 51 via a ground fault circuit interrupter 52 .

[0017] The fluid supply device 1 includes a check valve 15 and a flow switch 16 on the discharge side (downstream side) of the pump 12. The fluid supply device 1 shown in FIG. The fluid supply device 1 includes a motor 13, a check valve 15, and a flow switch 16. Even if the system has two or more sets of pumps 12, motors 13, check valves 15, and flow switches 16, The fluid supply device 1 includes a plurality of pumps 12, so that some of the pumps 12 can be In the event that one pump becomes inoperable, water supply can be continued using another pump 12 that is operable, thereby preventing water outages. This can be done.

[0018] The fluid supply device 1 also includes a backflow prevention device 11 upstream of the pump 12. The stop device 11 is provided in an introduction pipe 42 connected to the suction port of the fluid supply device 1, and the fluid The backflow of water from the supply device 1 to the fluid supply source 41 is prevented. In such a case, the fluid supplying device 1 does not need to include the backflow prevention device 11.

[0019] The fluid supplying device 1 has a check valve 15 in a discharge pipe connected to the discharge port of the pump 12. The stop valve 15 prevents water from flowing back to the pump 12 when the pump 12 stops. The body supply device 1 has a flow switch 16 downstream (secondary side) of the check valve 15. The flow switch 16 detects when the flow rate of water flowing through the discharge pipe drops to a predetermined value. Furthermore, the fluid supply device 1 includes a pressure tank downstream of the flow switch 16. The pressure tank 17 stores the discharge of the fluid supply device 1 while the pump 12 is stopped. The pressure tank is a pressure retainer that maintains the pressure on the outlet side. The pressure and the fill pressure of the pressure tank both refer to the pressure inside the air chamber.

[0020] The pressure tank 17 according to this embodiment has a diaphragm 17a that divides the inside of the container into an air chamber and a water chamber. The air chamber is filled with air in advance, and the water that flows into the water chamber from the pump 12 is The air is compressed through the diaphragm 17a. A check valve 1 is provided between the pressure tank 17 and the pump 12. 5 is provided, and this check valve 15 and the pressure tank 17 ensure that when the pump 12 stops, Even after the water supply, the water pressure in the water supply pipe is appropriately maintained, and water is sent to the fluid supply destination 44. When water is sent to the supply destination 44, the diaphragm 17a presses from the air chamber side to the water chamber side.

[0021] The fluid supplying device 1 includes a control unit 20 that controls the water supplying operation. If a circuit board with a central processing unit (Central Processing Unit) or a dedicated circuit board is used, The control unit 20 is supplied with power from a commercial power source 51 via a power line (not shown). However, the control unit 20 controls the flow rate of the fluid supplying device 1 through other components such as the inverter 14. Electrical power may be provided.

[0022] The control unit 20 and the inverter 14 are connected to each other through a communication line 21. The control unit 20 sends control commands to the inverter 14 regarding various setting values ​​and the rotation speed of the motor 13. The inverter 14 sends signals such as a frequency command value, an on / off signal, etc. to the control unit. The actual operating conditions such as frequency value and current value are sequentially sent to 20.

[0023] The control signal transmitted between the control unit 20 and the inverter 14 is an analog signal. For example, analog signals may be used for the rotation frequency, etc. A signal can be used for an operation stop command, and a digital signal can be used for an operation stop command, etc.

[0024] A pressure tank for estimating the state of the pressure tank 17 for the fluid supply device 1 having the above configuration. A state estimation system 2 is provided. The pressure tank state estimation system 2 includes a measuring device 3 and The system is equipped with a machine learning device 4 and a pressure tank state estimation device 5.

[0025] The measuring device 3 is installed in each part of the fluid supplying device 1 and measures the physical quantity and state of each part of the fluid supplying device 1. In the learning phase, the measurement device 3 is connected to the machine learning device 4 to measure the state of the object. In the inference phase, the pressure tank state estimation device 5 is connected to the inference The measurement device 3 is used to measure data for the measurement. The pressure sensor 32, the pump operation status sensor 33, and the usage flow rate sensor 34 are provided. The pump primary pressure sensor 31 and the pump secondary pressure sensor 32 are pumps that measure the pressure related to the pump. At least one of the flow rate sensors 3 may be provided as the pressure sensor. Item 4 does not need to be provided.

[0026] The pump primary pressure sensor 31 is a pressure measuring device for measuring the suction side pressure of the pump 12. The pump primary pressure sensor 31 of this embodiment is disposed between the fluid supply source 41 and the backflow prevention device 11. The pump primary pressure sensor 31 is capable of measuring time series data for a predetermined period of time. Preferably, the predetermined period of time may include a start-up time point when the pump is started.

[0027] The pump secondary pressure sensor 32 is a pressure measuring device for measuring the pressure on the discharge side of the pump 12. The pump secondary pressure sensor 32 of this embodiment is connected to the flow switch 16 and the pressure tank 17. The pump secondary pressure sensor 32 is capable of measuring time series data for a predetermined period of time. Preferably, the predetermined period of time may include a start-up time when the pump is started.

[0028] The pump operation status sensor 33 acquires the operation status of the pump 12. The conditions include, for example, ON / OFF of the pump 12, the rotation speed of the pump 12, the current value of the motor 13, etc. The pump operation status sensor 33 can measure time series data for a predetermined period of time. Preferably, the predetermined period of time may include the start-up time when the pump is started. By acquiring the operating status of the pump 12, the amount of fluid discharged from the pump 12 can be estimated. do.

[0029] The pump operation status sensor 33 detects the operation of the pump 12, the motor 13, or the inverter 14. It is preferable to be able to obtain time series data of the operating state indicating whether the vehicle is in operation or stopped. The operation state is determined by ON / OFF control from the control unit 20 to the pump 12, the motor 13, or the inverter 14. The FF command is acquired. When the command of the control unit 20 is ON, the operation state is in operation, and when the command of the control unit 20 is ON, the operation state is O. In the case of FF, it is sufficient to recognize that the vehicle is stopped.

[0030] In addition, the pump operation status sensor 33 may acquire time series data of the rotation speed of the pump 12. The rotation speed of the pump 12 can be obtained by measuring the rotation speed of the pump 12 or the motor 13. In addition, the rotation speed of the motor 13 is determined by the output frequency of the inverter 14. , the output frequency of the inverter 14 may be obtained.

[0031] Furthermore, the pump operation status sensor 33 is connected to the motor 13 which is the driving source of the pump 12. By acquiring the motor current value, the power supply current can be calculated. The amount of work done by the pump 12 can be obtained. It is also possible to recognize that the flow rate is excessive.

[0032] The usage flow rate sensor 34 is a flow rate sensor for measuring the amount of fluid supplied from the pressure tank 17 during a predetermined period of time. The flow rate of the fluid used at the supply destination 44 is measured. When there are multiple fluid supply destinations 44, The usage flow rate sensor 34 measures the total usage flow rate obtained by adding up the flow rates of the multiple fluid supply destinations 44. It is preferable that the flow rate can be measured using time series data. The source 34 may be provided on the upstream side where the fluid is supplied to a plurality of fluid destinations 44, as shown in FIG. The usage flow rate sensor 34 detects the fluid supply destination of each room in a collective housing unit such as an apartment or condominium. The usage flow rate sensor 34 may be provided in each room as a water meter 44. The predetermined period of time may include a start-up time when the pump is started.

[0033] The usage flow rate sensor 34 may estimate the usage flow rate from other measurement data instead of directly measuring the usage flow rate. For example, The entry and exit records of people at a plurality of fluid supply destinations 44 are measured, and the flow rate used is calculated from the time series data. In addition, sensors can be installed in the toilets of each room, the showers in the bathroom, the faucets in the kitchen, etc. The amount of fluid used may be estimated from the amount of fluid used. The usage flow rate of the fluid may be estimated from the usage rate of the gas.

[0034] The measuring device 3 measures the physical quantities and state quantities of each part at a predetermined measurement period, for example. Each time time passes, the measured value at that time is input to the machine learning device 4 and the pressure tank state estimation device The measurement device 3 is configured to be able to output the pressure to the pump primary pressure sensor 31 and the pump The pump secondary pressure sensor 32, the pump operation status sensor 33, and the usage flow rate sensor 34 each The measurement periods may be the same or different. Instead of outputting a discrete measurement value every time the measurement period elapses, the The device may output continuous measurements as follows:

[0035] The machine learning device 4 operates as a subject of the learning phase and estimates the state of the pressure tank 17. The machine learning device 4 generates a learning model 6 used in the process of the machine learning. As a method for this, "supervised learning" can be adopted.

[0036] The pressure tank state estimation device 5 operates as the subject of the inference phase and performs the machine learning A trained learning model 6 is used to generate a trained learning model 6 based on the measured values ​​obtained by the measuring device 3. The state of the pressure tank 17 is estimated. The trained learning model 6 is arbitrarily input from the machine learning device 4. The pressure tank state estimation device 5 receives the pressure tank state information via a communication network, a recording medium, or the like.

[0037] FIG. 2 is a diagram showing a computer 200 constituting the machine learning device 4 and the pressure tank state estimation device 5. FIG. 1 is a hardware configuration diagram showing an example of the

[0038] Each of the machine learning device 4 and the pressure tank state estimation device 5 is a general-purpose or dedicated computer. As shown in FIG. 2, the computer 200 is mainly composed of The components include a bus 210, a processor 212, a memory 214, an input device 216, a display display device 218, storage device 220, communication I / F (interface) unit 222, An external device I / F unit 224, an I / O (input / output) device I / F unit 226, and a media input The above components are included in the computer 200 depending on the application. It may be omitted as appropriate.

[0039] The processor 212 may be one or more processing units (CPU, MPU, GPU, DSP, etc.). The memory 21 is made up of a plurality of memory units 21a, 21b, 21c, 21d, 21e, 21f, 21g, 21h, 21h, 21i, 21m ... 4 stores various data and programs 230 and functions as, for example, a main memory. There are two types of memory: volatile memory (DRAM, SRAM, etc.) that can be used with a single chip, and non-volatile memory (ROM, flash memory, etc.) that can be used with a single chip. It consists of the following:

[0040] The input device 216 is composed of, for example, a keyboard, a mouse, a numeric keypad, an electronic pen, etc. The display device 218 may be, for example, a liquid crystal display, an organic EL display, or an electronic The input device 216 and the display device 218 are composed of paper, a projector, etc. The storage device may be integrally configured with the display, such as a touch panel display. 220 is composed of, for example, a HDD, SSD, etc., and contains an operating system and programs. It stores various data necessary for the execution of program 230.

[0041] The communication I / F unit 222 is connected to a network 240 such as the Internet or an intranet. It is connected by wire or wirelessly and transmits data between other computers according to a specified communication standard. The external device I / F unit 224 is connected to an external device 250 such as a printer or a scanner. The device is connected by wire or wirelessly, and data is transmitted between the device and the external device 250 in accordance with a predetermined communication standard. The means of wireless communication typically conforms to international standards. As communication means, IEEE802.15.4, IEEE802.15.1, IEEE8 02.15.11a, 11b, 11g, 11n, 11ac, 11ad, ISO / IEC1 IEEE 802.15.4g and 4513-3-10. ooth (registered trademark), Bluetooth Low Energy, Wi-Fi, ZigB ee (registered trademark), Sub-GHz, EnOcean (registered trademark), etc. can also be used. The I / O device I / F unit 226 is a unit for connecting various I / O devices such as sensors and actuators. The I / O device 260 is connected to the I / O device 260, and, for example, a detection signal from a sensor or Sends and receives various signals and data, such as control signals to the actuator. Media I / O section 228 is composed of a drive device such as a DVD drive, a CD drive, etc. The optical disc drive 240 reads and writes data from and to a medium 270 such as a CD.

[0042] In the computer 200 having the above configuration, the processor 212 executes the program 23. 0 is loaded into the work memory area of ​​the memory 214 and executed, and the program is then transmitted to the computer via the bus 210. The program 230 controls each part of the computer 200 instead of the memory 214. The program 230 may be installed in the storage device 220. Recorded in file format or executable file format on non-transient recording media such as CDs and DVDs and provided to the computer 200 via the media input / output unit 228. The RAM 230 downloads the data via the network 240 through the communication I / F unit 222. The information may be provided to the computer 200 by

[0043] The computer 200 may be, for example, a desktop computer or a portable computer. The computer 200 is a client-type computer. Alternatively, a server-type computer or a cloud-type computer may be used. 0 may be applied to devices other than the machine learning device 4 and the pressure tank state estimation device 5. .

[0044] (Machine Learning Device 4) FIG. 3 is a block diagram showing an example of the machine learning device 4 according to the first embodiment.

[0045] The machine learning device 4 includes a learning data acquisition unit 400, a learning data storage unit 401, and a machine The machine learning device 4 includes a learning unit 402 and a trained model storage unit 403. The machine learning device 4 includes, for example, The learning data acquisition unit 400 is configured as the computer 200 shown in FIG. The machine learning unit 402 is configured by the communication I / F unit 222 or the I / O device I / F unit 226. , a processor 212, a learning data storage unit 401 and a learned model storage unit 4 03 is composed of a storage device 220.

[0046] The learning data acquisition unit 400 is connected to various external devices via the communication network 300 and receives input. The external device is an interface unit for acquiring learning data including at least the data. The device may be, for example, a measuring device 3 provided in the fluid supply device 1, a test device simulating the fluid supply device 1, These are the measuring device 3 provided in the device 7, and the worker terminal 8 used by the worker. When the pressure tank state estimation device 5 is connected to the communication network 300, the learning data The acquisition unit 400 is provided in the fluid supply device 1 that is the estimation target of the pressure tank state estimation device 5. Learning data may also be acquired from the measurement device 3.

[0047] The test device 7 includes a pump primary pressure sensor 31, a pump secondary pressure sensor 32, and a The test equipment includes a test sensor 32, a pump operation status sensor 33, and a usage flow rate sensor 34. The test device 7 has a pressure sensor 35 in the pressure tank 17 for measuring the pressure inside the pressure tank. The pressure sensor 35 in the pressure tank is too expensive to install in every fluid supply device 1. Therefore, a pressure sensor 35 in the pressure tank is used in the test device 7, and the data By applying this to the fluid supply device 1 that does not have a pressure sensor 35 in the pressure tank, Estimate the pressure in the tank.

[0048] The learning data storage unit 401 stores the learning data acquired by the learning data acquisition unit 400 in a The learning data storage unit 401 is a database that stores multiple sets of learning data. The specific configuration of the database can be designed as appropriate.

[0049] The machine learning unit 402 performs machine learning using the learning data stored in the learning data storage unit 401. That is, the machine learning unit 402 inputs a plurality of pieces of learning data to the learning model 6. By inputting a set of input data, the input data included in the learning data and the estimated information of the fluid supplying device 1 are By having the learning model 6 learn the correlation, a trained learning model 6 is generated. In this embodiment, as a specific method of supervised learning by the machine learning unit 402, a neural network is used. This section explains the case where a network is adopted.

[0050] The trained model storage unit 403 stores the trained model generated by the machine learning unit 402. The trained model storage unit 403 stores trained models 6. The learning model 6 is transmitted to a real system (e.g., a pressure tank) via any communication network or recording medium. In FIG. 3, a learning data storage unit 401 and a learning Although the completed model storage unit 402 and the completed model storage unit 403 are shown as separate storage units, they are actually a single storage unit. It may be composed of:

[0051] FIG. 4 is an example of data (supervised learning) used in the machine learning device 4 according to this embodiment. FIG.

[0052] The learning data is input data, which is the primary pressure data on the suction side of the pump 12 for a predetermined period. data on the secondary pressure of the discharge side of the pump 12 for a predetermined period of time; data on the secondary pressure of the discharge side of the pump 12 for a predetermined period of time; Data on the operating conditions and data on the flow rate of the fluid used at the fluid supply destination 44 are collected. The input data includes the primary pressure data on the suction side of the pump 12, the secondary pressure data on the discharge side of the pump 12, The pressure data, the operating status data of the pump 12, and the data of the fluid used at the fluid supply destination 44. The predetermined period for each of the flow rate data is basically set to the same period. If a time lag is found between the data, the period is set to the period corresponding to the time lag. The input data may include the primary pressure data on the suction side of the pump 12, the discharge The secondary pressure data, the operating status data of the pump 12, and the data of the fluid used at the fluid supply destination 44 In addition to the data on the flow rate of the fluid, other data may also be included. For example, The sound generated during operation, the temperature of at least one of the pump 12 and the motor 13, the pump 1 2. Lubrication pressure, lubricant flow rate, lubricant temperature, and lubricant property values ​​(color, particle size distribution) The information may further include at least one of the data.

[0053] The primary pressure data on the suction side of the pump 12 during a predetermined period is collected at multiple measurement points during the predetermined period. A plurality of measured values ​​(pump suction side primary pressure values) are measured by the pump primary pressure sensor 31 in the The primary pressure data on the suction side of the pump 12 is composed of, for example, a measurement time indicating the time of measurement. The data is linked to the measurement time and arranged in chronological order. When the pump primary pressure sensor 31 outputs continuous measured values, the learning data is acquired. The measurement value is converted into data at a predetermined sampling period by the unit 400. The primary pressure data on the suction side of the pump 12 is collected at multiple sampling points within a predetermined period. It is composed of multiple measured values ​​(primary pressure values ​​on the pump suction side)

[0054] The discharge side secondary pressure data of the pump 12 during a predetermined period is collected at a plurality of measurement points during the predetermined period. A plurality of measurement values ​​(pump discharge side secondary pressure values) measured by the pump secondary pressure sensor 32 in The discharge side secondary pressure data of the pump 12 is composed of, for example, a measurement time indicating the time of measurement. The data is linked to the measurement time and arranged in chronological order. When the pump secondary pressure sensor 32 outputs continuous measured values, the learning data is acquired. The measurement value is converted into data at a predetermined sampling period by the unit 400. The discharge side secondary pressure data of the pump 12 in the above is obtained at multiple sampling points within a predetermined period. It is composed of multiple measured values ​​(pump discharge side secondary pressure values) at the same time.

[0055] The data on the operation status of the pump 12 during a predetermined period is stored as data on the operation state indicating either one of the above, data on the number of revolutions of the pump 12 during a predetermined period, and The data of the motor current value supplied to the motor 13 which is the driving source of the pump 12 during a predetermined period of time. In this embodiment, the data on the operation status of the pump 12 is preferably at least one of the following: Data on the operation state of the pump 12, indicating whether it is operating or stopped, data on the pump 12 during a predetermined period, The data on the number of revolutions of the pump 12 and the motor 1 which is the driving source of the pump 12 during a predetermined period of time. In the following description, the data is assumed to consist of motor current value data supplied to 3.

[0056] The operation status data indicating whether the pump 12 is operating or stopped during a predetermined period is At multiple measurement points within a predetermined period, multiple measurements are made by the pump operation status sensor 33. The data on the operating state is, for example, The data is linked to the measurement time and arranged in chronological order by measurement time. In addition, when the pump operation status sensor 33 outputs continuous measured values, the learning The measurement data acquisition unit 400 converts the measurement values ​​into data at a predetermined sampling period. The data on the operating state during a given period is collected at multiple sampling points during the given period. It consists of multiple measured values ​​(operation / stop, etc.)

[0057] The data on the number of revolutions of the pump 12 during a predetermined period is collected at a plurality of measurement points during the predetermined period. The rotation speed is composed of a plurality of measured values ​​(rotation speeds) measured by the pump operation status sensor 33. The data of the number of turns is, for example, linked to the measurement time indicating the time of measurement and arranged in chronological order by the measurement time. The pump operation status sensor 33 continuously detects the When a measurement value is to be output, the learning data acquisition unit 400 outputs the measurement value at a predetermined sampling frequency. This allows the data on the number of rotations for a given period to be recorded. It is composed of multiple measurement values ​​(number of rotations) at multiple sampling points within the interval.

[0058] The data of the motor current value supplied to the motor 13 which is the driving source of the pump 12 during a predetermined period The data is measured by the pump operation status sensor 33 at multiple measurement points within a predetermined period. The data of the current value is composed of a number of measured values ​​(current values) that indicate the time of measurement, for example. The data is linked to a fixed time and arranged in chronological order according to the measurement time. In the case where the pump operation status sensor 33 outputs continuous measured values, the learning data The data acquisition unit 400 converts the measurement values ​​into data at a predetermined sampling period. The data of the current value during a predetermined period is a plurality of data at a plurality of sampling points during the predetermined period. It is composed of the measured value (current value).

[0059] Used at the fluid supply destination 44 to which the fluid is supplied from the pressure tank 17 during a specified period The fluid flow data is a measurement of the total usage flow rate, calculated by adding up the flow rates of multiple fluid supply destinations 44. The fluid is supplied from the pressure tank 17 to the fluid supply destination 44. For example, the flow rate data is linked to the measurement time, which indicates the time of measurement, and is sorted in chronological order by the measurement time. The flow rate sensor 34 measures continuous measured values. In the case where the learning data acquisition unit 400 outputs The measured values ​​are converted into data. This allows the amount of fluid supplied from the pressure tank 17 during a given period to be calculated. The data on the flow rate of the fluid used at the fluid supply destination 44 is collected from multiple samples taken over a given period. The flow rate of the fluid used at the fluid supply destination 44 is calculated by multiplying the measured values ​​by the flow rate of the fluid used at the fluid supply destination 44. do.

[0060] When using "supervised learning" as machine learning, the learning data corresponds to the input data. The output data is measured by the pressure sensor 35 in the pressure tank of the test device 7. The output data further includes data on the charged pressure of the pressure tank 17 during the specified period. In supervised learning, for example, these are called teacher data or correct labels.

[0061] The data of the charged pressure of the pressure tank 17 is a data of a time point included in a predetermined period or after a predetermined period. Point data of the charged pressure at a time point, or a time point during or after a predetermined period of time 4 is time series data of the charged pressure during the period.

[0062] Therefore, the learning data according to this embodiment is, as shown in FIG. Primary pressure data on the suction side of the pump 12, secondary pressure data on the discharge side of the pump 12 for a predetermined period of time, Data on the operation status of the pump 12 during a specified period and data on the fluid used at the fluid supply destination 44 Input data including data on the flow rate of the fluid and the charged pressure of the pressure tank 17 for a specified period of time The output data includes the data.

[0063] Here, the input data included in the learning data (the suction side 1 of the pump 12 in a predetermined period) Next data, secondary pressure data on the discharge side of the pump 12 during a predetermined period, pump 12 operating status data and data on the flow rate of the fluid used at the fluid supply destination 44) , the correlation between the data on the charged pressure of the pressure tank 28 will be explained.

[0064] The secondary pressure on the discharge side of the pump 12 during a given period is calculated by the suction pressure of the pump 12 during a given period. The primary pressure in the pressure tank 17, the operating status of the pump 12 during a predetermined period, and the charged pressure of the pressure tank 17 However, measuring the pressure of the pressure tank 17 requires expensive sensors. In the past, it was highly dependent on the experience of the operator. The input data includes the primary pressure on the suction side of the pump 12 during a predetermined period, The operating condition of the pump 12 and the secondary pressure on the discharge side of the pump 12 for a predetermined period are measured. By setting the pressure tank 17, it is possible to estimate the pressure of the pressure tank 17. By measuring the flow rate of the fluid used at the fluid supply destination 44 in the It becomes possible to estimate the pressure contained in the tank 17.

[0065] When acquiring the learning data, the learning data acquiring unit 400 acquires the learning data from the fluid supplying device 1 or is the suction of the pump 12 during a predetermined period measured by the measuring device 3 provided in the test device 7. data on the primary pressure of the pump 12 during a predetermined period of time; data on the operation status of the pump 12 during a predetermined period of time; The secondary pressure data of the discharge side of the pump 12 and the data of the fluid used at the fluid supply destination 44 during a specified period The data of the flow rate of the fluid is obtained as input data from the measuring device 3. The pressure measurement device 3 estimates the charged pressure of the pressure tank 17 during a specified period when the input data was measured. When the estimated value is input to the worker terminal 8, the learning data acquisition unit 400 The estimated value inputted at the terminal 8 is taken as output data (teaching data) from the worker terminal 8. The learning data acquisition unit 400 then compares the input data with the output data. By associating these, one learning data set is constructed and stored in the learning data storage unit 401 .

[0066] FIG. 5 shows a neural network model used in the machine learning device 4 according to this embodiment. FIG.

[0067] The learning model 6 is configured as a neural network model shown in FIG. The neural network model consists of l neurons (x1 to xl) in the input layer, m neurons (y11~y1m) in the first hidden layer, n neurons (y2 It consists of o neurons (z1 to zo) in the output layer.

[0068] Each neuron in the input layer is associated with one input data set contained in the training data. Each neuron in the output layer is assigned a corresponding output data set included in the training data. In addition, a certain pre-processing may be performed on the input data before it is input to the input layer. Alternatively, a predetermined post-processing may be performed on the output data after it is output from the output layer.

[0069] The first and second hidden layers are also called hidden layers, and are used in the neural network. In addition to the first and second hidden layers, the neural network may further include a plurality of hidden layers. Only the first hidden layer may be a hidden layer. Between the intermediate layer and the second hidden layer, and between the second hidden layer and the output layer, there are connections between the neurons of each layer. Each synapse has a weight wi (i is a natural number). Can be attached.

[0070] The neural network model uses the training data to The input data is input to the input layer, and the inference result is output from the output layer. By comparing the input data with the output data (teaching data) contained in the learning data, The system learns the correlation between the input data and the output data.

[0071] Specifically, each neuron in the input layer is assigned a unique number of input data points included in the training data. The value of each neuron in the output layer is the input The value of the input neuron and the synapse connecting the output neuron and the input neuron The process of calculating the sum of the sequence of multiplication values ​​of the weights wi associated with the input layers is performed on the layers other than the input layer. It is calculated by performing this on all neurons.

[0072] Then, the values ​​(z1 to zo) output to each neuron in the output layer as inference results and the learning The values ​​of the teacher data (t1 to to) corresponding to each of the output data included in the use data are The error is calculated by comparing each of them, and each synapse is assigned a value that minimizes the error. Then, a process of adjusting the weights wi (backpropagation) is performed.

[0073] The above series of steps is repeated a predetermined number of times, or the above error is reduced to a value smaller than the allowable value. When a certain condition for ending the learning is met, such as the machine learning being stopped, the trained Neural network model (all weights wi associated with each synapse) is generated as:

[0074] (Machine learning methods) FIG. 6 is a flowchart showing an example of a machine learning method performed by the machine learning device 4 according to this embodiment. The machine learning method corresponds to the learning phase in Figure 4.

[0075] First, in step S11, the learning data acquisition unit 400 executes As a preliminary step, a desired number of training data are prepared and the prepared training data are used for learning. The training data is stored in the training data storage unit 401. The number of training data to be prepared here is It is only necessary to set it taking into consideration the inference accuracy required for the learning model 6 obtained in the intended manner.

[0076] There are several methods for preparing training data. For example, When estimating the charged pressure of the pressure tank 17 in a specific fluid supply device 1 or test device 7, Various measurement values ​​are obtained before and after the measurement using a measuring device 3. The worker inputs the estimation results in correspondence with these measured values ​​using the worker terminal 8. By doing so, the input data and output data that constitute the training data are prepared. By repeating such a process, it is possible to prepare multiple sets of training data. As a method, for example, the pressure of the pressure tank of the fluid supply device 1 or the test device 7 is intentionally set to a predetermined value. It is also possible to obtain learning data by setting it to a constant value.

[0077] Next, in step S12, the machine learning unit 402 performs pre-learning The learning model 6 before learning is prepared as shown in FIG. It is composed of a neural network model, in which the weights of each synapse are set to their initial values. Each neuron in the input layer is assigned a predetermined period of time as input data included in the learning data. suction side primary pressure data of the pump 12 in a predetermined period, and discharge side secondary pressure data of the pump 12 in a predetermined period data, data on the operation status of the pump 12 during a predetermined period, and data on the fluid The data of the flow rate of the fluid used at the supply destination 44 is associated with each of the output layers. The neuron receives the filled pressure of the pressure tank 17 as output data included in the learning data. Each of the data is associated with each other.

[0078] Next, in step S13, the machine learning unit 402 stores the learning data in the learning data storage unit 401. For example, one piece of learning data is randomly selected from the plurality of sets of learning data stored.

[0079] Next, in step S14, the machine learning unit 402 performs The force data is input to the input layer of the prepared learning model 6 before learning (or during learning). As a result, output data is output from the output layer of the learning model 6 as an inference result. The data is generated by a learning model6 before (or during) learning. Before learning (or during learning), the output data output as the inference result is the learning data. This indicates information different from the output data (teacher data) contained in the data.

[0080] Next, in step S15, the machine learning unit 402 uses the In step S130, the output data (teacher data) included in the one learning data set is The output data is compared with the inference results from the output layer, and the weights of each synapse are adjusted. By this, the machine learning unit 402 performs machine learning by synthesizing input data and output data. The learning model 6 is made to learn correlations with the data.

[0081] Next, in step S16, the machine learning unit 402 determines whether it is necessary to continue machine learning. For example, the error between the output data and the teacher data, or the learning data storage unit 401 The determination is made based on the remaining number of stored unlearned learning data.

[0082] In step S16, if the machine learning unit 402 determines to continue machine learning (step S16), If the result of step S16 is No, the process returns to step S13 and the step Steps S13 to S15 are performed multiple times using unlearned learning data. In S16, when the machine learning unit 402 determines that the machine learning is to be ended (step S1 6), proceed to step S17.

[0083] Then, in step S17, the machine learning unit 402 The trained learning model 6 generated by adjusting the weights is stored in the trained model storage unit 43. The machine learning method shown in FIG. 6 is terminated. S11 is a learning data storage step, steps S12 to S16 are machine learning steps, and step S1 Step 7 corresponds to the learned model storage process.

[0084] As described above, according to the machine learning device 4 and the machine learning method of this embodiment, data on the suction side primary pressure of the pump 12 for a predetermined period as input data included in the data; Discharge side secondary pressure data of pump 12 for a predetermined period, operation of pump 12 for a predetermined period Status data and data on the flow rate of fluid used at the fluid destination 44 over a given period of time. A learning model that can infer (estimate) the pressure of the pressure tank 17 with high accuracy from the data. 6 can be provided.

[0085] (Pressure tank state estimation device 5) FIG. 7 is a block diagram showing an example of the pressure tank state estimating device 5 according to this embodiment.

[0086] The pressure tank state estimation device 5 includes an input data acquisition unit 500, an inference unit 501, and a trained The pressure tank state estimation device 5 includes a model storage unit 502 and an output processing unit 503. For example, the computer 200 shown in FIG. 2 may be used. In this case, the input data acquisition unit 500 The inference unit 501 and the I / O device I / F unit 226 are configured as the communication I / F unit 222 and the I / O device I / F unit 226. The output processing unit 503 is configured with the processor 212, and the trained model storage unit 502 is configured with The pressure tank state estimation device 5 is configured as a fluid supply device 1, or may be incorporated in a higher-level management device (for example, a facility control device) of the fluid supplying device 1. The equipment may be incorporated into a system (such as a controller or an equipment management system for managing multiple equipment).

[0087] The input data acquisition unit 500 is a measuring device 3 (pump primary pressure sensor) provided in the fluid supply device 1. a pump secondary pressure sensor 32, a pump operation status sensor 33, and a usage flow rate sensor 34) and inputs input data (for a predetermined period) based on the measured values ​​measured by the measuring device 3. In the above, the primary pressure data of the suction side of the pump 12, the secondary pressure data of the discharge side of the pump 12, 2 and data on the flow rate of the fluid used at the fluid supply destination 44) are collected. It is a useful interface unit.

[0088] The inference unit 501 converts the input data acquired by the input data acquisition unit 500 into a learning model 6. The inference process is performed by inputting the information into the machine learning device. A pre-trained learning model 6 in which supervised learning was performed using a machine learning method was used. do.

[0089] The inference unit 501 not only performs inference processing using the learning model 6, but also performs As a process, the input data acquired by the input data acquisition unit 500 is adjusted to a desired format, etc. The pre-processing function inputs the learning model 6 to the learning model 6 as a post-processing of the inference process. By applying a predetermined logical formula or calculation formula to the output data value, the pressure tank 17 It also includes a post-processing function for finally estimating the state. The inference result of the inference unit 501 is It is preferable to store the trained model in the trained model storage unit 502 or another storage device (not shown). The inference results of the learning model 6 can be further improved by online learning and It can be used as learning data for re-learning.

[0090] The trained model storage unit 502 stores trained models used in the inference process of the inference unit 501. The trained model storage unit 502 stores the trained model 6. The number of learning models 6 to be used is not limited to one. For example, the number of input data may be different, or A plurality of learning models 6 using different machine learning techniques may be stored and selectively available. .

[0091] The output processing unit 503 outputs the inference result of the inference unit 501, that is, the state of the pressure tank 17. As a specific output means, various means can be adopted. The processing unit 503 notifies the operator of the state of the pressure tank 17 by display or sound, for example, The state of the power tank 17 is recorded as a history, for example, in the fluid supply device 1 or a higher-level The information is transmitted to a management device (not shown) and stored in a storage unit of the management device of the fluid supply device 1. It may also be used for drive control of 13.

[0092] (Pressure tank status estimation method) FIG. 8 illustrates a pressure tank state estimation method performed by the pressure tank state estimation device 5 according to this embodiment. 11 is a flowchart illustrating an example.

[0093] When a current is supplied to the motor 13, the motor 13 is rotated, and the pump 12 At this time, the series of pressure tank state estimation methods shown in FIG. The estimation may be performed at a predetermined timing by the link state device 5. For example, it may be at a predetermined time interval, or when a predetermined event occurs (when an operator gives an instruction to operate, when a maintenance This may be during operation of the sensor.

[0094] First, in step S21, the input data acquisition unit 500 acquires the measurement data of the measurement device 3 (the pump primary A pressure sensor 31, a pump secondary pressure sensor 32, a pump operation status sensor 33, and a usage flow rate sensor Input data based on measurements taken by a sensor 34 (pump 12 over a given period of time) The primary pressure data on the suction side of the pump 12, the secondary pressure data on the discharge side of the pump 12, and the operating status data of the pump 12 and data on the flow rate of the fluid used at the fluid supply destination 44). The input data is data measured by the measuring device 3 during the same period as the predetermined period or a period longer than the predetermined period. The requirements include measurements that have been established.

[0095] Next, in step S22, the inference unit 501 performs preprocessing on the input data to generate a learning model. The input data is input to the input layer of learning model 6, and the output data is obtained from the output layer of learning model 6. do.

[0096] Next, in step S23, the inference unit 501 estimates the pressure in the pressure tank 17. .

[0097] As described above, the pressure tank state estimating device 5 and the pressure tank state estimating method according to the present embodiment According to this method, the state of the pressure tank 17 can be estimated with high accuracy without relying on the experience of the operator. It is possible.

[0098] (Other embodiments) The present invention is not limited to the above-described embodiment, and the scope of the present invention is not limited to the above-described embodiment. Various modifications and variations are possible within the scope of the present invention. It is contained in the idea.

[0099] In the above embodiment, the machine learning unit 402 performs machine learning using neural network. The machine learning unit 402 may be any of the following: Other machine learning techniques such as decision trees may be used. , tree-based learning such as regression trees, ensemble learning such as bagging and boosting, recurrent neural Neural network type (deep neural network, convolutional neural network, etc.) learning), hierarchical clustering, non-hierarchical clustering, k-nearest neighbors, k-means Clustering methods such as the average mean square method, principal component analysis, factor analysis, logistic regression, and other multivariate analyses; Examples include support vector machines.

[0100] In the present invention, the computer 200 shown in FIG. 2 is provided with the machine learning method according to the above embodiment. The machine learning program 230 is provided to execute each step of the process. In addition, the present invention relates to a computer 200 shown in FIG. A program for executing each step of the pressure tank state estimation method (pressure tank state estimation The information may also be provided in the form of a program 230.

[0101] (Inference device, inference method, and inference program) The present invention relates to a pressure tank state estimating device 5 (pressure tank state estimating method or The pressure tank state estimation program is not the only program that estimates the state of the pressure tank 17. In the form of an inference device (inference method or inference program) used to infer In this case, the inference device (inference method or inference program) may include a memory and , and a processor, among which the processor executes a series of processes. The series of processes is the data of the primary pressure on the suction side of the pump 12 during a predetermined period. The secondary pressure data on the discharge side of the pump 12, the operating status data of the pump 12, and the fluid supply destination 44 Input data acquisition process (input When the input data is acquired in the input data acquisition process, the pressure tank 1 and an inference process (inference step) for inferring state estimation information of the seventh embodiment.

[0102] By providing the inference device (inference method or inference program), the pressure tank state inference This makes it easier to apply the inference device to various devices than implementing the inference device 5. When the logic method or inference program infers the state of the pressure tank 17, A machine learning device 4 and a trained learning model 6 generated by a machine learning method are used to The inference method implemented by the inference unit 501 of the power tank state estimation device 5 may be applied. This is something that can be naturally understood by businesses. [Explanation of symbols]

[0103] 1...fluid supply device, 12...pump, 13...motor, 14...inverter, 17...pressure tank, 41...fluid supply source, 44...fluid supply destination, 2...Pressure tank status estimation system, 3... Measuring device, 31... Pump primary pressure sensor, 32... Pump secondary pressure sensor, 33... Pump operation rotation status sensor, 34...usage flow rate sensor, 4...Machine learning device, 5...Pressure tank state estimation device, 6...Learning model, 400...learning data acquisition unit, 401...learning data storage unit, 402...machine learning unit, 40 3...Model storage unit, 500: input data acquisition unit, 501: inference unit, 502: model storage unit, 503: output processing Department, 200…Computer

Claims

1. A machine learning device that generates a learning model for estimating a pressure of a pressure tank in a fluid supply device including a pump that transfers a fluid and a pressure tank provided on a discharge side of the pump, the machine learning device comprising: a learning data storage unit that stores a plurality of sets of learning data in which the input data includes pump pressure data for the pump during a predetermined period, data on the operating status of the pump during the predetermined period, and data on the usage flow rate of the fluid used at a fluid supply destination to which the fluid is supplied from the pressure tank during the predetermined period, and data on the filled pressure of the pressure tank as output data, and in which the input data and the output data are associated with each other; a machine learning unit that causes the learning model to learn a correlation between the input data and the output data by inputting a plurality of sets of the learning data to the learning model; A trained model storage unit that stores the learning model trained by the machine learning unit. Machine learning device.

2. The data on the usage flow rate of the fluid included in the input data is time series data of a total usage flow rate obtained by summing up the measurement results of the usage flow rate measuring meters provided at the plurality of fluid supply destinations; Time series data of entry and exit records provided at the plurality of fluid supply destinations; Time series data of faucet usage amounts provided at the plurality of fluid supply destinations; Time series data of electricity usage provided at the plurality of fluid supply destinations, or Time series data of gas usage provided at the plurality of fluid supply destinations; That is, The machine learning device according to claim 1 .

3. The pump pressure data included in the input data is Time series data of the suction side pressure of the pump; At least one of time series data of the discharge side pressure of the pump, The machine learning device according to claim 1 or 2.

4. The data on the operation status of the pump included in the input data is Time series data of the operation state of the pump, which indicates whether the pump is operating or stopped; Time series data of the rotation speed of the pump, or Time series data of a motor current value supplied to a motor which is a driving source of the pump. The machine learning device according to claim 1 .

5. The data on the charged pressure of the pressure tank included in the output data is Point data of the charged pressure at a time point included in the predetermined period or at a time point after the predetermined period, or The time series data of the charged pressure in the predetermined period or a period including a point after the predetermined period. The machine learning device according to claim 1 .

6. The predetermined period is: a start time when the pump is started; The machine learning device according to claim 1 .

7. 7. A pressure tank state estimation device for estimating a pressure of a pressure tank in a fluid supply device including a pump for transporting a fluid and a pressure tank provided on a discharge side of the pump, using a learning model generated by the machine learning device according to any one of claims 1 to 6, an input data acquisition unit that acquires input data including pump pressure data for the pump during a predetermined period and data on an operating status of the pump during the predetermined period; an inference unit that inputs the input data acquired by the input data acquisition unit into the learning model and infers the charged pressure of the pressure tank; Pressure tank condition estimator.

8. 1. An inference device used to estimate a pressure of a pressure tank in a fluid supply device including a pump for transferring a fluid and a pressure tank provided on a discharge side of the pump, comprising: The inference device comprises a memory and a processor; The processor, an input data acquisition process for acquiring input data including pump pressure data for the pump during a predetermined period, data on the operating status of the pump during the predetermined period, and data on the usage flow rate of the fluid used at a fluid supply destination to which the fluid is supplied from the pressure tank during the predetermined period; When the input data is acquired in the input data acquisition process, an inference process is executed to infer the charged pressure of the pressure tank. Reasoning device.

9. A machine learning method for generating a learning model for estimating a pressure of a pressure tank in a fluid supply device including a pump for transporting a fluid and a pressure tank provided on a discharge side of the pump, the method comprising: a learning data storage step of storing a plurality of sets of learning data in a learning data storage unit, the learning data including input data including pump pressure data for the pump during a predetermined period, data regarding the operating status of the pump during the predetermined period, and data regarding a usage flow rate of the fluid used at a fluid supply destination to which the fluid is supplied from the pressure tank during the predetermined period, and output data including data regarding the filled pressure of the pressure tank, the input data and the output data corresponding to each other; By inputting a plurality of sets of the learning data to the learning model, the input data and the output data are a machine learning step of having a machine learning unit learn a correlation between the learning model and force data; A trained model storage step of storing the learning model trained by the machine learning unit in a trained model storage unit. Machine learning methods.

10. 7. A pressure tank state estimation method for estimating a pressure of a pressure tank in a fluid supply device including a pump for transporting a fluid and a pressure tank provided on a discharge side of the pump, using a learning model generated by the machine learning device according to any one of claims 1 to 6, comprising: an input data acquisition step of acquiring input data including time series data of a pump pressure related to the pump during a predetermined period and time series data of an operating status of the pump during the predetermined period; an inference step of inputting the input data acquired by the input data acquisition step into the learning model and inferring the charged pressure of the pressure tank, Pressure tank condition estimation method.

11. 1. An inference method used to estimate a pressure of a pressure tank in a fluid supply device including a pump for transferring a fluid and a pressure tank provided on a discharge side of the pump, comprising: an input data acquisition step of acquiring input data including pump pressure data for the pump during a predetermined period, data on the operating status of the pump during the predetermined period, and data on the usage flow rate of the fluid used at a fluid supply destination to which the fluid is supplied from the pressure tank during the predetermined period; When the input data is acquired in the input data acquisition step, an inference step of inferring a charged pressure of the pressure tank is executed. Reasoning method.

12. A method for causing a computer to execute each step of the machine learning method according to claim 9, Machine learning programs.

13. A method for causing a computer to execute each step of the pressure tank state estimation method according to claim 10, Pressure tank condition estimation program.

14. A method for causing a computer to execute each step of the inference method according to claim 11, Inference program.

Citation Information

Patent Citations

  • Inspection device for feed water unit

    JP1996086278A

  • Method for detecting abnormality in pressure tank, and water supply system

    JP2006161337A

  • Diaphragm inspection method for liquid supply system and apparatus therefor

    JP3596909B2