JP7576483B2 - Machine learning device, setting device, inference device, machine learning method, setting method, inference method, machine learning program, setting program, and inference program - Google Patents

Description

The present invention relates to a machine learning device, a setting device, an inference device, a machine learning method, a setting method, an inference method,
This relates to machine learning programs, configuration programs, and inference programs.

A water supply device, which is an example of a fluid supply device, is used to supply water (
Water supply devices are widely used as devices for supplying water (fluids). A water supply device generally includes a pump for transporting water, a control unit for controlling the operation of the pump, and a pressure tank provided in the water supply pipe on the discharge side of the pump.

Recently, direct-connected water supply systems have been used to supply water to high-rise buildings. Such water supply systems include a first water supply unit for the lower floors and a second water supply unit for the higher floors, which are connected in series. The suction port of the first water supply unit is directly connected to the water main, and water is supplied to the lower floors of the building by this first water supply unit. Meanwhile, the second water supply unit is located on the middle floors of the building, and boosts the water supplied from the first water supply system and supplies it to the higher floors.

Although the basic internal configuration of the water supply unit is the same as that of a water supply device that is used alone and not connected, in this specification, the water supply device is defined as being composed of multiple water supply units that are directly connected. Also, the number of stages to which the water supply units are directly connected is not limited to two as described above, and may be three or more. Also, in this specification, the water supply unit is defined as an example of a fluid supply unit.

For example, Patent Document 1 (JP 2020-45747 A) describes a water supply system in which multiple water supply devices supply water to water supply targets within a building, in which the multiple water supply devices are capable of communicating with each other via power line communication that uses the building's power lines as a communication line.

JP 2020-45747 A

In a water supply device that supplies water to a building using multiple water supply units, a first water supply unit and a second water supply unit are configured to communicate with each other to exchange their operating status. This allows each water supply unit to control its own operation according to the operating status of the other water supply unit. In the invention described in Patent Document 1, when a communication abnormality is detected, the pump in the second water supply unit is configured to start at a startup speed slower than the normal startup speed.

However, in the conventional technology described in Patent Document 1, each water supply unit exchanges data related to the pressure on the discharge side of the pump and data related to the rotation speed of the motor that drives the pump via a communication line to control its own water supply unit, but since each water supply unit is basically operated independently, it does not enable efficient operation from a bird's-eye view of the entire water supply system.

In view of the above-mentioned problems, the present invention provides a machine learning device, an inference device, a machine learning method, an inference method, a machine learning program, and the like that are suitable for a fluid supply device that operates each water supply unit by referring to more extensive data rather than just referring to data related to the operating status of each water supply unit.
The present invention aims to provide an inference program.

In order to achieve the above object, a machine learning device according to an aspect of the present invention comprises:
A fluid supply unit that is installed in a building and includes a motor whose rotation is controlled by an inverter and a pump that uses the motor as a drive source to transport a fluid;
A machine learning device that generates a learning model of settings for controlling a fluid supply device including a usage trend acquisition unit that acquires usage trends of a fluid in a consumer in the building,
a learning data storage unit configured to store a plurality of sets of learning data including at least data related to the fluid usage trend acquired by the usage trend acquisition unit as input data;
a machine learning unit that inputs a plurality of sets of the learning data to the learning model, thereby causing the learning model to learn a correlation between the input data and a parameter when the inverter controls the motor;
A trained model storage unit that stores the learning model trained by the machine learning unit;
Equipped with.

According to the machine learning device of the present invention, control settings for operating the water supply unit are machine-learned based on fluid usage trends acquired by a usage trend acquisition unit (sensor group 45) that acquires fluid usage trends among consumers within the building, thereby enabling efficient operation from a bird's-eye view of the entire water supply system.

Other objects, configurations and effects will become apparent from the detailed description of the invention described below.

1 is a schematic diagram showing an example of a building 1 to which a water supply device 10 according to the present invention is applied. 2 is a schematic diagram illustrating the structure of a water supply unit WSU _n in the water supply device 10. FIG. FIG. 2 is a hardware configuration diagram showing an example of a computer 200. FIG. 2 is an image diagram illustrating measuring instruments, sensors, etc. installed at a consumer's facility. 1 is a table showing an example of data acquired by measuring instruments, sensors, etc. installed in a consumer facility. FIG. 1 is a block diagram showing an example of a machine learning device 14 according to the present invention. FIG. 2 is a data configuration diagram showing an example of data (supervised learning) used in the machine learning device 14 according to the present invention. 13 is a diagram showing an example of status data of the water supply unit _WSU1 . FIG. 11 is a diagram showing an example of the start/stop timing of each water supply unit WSU _n (the operation sequence of each INV _n ). FIG. FIG. 2 is a schematic diagram showing an example of a neural network model used in the machine learning device 14 according to the present invention. 4 is a flowchart showing an example of a machine learning method by the machine learning device 14 according to the present invention. FIG. 2 is a block diagram showing an example of an inference device 19 according to the present invention. 4 is a flowchart showing an example of control of the water supply unit _WSU1 based on inference by the inference device 19 according to the present invention. FIG. 2 is a diagram for explaining control based on the inference device 19 according to the present invention.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
The scope necessary for the explanation to achieve the objective of the present invention is shown diagrammatically, and the scope necessary for the explanation of the relevant parts of the present invention will be mainly explained, and the parts where explanation is omitted will be based on publicly known technology.

The water supply device 10 according to the present invention can be installed and used in buildings and facilities such as apartment buildings, office buildings, public facilities, commercial facilities, factories, etc. Fig. 1 is a schematic diagram showing an example of a building 1 to which the water supply device 10 according to the present invention is applied. In the water supply device 10 according to the present invention, a direct-connected booster water supply system is adopted. Such a direct-connected booster water supply system is used in high-rise buildings (for example, buildings with 16 or more floors).

In the following embodiment, the water supply device 10 according to the present invention is applied to one building 1, but the concept of the present invention can be applied to a plurality of buildings. The concept of the present invention can also be applied to the control of a single water supply unit.

As shown in FIG. 1, the water supply device 10 includes a first water supply unit WSU ₁ connected to a water main 2 and a second water supply unit WSU 2 connected in series to the first water supply unit WSU ₁ .
The building ₁ has a first water supply unit WSU ₂ and a third water supply unit WSU ₃ connected in series to the second water supply unit WSU 2. The first water supply unit WSU ₁ is installed at ground level or underground, the second water supply unit WSU ₂ is installed, for example, on the 16th floor of the building 1, and the third water supply unit WSU ₃ is installed, for example, on the 30th floor of the building 1.

In this embodiment, the water supply device 10 will be described as an example in which three stages of water supply units WSU are connected as shown in Figure 1, but the number of stages of water supply units WSU may be two stages or four stages or more. Also, in this embodiment, the water supply units may be generalized and described as WSU _n . In this case, the water supply units WSU _n installed on the lower floors of the building 1 have a smaller number n. Also, the water supply pipe connected to the discharge side of the water supply unit WSU _n is given the reference symbol 7 _n , and the pipes branching off from the water supply pipe 7 _n and conducting water to each consumer in the building 1 are referred to as branch pipes 9 _n.
It is called.

In addition, in this embodiment, the water supply unit WSU _n is described as supplying water as a fluid, but the water supply unit WSU _n may also supply liquids other than water, such as fuel and chemicals, as well as fluids including gases such as gas, hydrogen, and oxygen.

The suction port of the first water supply unit WSU ₁ is connected to the water main 2 via a water distribution pipe 7 _0. In this water distribution pipe 7 ₀ , a pressure reducing backflow preventer 8 is connected in front of the suction port of the water supply unit WSU _1. The installation of this pressure reducing backflow preventer 8 is mandatory to reliably prevent backflow of water into the water main 2. Note that the pressure reducing backflow preventer is a backflow preventer configured such that two check valves are arranged to sandwich an intermediate chamber in which a relief valve is arranged.

The discharge outlet of the first water supply unit _WSU1 and the suction port of the second water supply unit _WSU2 are connected by a water distribution pipe ₇₁ , and this water distribution pipe ₇₁ is connected via branch pipes ₉₁ to each customer on the 15th floor (lower floors) of building 1.

The discharge outlet of the second water supply unit _WSU2 and the suction port of the third water supply unit _WSU3 are connected by a water distribution pipe ₇₂ , and this water distribution pipe ₇₂ is connected via branch pipes ₉₂ to each of the customers on the 16th to 29th floors (middle floors) of building 1.

A water distribution pipe ₇₃ is connected to the discharge outlet of the third water supply unit _WSU3 , and this water distribution pipe ₇₃ is connected via a branch pipe ₉₃ to each customer on floors higher than the 30th floor of building 1, including the 30th floor.

The first water supply unit WSU ₁ boosts the pressure of water from the water main 2 to supply water to the consumers on the lower floors of the building 1 and also supplies water to the water supply unit WSU ₂ .
The first water supply unit WSU ₂ pressurizes the water transferred from the first water supply unit WSU ₁ to supply water to each consumer on the middle floors of the building 1 and also supplies water to the third water supply unit WSU ₃ , and the third water supply unit WSU ₃ pressurizes the water transferred from the second water supply unit WSU ₂ to supply water to each consumer on the upper floors of the building 1.

FIG. 2 is a schematic diagram illustrating the structure of the water supply unit WSU _n in the water supply apparatus 10. As shown in FIG.
FIG. 2A shows the structure of the water supply unit WSU _n used in the learning phase.
FIG. 2B shows a structure suitable for the water supply unit WSU _n used in the inference phase. Note that the water supply unit WSU _n having the structure described in FIG. 2A can be used in the present invention, and it is not necessary to use the water supply unit WSU _n having the structure shown in FIG. 2B in the present invention. With regard to the piping system, the structure in FIG. 2B is a structure in which some parts have been omitted from the structure in FIG. 2A. In the following, the piping system will be described based on the structure in FIG. 2A.

As shown in Fig. 2(A), the water supply unit _WSUn includes a pump _Pn connected to the water main 2 and the water supply unit WSUn _-1 via a water distribution pipe _7n-1 , a motor _Mn as a drive source for driving the pump _Pn , an inverter _INVn for increasing or decreasing the rotation speed of the motor _Mn , a suction side pressure sensor _PSn1 disposed on the suction side of the pump _Pn , a check valve _CVn disposed on the discharge side of the pump _Pn , a flow switch _SWn disposed on the discharge side of the check _{valve CVn} , a discharge side pressure sensor _PSn2 , and a pressure tank _PTn . These components are housed, for example, in a cabinet. In this embodiment, a set of the pump _Pn , the motor M
_n , inverter INV _n , check valve CV _n , and flow switch SW _n are arranged in parallel, but two or more sets of pumps, motors, inverters, check valves, and flow switches may be provided in parallel.

The check valve CV _n is provided in a discharge pipe connected to the discharge port of the pump P _n .
n is a valve for preventing backflow of water when the pump _{P n} stops. The flow switch SW _n is a flow detector that detects when the flow rate of water flowing through the discharge pipe drops to a predetermined value. The suction side pressure sensor PS _n1 is a water pressure gauge for measuring the suction side pressure of the water supply unit WSU _n . The discharge side pressure sensor PS _n2 is a water pressure gauge for measuring the discharge side pressure of the water supply unit WSU _n . The pressure tank PT _n is a pressure retainer for retaining the discharge side pressure while the pump P _n is stopped. The pressure tank PT _n has a diaphragm (not shown) that divides the inside of the container into an air chamber and a water chamber. The air chamber has air sealed therein beforehand, and the water chamber is connected to the discharge pipe.

The water supply unit WSU _n further includes a control unit CU _n for controlling the water supply operation, and the inverter INV _n , flow switch SW _n , suction side pressure sensor PS _n1 , and discharge side pressure sensor PS _n2 are connected to the control unit CU _n via signal lines. In the drawing, the dashed line indicates a data communication line. In the structure of FIG. 2(A), when the flow switch SW _n detects that the water flow rate has dropped to a predetermined value, the control unit CU _n issues a command to the inverter INV _n to temporarily increase the operating speed of the pump P _n , and stops the operation of the pump P _n after increasing the discharge side pressure to a predetermined stop pressure. On the other hand, when the discharge side pressure drops to a predetermined start pressure, the control unit CU _n issues a command to the inverter INV _n to start the operation of the pump P _n . This start pressure is a threshold value that serves as a trigger to start the pump P _n .

In the water supply unit _WSUn in Fig. 2(A), the pump _Pn is driven at a variable speed by the inverter _INVn based on the output signals of the flow switch _{SWn and} the discharge pressure sensor _PSn2 . The operating status of the pump _Pn can be acquired based on control _signals (control command value for turning the pump _Pn ON or OFF, control command value for the rotation speed of the motor _Mn , etc., voltage applied to the motor _Mn and its frequency, etc.) sent from the control unit CUn to the inverter INVn. The operating status of the pump _Pn may also be acquired from the output frequency of the inverter _INVn .

In general, constant discharge pressure control is performed to control the operating speed of the pump _Pn so that the pressure signal measured by the discharge pressure sensor _PSn2 coincides with a set target pressure, thereby controlling the discharge side pressure of the pump _Pn to be constant, and constant estimated terminal pressure control is performed to control the supply water pressure at the terminal water faucet to be constant by appropriately changing the target value of the discharge side pressure of the pump _Pn . According to these controls, the pump _Pn is driven at a rotation speed appropriate to the water demand at each time, thereby achieving energy saving. The above constant discharge pressure control and constant estimated terminal pressure control are appropriately performed when constructing the learning model 16 by machine learning.

The control unit CU _n is connected to the network module NWM _n so as to be able to perform data communication, and the network module NWM _n is further able to perform data communication with the network 240. This makes it possible, for example, for the control unit CU _n to transmit data relating to the detection results of the flow switch SW _{n , the suction side pressure sensor PS n1 , and the discharge side pressure sensor PS n2 to the computer 200, and also makes it possible for the computer 200 to transmit data relating to the detection results of the flow switch SW n} , the suction side pressure sensor PS _n1 , and the discharge side pressure sensor PS _n2 to the control unit CU _n.
It is possible to transmit parameters (inverter ON/OFF, rotation speed command value, frequency, voltage, etc.) when the inverter INV _n controls the motor M _n to the inverter INV n.
The computer 200 may be a general-purpose computer, or may be a dedicated computer specialized for the water supply device 10 according to the present invention.

As shown in FIG. 1, a water supply unit WSU constituting a water supply device 10 according to the present invention
The water supply unit _{WSU 1} , the water supply unit WSU ₂ , and the water supply unit WSU ₃ are all connected to a computer 200 via a network 240 so as to be capable of data communication.
00. In this embodiment, an example is shown in which each water supply unit WSU _n is connected to the computer 200 via the network 240, but each water supply unit WSU _n may be connected to the computer 200 without going through the network 240.

Each consumer facility is provided with a sensor group 45 capable of detecting data related to fluids including water and detecting other trends at the consumer facility, and data collected by the sensor group 45 is transmitted to the computer 200 via the network 240. The computer 200 receives the data collected by the sensor group 45 and:
Based on this, settings for controlling each water supply unit WSU _n in the water supply apparatus 10 are calculated.

An example of a sensor included in the sensor group 45 will be described. A flow meter 47 is provided in the branch pipe _9n that supplies water to each customer, and the flow meter 47 measures the amount of water used at each customer.
is provided. When there are multiple supply destinations, the flow meter 47 can also measure a total usage flow rate obtained by adding up the flow rates of the multiple supply destinations. Furthermore, a water meter, for example, can be used as the flow meter 47. Data measured by the flow meter 47 is transmitted to the computer 200 via the network 240. Note that the sensor group 45 is not limited to such a flow meter 47, but is intended to obtain data related to a wider range of consumers, and other examples of the sensor group 45 will be described later.

FIG. 3 is a hardware configuration diagram showing an example of the computer 200.
Either a general-purpose or dedicated water supply unit can be used.
The control unit CU _n in the SU _n can also use the computer 200 .

As shown in FIG. 2, the computer 200 includes, as its main components, a bus 210,
A processor 212, a memory 214, an input device 216, a display device 218, a storage device 220, a communication I/F (interface) unit 222, an external device I/F unit 224, an I
The computer 200 includes an I/O (input/output) device I/F unit 226, and a media input/output unit 228. Note that the above components may be omitted as appropriate depending on the application in which the computer 200 is used.

The processor 212 is one or more arithmetic processing devices (CPU, MPU, GPU, DSP, etc.).
The memory 21 is made up of a plurality of memory modules, each of which functions as a control unit for controlling the entire computer 200.
The internal memory 4 stores various data and programs 230 and is composed of, for example, a volatile memory (DRAM, SRAM, etc.) that functions as a main memory, and a non-volatile memory (ROM, flash memory, etc.).

The input device 216 is, for example, a keyboard, a mouse, a numeric keypad, an electronic pen, etc. The display device 218 is, for example, a liquid crystal display, an organic EL display, electronic paper, a projector, etc. The input device 216 and the display device 218 may be integrated together, such as a touch panel display. The storage device 220 is, for example, a HDD, an SSD, etc., and stores various data required for the execution of the operating system and the program 230.

The communication I/F unit 222 is connected to a network 240 such as the Internet or an intranet by wire or wirelessly, and transmits and receives data to and from other computers in accordance with a predetermined communication standard. The external device I/F unit 224 is connected to an external device 250 such as a printer or scanner by wire or wirelessly, and transmits and receives data to and from the external device 250 in accordance with a predetermined communication standard. Note that an international standard communication means can typically be used as the wireless communication means of the external device I/F unit 224. An example of an international standard communication means is IEEE 802.
15.4, IEEE802.15.1, IEEE802.15.11a, 11b, 11g
, 11n, 11ac, 11ad, ISO/IEC14513-3-10, IEEE802
15.4g and the like. In addition, the external device I/F unit 224 supports wireless communication using Bluetooth (registered trademark), Bluetooth Low Energy, Wi-Fi, etc.
i, ZigBee (registered trademark), Sub-GHz, EnOcean (registered trademark), etc. can also be used. The II/O device I/F unit 226 is connected to an I/O device 260 such as various sensors and actuators, and transmits and receives various signals and data, such as detection signals from sensors and control signals to actuators, between the I/O device 260. The media input/output unit 228 is composed of a drive device such as a DVD drive, a CD drive, etc., and reads and writes data from and to media 270 such as DVDs and CDs.

In the computer 200 having the above configuration, the processor 212 executes the program 23.
2, the processor 212 loads the program 230 into a work memory area of the memory 214, executes the program 230, and controls each unit of the computer 200 via the bus 210. The program 230 may be stored in the storage device 220 instead of the memory 214. The program 230 may be recorded on a non-transitory recording medium such as a CD or DVD in an installable file format or an executable file format, and provided to the computer 200 via the media input/output unit 228. The program 230 may be provided to the computer 200 by being downloaded via the network 240 through the communication I/F unit 222. The computer 200 may also control various functions that are realized by the processor 212 executing the program 230, for example, as an FP.
It may be realized by hardware such as GA or ASIC.

The computer 200 is, for example, a desktop computer or a portable computer, and is an electronic device of any type. The computer 200 may be a client computer, a server computer, or a cloud computer.

Next, the sensor group 45 that acquires various data at the consumer will be described. FIG. 4 is an image diagram illustrating measuring instruments, sensors, etc. installed at the consumer, and shows a perspective view of a room in an apartment or the like that is a consumer. FIG. 5 is a table showing an example of data acquired by measuring instruments, sensors, etc. installed at the consumer. The main ones in the sensor group 45 are those that directly measure the flow rate of a fluid (water in this embodiment), but the present invention is not limited to this and also includes measuring instruments, sensors, etc. that can indirectly grasp the usage trend of the fluid at each consumer. These sensor groups 45 are also referred to as a fluid usage trend acquisition unit.

The main components of the sensor group 45 are a flow meter 47 and a pressure meter 49 connected to the branch pipe 9 _n of each customer. The measurement data obtained by these sensors is input to the computer 20.
The pressure data of the fluid (water) acquired by the pressure gauge 49 at each consumer provided at the end of the branch pipe 9 _n is transmitted to the customer's local network 100.
This data is directly related to the feel of the liquid (water) when in use, and is one of the most important pieces of data that can be obtained by the sensor group 45.

The sensor group 45 may also include sensors that acquire data related to lifelines other than water.
A power meter 54 and a gas flow meter 56 are attached to each of the fluids, and the power data and gas flow data acquired by these are also transmitted to computer 200. It has been found that data relating to the usage of fluids other than the fluid of interest (water), i.e., data relating to the amount of electricity usage and data relating to the amount of gas usage acquired by power meter 54 and gas flow meter 56, are closely correlated with the amount of fluid (water) used by the consumer, and in the present invention, these data are used as data for understanding the usage trends of fluids at each consumer.

Furthermore, the sensor group 45 may include a meter that detects the trend of water usage at each location in the demand. Specifically, the sensor group 45 may include a washbasin water flow meter 61 that measures the amount of water usage at the washbasin, a washing machine water flow meter 62 that measures the amount of water usage at the washing machine,
The flow rate data measured by these devices can include a bathroom water flowmeter 63 that measures the amount of water used in the bathroom, a kitchen water flowmeter 64 that measures the amount of water used in the kitchen, and a toilet water flowmeter 65 that measures the amount of water used in the toilet.
It is transmitted to the computer 200.

Furthermore, the sensor group 45 may include sensors that detect the opening and closing of doors and the like at various locations within the consumer. Specifically, the sensor group 45 may include a front door opening and closing sensor 67 that detects the opening and closing of the front door, a bathroom door opening and closing sensor (not shown) that detects the opening and closing of the bathroom door, a shower curtain sensor (not shown) that detects the opening and closing of the shower curtain, and a toilet door opening and closing sensor (not shown) that detects the opening and closing of the toilet door. The detection data acquired by these sensors is also transmitted to the computer 200. For example, if a record of people entering and leaving the room can be grasped based on the opening and closing of the bathroom door, it is possible to estimate the usage flow rate of the fluid from the record of people entering and leaving the room. In the present invention, the management information that can be acquired by each of the above sensors, such as the front door opening and closing sensor 67, is referred to as data regarding people entering and leaving the consumer space.

Furthermore, the sensor group 45 may include sensors that detect the lighting status at each location within the demand. Specifically, the sensor group 45 may include a bathroom lighting switch state detection sensor (not shown) that detects the lighting status of the bathroom lighting, and a toilet lighting switch state detection sensor (not shown) that detects the lighting status of the toilet lighting. The detection data acquired by these sensors is also transmitted to the computer 200. For example, if it is known that the toilet lighting is turned on, it is possible to estimate the flow rate used in the toilet. The data acquired by the sensors that detect the lighting status of the lighting as described above can be understood as data related to human activities in the demand.

Furthermore, the sensor group 45 may include sensors that are not shown in Figures 4 and 5 and that can indirectly estimate the fluid usage trends of the consumer. For example, a human presence sensor may be installed in each room where the fluid may be used, i.e., the toilet, bathroom, kitchen, etc., and the usage flow rate may be estimated from the detection results of the human presence sensor. Furthermore, the usage flow rate may also be estimated from the amount of electricity usage measured by an electric meter or the amount of gas usage measured by a gas meter installed in each room.

In this embodiment, a learning model 16 based on machine learning is used to derive control settings for each water supply unit WSU _n constituting the water supply apparatus 10. The phase in which the control settings are obtained using the learning model 16 can be referred to as an inference phase.
As a machine learning method in the learning phase of constructing the model, "supervised learning" can be adopted. As a specific method of supervised learning, a neural network can be adopted.

When a neural network is used for machine learning, each piece of data collected by the sensor group 45 of the consumer in the building 1 can be input to the nodes (neurons) of the input layer of the neural network. On the other hand, the nodes (neurons) of the output layer can be input with the water supply unit WSU.
The output can be configured to output values related to the status data of the water supply unit WSU _n . Here, the status data of the water supply unit WSU _n can include ON/OFF data of the INV _n , data such as the rotation speed command value, frequency, and voltage, and detection data of the discharge side pressure sensor PS _n2 of the pump P _n . (The detection data of the suction side pressure sensor PS _n1 of the pump P _n is omitted because it is less important in terms of the operation of the water supply unit WSU _n than the detection data of the discharge side pressure sensor PS _n2 .)

When the learning model 16 trained by machine learning is stored in the computer 200 and the inference phase is executed by the computer 200, the computer 200 operates as the subject of the inference phase of the machine learning. In addition, when the computer 200 is used as the machine learning device 14, the computer 200 operates as the subject of the learning phase of the machine learning, and generates the learning model 16 used for the control settings of the water supply unit WSU _n by machine learning.

(Machine learning device 14)
FIG. 6 is a block diagram showing an example of a machine learning device 14 according to the present invention.

The machine learning device 14 includes a learning data acquisition unit 40, a learning data storage unit 41, a machine learning unit 42, and a trained model storage unit 43. The machine learning device 14 includes, for example, a
In this case, the learning data acquisition unit 40 is configured with a computer 200 shown in FIG.
The machine learning unit 42 is composed of the processor 212, and the learning data memory unit 41 and the learned model memory unit 43 are composed of the storage device 220.

The learning data acquisition unit 40 is an interface unit that is connected to various external devices via a communication network and acquires learning data including at least input data. In this embodiment, the external devices include the water supply unit WSU _n and a sensor group 4 in the consumer's house in the building 1.
5, etc. The input data is data collected by a sensor group 45 of the consumer in the building.

The learning data storage unit 41 is a database that stores a plurality of sets of learning data acquired by the learning data acquisition unit 40. The specific configuration of the database that constitutes the learning data storage unit 41 may be designed as appropriate.

The machine learning unit 42 performs machine learning using the learning data stored in the learning data storage unit 41. In other words, the machine learning unit 42 inputs a plurality of sets of learning data to the learning model 16, and causes the learning model 16 to learn the correlation between the input data included in the learning data and the status data (teacher data) of the water supply unit WSU _n , thereby generating a trained learning model 16. Here, the status data of the water supply unit WSU _n includes the ON status of INV _n,
The data may include ON/OFF data, a rotation speed command value, frequency, voltage data, and detection data of the discharge side pressure sensor PS _n2 of the pump P _n . In this embodiment, a case where a neural network is adopted as a specific method of supervised learning by the machine learning unit 42 will be described.

The trained model storage unit 43 is a database that stores the trained learning model 16 generated by the machine learning unit 42. The trained learning model 16 stored in the trained model storage unit 43 can be transmitted to a real system (e.g., a water supply device 1) via an arbitrary communication network, a recording medium, or the like.
2. The water supply units WSU _n of the system 10 are provided with a computer 200 (see FIG. 2) for controlling the water supply units WSU n of the system 10.
In addition, in FIG. 6, the learning data storage unit 41 and the trained model storage unit 43 are shown as separate storage units, but these may be configured as a single storage unit.

FIG. 7 is a data configuration diagram showing an example of data (supervised learning) used by the machine learning device 14 according to the present invention.

The learning data may include, as input data, a data set based on a sensor group 45 that acquires usage trends of fluid (water) at consumers.
In the data processing device 14 according to the present invention, machine learning is performed to learn correlations between data on fluid pressure acquired at the consumer and output data, correlations between data on people entering and leaving the consumer space and output data, correlations between data on electricity usage at the consumer and output data, correlations between data on gas usage at the consumer and output data, and correlations between data on people's activities at the consumer and output data. In the machine learning device 14 according to the present invention, at least one of these correlations may be used. The input data may further include other data in addition to those described above. For example, the input data may include temperature data and humidity data in the consumer's room, and weather data for the location where the building 1 is located.

The data set acquired by the sensor group 45 as input data may be a current data set, or a time series data set (e.g., every second) going back a certain amount of time (e.g., 10 minutes ago) to the present. The data set as input data may also be a processed data set, such as an average of a time series data set going back a certain amount of time to the present.

When "supervised learning" is adopted as the machine learning, the learning data includes the status data of the water supply unit WSU _n as output data associated with the input data. Here,
The status data of the water supply unit WSU _n is assumed to be a data set including ON/OFF data of the INV _n , data such as a rotation speed command value, frequency, and voltage, and detection data of the discharge side pressure sensor PS _n2 of the pump P _n . In supervised learning, the output data is called, for example, teacher data or correct answer label.

In the learning data, the acquisition timing of the data set that is input data and the timing that the data set that is output data are not required to be one-to-one, and a certain time span can be provided. For example, by using learning data such as learning data based on input data at time T and output data at time T, learning data based on input data at time T and output data at time T+Δt, learning data based on input data at time T and output data at time T+2Δt, etc., the machine learning device 14 can learn the correlation between the present time and the near future (e.g., about 10 minutes).

As described above, by performing learning with a certain time gap between the timing of acquiring the data set that will be input data and the timing expected for the data set that will be output data, it is possible to infer, for example, the start/stop (ON/OFF) timing of the inverter from the present to the near future using the inference result Ra in the inference phase of Figure 7.

Here, an example of time series data that is machine-learned by the machine learning device 14 in the present invention will be described. FIG. 8 is a diagram showing an example of status data of the water supply unit _WSU1 . As an example, the status data of the first water supply unit _WSU1 is shown. In FIG. 8, the horizontal axis indicates the time axis. Also, FIG. 8(A) shows a rotation speed command value in percentage as an example of a control setting parameter of the inverter _INV1 . (Note that other parameters such as frequency and voltage may also be used as appropriate as the control setting parameters of the inverter.) Also, FIG. 8(B) shows a rotation speed command value in percentage as an example of a control setting parameter of the inverter INV1.
FIG. 8 shows the ON/OFF timing of the power supply ₁ , i.e., the start and stop timing.
1C) shows detection data from the discharge side pressure sensor _PS12 of the pump _P1 . The machine learning device 14 learns these data as output data to generate a learning model 16. The data surrounded by the dashed dotted line shows, as an example, a data set of output data at time T+2Δt. The data set of output data is not limited to one point in time at time T+2Δt, and for example, a series of time series data sets from time T+4Δt to time T+7Δt surrounded by the dashed dotted line can also be treated as a single piece of output data for learning.

FIG. 9 is a diagram showing an example of the start/stop timing of each water supply unit WSU _n (the operation sequence of each INV _n ). In FIG. 9, the horizontal axis indicates the time axis.
) indicates the ON/OFF timing of the inverters _INV1 to _INV3 .
9 as a whole shows the operation sequence of each inverter. A learning model 16 is generated based on such output data. The data surrounded by the dashed dotted line shows, as an example, a data set of output data at time T+2Δt. The data set of output data is not limited to one point in time at time T+2Δt, and for example, a series of time series data sets from time T+4Δt to time T+7Δt surrounded by the dashed dotted line can also be treated as a single piece of output data for learning.

FIG. 10 is a schematic diagram showing an example of a neural network model used in the machine learning device 14 according to the present invention.

The learning model 16 is configured as a neural network model shown in Fig. 10. The neural network model includes l neurons (x1 to xl) in an input layer, m neurons (y11 to y1m) in a first hidden layer, and n neurons (
y21 to y2n) and o neurons (z1 to zo) in the output layer.

Each neuron in the input layer is associated with a respective input data included in the learning data. Each neuron in the output layer is associated with a respective output data included in the learning data. Note that a predetermined pre-processing may be performed on the input data before it is input to the input layer, and a predetermined post-processing may be performed on the output data after it is output from the output layer.

The first and second hidden layers are also called hidden layers. 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. Synapses that connect neurons in each layer are laid between the input layer and the first hidden layer, between the first hidden layer and the second hidden layer, and between the second hidden layer and the output layer, and a weight wi (i is a natural number) is assigned to each synapse.

A neural network model uses training data to learn the correlation between the input data and output data by inputting input data contained in the training data into an input layer, and comparing the output data output from the output layer as the inference result with the output data (teacher data) contained in the training data.

Specifically, input data included in the learning data is input to each neuron in the input layer, and the value of each neuron in the output layer is calculated by performing a process for all neurons other than the input layer, in which the value of the neuron on the input side connected to the neuron in question is calculated as the sum of a sequence of multiplication values of the value of the neuron on the input side connected to the neuron in question and the weight wi associated with the synapse connecting the neuron on the output side and the neuron on the input side.

Then, the values (z1 to zo) output to each neuron in the output layer as the inference results are compared with the values (t1 to to) of the teacher data corresponding to each of the output data included in the learning data to determine the error, and a process (backpropagation) is performed to adjust the weights wi associated with each synapse so that the error is reduced.

When a predetermined learning end condition is met, such as the above series of steps being repeated a predetermined number of times or the above error being smaller than a tolerance, the machine learning is terminated and the trained neural network model (all weights w i associated with each synapse)
is generated as:

(Machine learning methods)
11 is a flowchart showing an example of a machine learning method according to the present invention, which is performed by the machine learning device 14. The machine learning method corresponds to the learning phase of FIG.

First, in step S100, the learning data acquisition unit 40 prepares a desired number of pieces of learning data as advance preparations for starting machine learning, and stores the prepared learning data in the learning data storage unit 41. The number of pieces of learning data to be prepared here may be set in consideration of the inference accuracy required for the learning model 16 to be finally obtained.

As a method of preparing learning data, while acquiring data from the sensor group 45, the above-mentioned constant discharge pressure control and constant estimated end pressure control are executed in each water supply unit WSU _n , and status data in each water supply unit WSU _n (INV _n ON/OFF data, rotation speed command value, frequency, voltage data, etc., detection data from the discharge side pressure sensor PS _n2 of the pump P _n ) are acquired.
Make sure to obtain the following.

Next, in step S110, the machine learning unit 42 prepares a pre-learning model 16 to start machine learning. The pre-learning model 16 prepared here is configured with the neural network model exemplified in FIG. 10, and the weights of each synapse are set to an initial value. Each neuron in the input layer is associated with data on the amount of power generated during reverse flow as input data contained in the learning data. Each neuron in the output layer is associated with diagnostic information as output data contained in the learning data.

Next, in step S<b>120 , the machine learning unit 42 acquires, for example, one piece of learning data randomly from the multiple sets of learning data stored in the learning data storage unit 41 .

Next, in step S130, the machine learning unit 42 inputs the input data included in one learning data to the input layer of the prepared learning model 16 before learning (or during learning). As a result, output data is output as an inference result from the output layer of the learning model 16, but the output data is generated by the learning model 16 before learning (or during learning). Therefore, in the state before learning (or during learning), the output data output as an inference result indicates information different from the output data (teacher data) included in the learning data.

Next, in step S140, the machine learning unit 42 performs machine learning by comparing the output data (teacher data) included in one of the learning data acquired in step S120 with the output data output from the output layer as an inference result in step S130 and adjusting the weight of each synapse. In this way, the machine learning unit 42 causes the learning model 16 to learn the correlation between the input data (data collected by the sensor group 45) and the output data (status data at the water supply unit WSU _n) .

Next, in step S150, the machine learning unit 42 determines whether or not it is necessary to continue machine learning based on, for example, the error between the output data and the teacher data and the remaining number of unlearned learning data stored in the learning data memory unit 41.

In step S150, if the machine learning unit 42 determines to continue the machine learning (No in step S150), the process returns to step S120, and the process of steps S120 to S140 is performed multiple times using unlearned learning data for the learning model 16 under learning. On the other hand, in step S150, if the machine learning unit 42 determines to end the machine learning (Yes in step S150), the process proceeds to step S160.

Then, in step S160, the machine learning unit 42 stores the trained learning model 16 generated by adjusting the weights associated with each synapse in the trained model storage unit 4.
7. In the machine learning method, step S100 corresponds to a learning data storage step, steps S110 to S150 correspond to a machine learning step, and step S160 corresponds to a trained model storage step.

As described above, according to the machine learning device 14 and the machine learning method according to the present embodiment, the building 1
The control settings for operating the water supply unit are machine-learned based on the fluid usage trends acquired by the usage trend acquisition unit (sensor group 45), which acquires the fluid usage trends of consumers within the facility, thereby enabling efficient operation from a bird's-eye view of the entire water supply device 10.

(Machine Learning Program)
The present invention can also be provided in the form of a program (machine learning program) 230 for causing a computer 200 shown in FIG. 3 to execute each step of the machine learning method according to the above embodiment.

(Inference device 19)
FIG. 12 is a block diagram showing an example of an inference device 19 according to the present invention.

The inference device 19 includes an input data acquisition unit 20, an inference unit 21, and a trained model storage unit 22.
and an output processing unit 23. The inference device 19 is, for example, a computer 20 shown in FIG.
In this case, the input data acquisition unit 20 is configured with a communication I/F unit 222 or an I/O device I/F unit 226, and the inference unit 21 and the output processing unit 23 are configured with a processor 212.
and the trained model storage unit 22 is configured with a storage device 220. In this embodiment, the inference device 19 may be configured with either a general-purpose or dedicated computer (see FIG. 3).

The input data acquisition unit 20 is an interface unit that is communicatively connected to the sensor group 45 and acquires the usage trends of the fluid (water) acquired by the sensor group 45 .

The inference unit 21 inputs the input data acquired by the input data acquisition unit 20 into the learning model 16, and performs an inference process to infer status data of the water supply unit WSU _n .
A machine learning device 14 and a trained learning model 16 in which supervised learning has been performed using a machine learning method are used.

The inference unit 21 not only performs an inference process using the learning model 16, but also includes a pre-processing function of adjusting the input data acquired by the input data acquisition unit 20 into a desired format, etc., and inputting the adjusted data to the learning model 16 as pre-processing of the inference process, and a post-processing function of applying a predetermined logical formula or calculation formula to the value of the output data output from the learning model 16 as post-processing of the inference process. Note that it is preferable to store the inference results of the inference unit 21 in the trained model storage unit 22 or another storage device (not shown), and past inference results can be used as learning data to be used for online learning or re-learning, for example, to further improve the inference accuracy of the learning model 16.

The trained model storage unit 22 is a database that stores trained learning models 16 used in the inference process of the inference unit 21. The number of learning models 16 stored in the trained model storage unit 22 is not limited to one. For example, multiple learning models 16 with different numbers of input data or different machine learning methods may be stored and selectively available.

The output processing unit 23 performs output processing to output the inference result of the inference unit 21, i.e., the status data of the water supply unit WSU _n . The status data of the water supply unit WSU _n output by the output processing unit 23 is transmitted to each water supply unit WSU _n , _and the I
It is used to set ON/OFF of _NVn , to set command values (rotation speed command value, frequency, voltage, etc.) for motor _Mn by _INVn , and to set a target value for discharge side pressure sensor _PSn2 of pump _Pn .

A description will now be given of a mode of use of the inference device 19 configured as above. Fig. 13 is a flow chart showing an example of control of the water supply unit _WSU1 based on inference by the inference device 19 according to the present invention.

In step S210, the sensor group 45 acquires measurement data and detection results such as door opening and closing from the sensor group 45 in the building 1, and then in step S220, transmits the acquired measurement data and detection results to the inference device 19.

In the inference device 19, in step S310, the input data acquisition unit 20 acquires the measurement data and detection results received from the sensor group 45, and the inference unit 21 executes an inference process using the learning model 16. In the following step S320, the output processing unit 23 transmits the obtained inference result (status data) to the water supply units _WSU1 , _WSU2 , and _WSU3 .

In the water supply units WSU ₁ , WSU ₂ , and WSU ₃ , control of the water supply units WSU ₁ , WSU ₂ , and WSU ₃ is carried out based on the inference result received from the inference device 19 in step S 410 .

Next, an example of control of the water supply apparatus 10 by the inference device 19 will be described. Fig. 14 is a diagram for explaining control based on the inference device 19 according to the present invention. Fig. 14 shows the main components of the water supply apparatus 10.

The trends in the use of fluid (water in this embodiment) at the consumer facility collected by the sensor group 45 as described above are transmitted to the inference device 19. Based on this, the inference device 19 infers parameters when the inverter INV _n in each water supply unit WSU _n controls the motor M _n . Here, the parameters in question can include the rotation speed command value, frequency, voltage, etc. for the motor M _n . The above inference results are transmitted to each water supply unit WSU
_n .

In the example of this embodiment shown in FIG. 14, the parameters when the inverter _INV1 controls the motor _M1 , the parameters when the inverter INV2 controls the motor M2, and the parameters when the inverter _INV3 controls the motor _M3 are set as parameters.
The parameters when _V3 controls the motor _M3 are inferred by the inference device 19, and each of the inferred parameters is transmitted to the water supply units _WSU1 , _WSU2 , and _WSU3 and used to control the inverter INV in each water supply unit WSU.

In addition, the inference device 19 infers the start/stop timing of the inverter INV _n in each water supply unit WSU _n based on the measurement data and detection results received from the sensor group 45, and transmits the inference results to each water supply unit WSU _n .

In the example of this embodiment shown in FIG. 14, the start timing and stop timing of each of the inverters INV ₁ , INV ₂ , and INV ₃ are inferred and transmitted to the water supply units WSU ₁ , WSU ₂ , and WSU ₃ , _respectively .
Based on the received start/stop timing, _the WSU ₃ controls the inverters INV ₁ ,
It starts/stops the inverters _INV2 and _INV3 .

In addition, the inference device 19 infers the target value of the discharge side pressure sensor PS _n2 of each pump P _n based on the measurement data and detection results received from the sensor group 45, and transmits the inference results to each water supply unit WSU _n .

In the example of this embodiment shown in FIG. 14, the discharge side pressure sensor PS ₁₂ and the discharge side pressure sensor PS ₂₂
The target values of the discharge side _pressure sensor PS32 and the discharge side pressure sensor _PS32 are inferred by _the inference device 19 and transmitted to the water supply units _WSU1 , _WSU2 , and _WSU3 , respectively.
_U3 is a discharge side pressure sensor _PS12 , a discharge side pressure sensor _PS22 , a discharge side pressure sensor PS
Based on the target value of inverter INV ₁ , inverter _{INV 2} , inverter INV
In this case, the control parameters of _{the three} parameters can be corrected.

As described above, according to the inference device 19 and inference method of this embodiment, inference processing is performed using a machine-learned model based on fluid usage trends acquired by the usage trend acquisition unit (sensor group 45) that acquires fluid usage trends among consumers in the building 1, thereby enabling efficient operation of the water supply device 10 from a bird's-eye view of the entire system.

(Inference method and inference program)
The present invention can be provided not only in the form of the inference device 19 according to the above embodiment, but also in the form of an inference method or an inference program. In this case, the inference device (inference method or inference program) can include a memory and a processor, and the processor executes a series of processes. The series of processes includes an input data acquisition process (input data acquisition step) for acquiring data acquired by the sensor group 45 as input data, and an inference process (inference step) for inferring status data of the water supply unit WSU _n when the input data is acquired by the input data acquisition process.

By providing it in the form of an inference device (inference method or inference program), it can be easily applied to various devices compared to the case of implementing the inference device 19. It can be naturally understood by those skilled in the art that when the inference device (inference method or inference program) infers the status data of the water supply unit WSU _n , an inference method executed by the inference unit 21 of the inference device 19 may be applied using the machine learning device 14 and the trained learning model 16 generated by the machine learning method according to the above embodiment.

Other Embodiments
The present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit and scope of the present invention, all of which are included in the technical concept of the present invention.

In the above embodiment, a case has been described in which a neural network is adopted as a specific method of machine learning by the machine learning unit 42, but the machine learning unit 42 may adopt any other machine learning method. Examples of other machine learning methods include tree types such as decision trees and regression trees, ensemble learning such as bagging and boosting, neural network types (including deep learning) such as recurrent neural networks and convolutional neural networks, clustering types such as hierarchical clustering, non-hierarchical clustering, k-nearest neighbors, and k-means, multivariate analyses such as principal component analysis, factor analysis, and logistic regression, and support vector machines.

1: Building 2: Water main 7 _n : Water distribution pipe 8: Pressure reducing backflow preventer 9 _n : Branch pipe 10: Water supply device (an example of a fluid supply device)
14... Machine learning device 16... Learning model 19... Inference device 20... Input data acquisition unit 21... Inference unit 22... Learned model memory unit 23... Output processing unit 40... Learning data acquisition unit 41... Learning data memory unit 42... Machine learning unit 43... Model memory unit 45... Sensor group 47... Flow meter 49... Pressure gauge 50... Power supply line 52... Gas piping 54... Power meter 56... Gas flow meter 61... Washbasin water flow meter 62... Washing machine water flow meter 63... Bathroom water flow meter 64... Kitchen water flow rate Total 65...Toilet water flow meter 67...Front door opening/closing sensor 200...Computer 210...Bus 212...Processor 214...Memory 216...Input device 218...Display device 220...Storage device 222...Communication I/F (interface) section 224...External device I/F section 226...I/O (input/output) device I/F section 228...Media input/output section 230...Program 240...Network 250...External device 260...I/O device 270...Media WSU _n (e.g., n=1, 2, 3)...Water supply unit (an example of a fluid supply unit)
NWM _n (e.g., n=1, 2, 3) - network module CU _n (e.g., n=1, 2, 3) - control unit INV _n (e.g., n=1, 2, 3) - inverter M _n (e.g., n=1, 2, 3) - motor P _n (e.g., n=1, 2, 3) - pump PS _n1 (e.g., n=1, 2, 3) - suction side pressure sensor PS _n2 (e.g., _n =1, 2, 3) - discharge side pressure sensor _{CV n (} e.g., n=1, 2, 3) - check valve SW _n (e.g., n=1, 2, 3) - flow switch PT _n (e.g., n=1, 2, 3) - pressure tank

Claims (14)

A fluid supply unit that is installed in a building and includes a motor whose rotation is controlled by an inverter and a pump that uses the motor as a drive source to transport a fluid;
A machine learning device that generates a learning model of settings for controlling a fluid supply device including a usage trend acquisition unit that acquires usage trends of a fluid in a consumer in the building,
a learning data storage unit configured to store a plurality of sets of learning data including at least data related to the fluid usage trend acquired by the usage trend acquisition unit as input data;
a machine learning unit that inputs a plurality of sets of the learning data to the learning model, thereby causing the learning model to learn a correlation between the input data and a parameter when the inverter controls the motor;
A trained model storage unit that stores the learning model trained by the machine learning unit;
A machine learning device comprising: The input data is
Data on fluid pressure obtained at the consumer's facility;
Data on people entering and leaving customer spaces;
Data on electricity usage by consumers,
Data on gas consumption at the consumer's premises or data on human activity at the consumer's premises;
The machine learning device according to claim 1 , wherein the machine learning device is at least one of the following: The learning data is
The output data corresponds to the input data and includes a pressure of a fluid on a discharge side of the pump;
The machine learning unit is
2. The method according to claim 1, wherein the correlation between the input data and the output data is learned by the learning model.
Or the machine learning device according to claim 2. The learning data is
The output data includes timings for starting and stopping the inverter, the timings being associated with the input data;
The machine learning unit is
2. The method according to claim 1, wherein the correlation between the input data and the output data is learned by the learning model.
The machine learning device according to claim 3 . The learning data is
The output data includes a rotation speed when the inverter controls the motor, the rotation speed being associated with the input data,
The machine learning unit is
2. The method according to claim 1, wherein the correlation between the input data and the output data is learned by the learning model.
The machine learning device according to claim 4 . A plurality of fluid supply units each having a motor whose rotation is controlled by an inverter and a pump that uses the motor as a drive source to transport a fluid, the fluid supply units being installed in a building ;
A machine learning device that generates a learning model of settings for controlling a fluid supply device including a usage trend acquisition unit that acquires usage trends of a fluid in a consumer in the building ,
a learning data storage unit configured to store a plurality of sets of learning data including at least data related to the fluid usage trend acquired by the usage trend acquisition unit as input data;
a machine learning unit that inputs a plurality of sets of the learning data into the learning model, thereby causing the learning model to learn a correlation between the input data and parameters when the inverters of the plurality of fluid supply units control the motors;
A machine learning device comprising: a trained model memory unit that stores the learning model trained by the machine learning unit. A setting device that sets parameters when the inverter controls the motor by 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 data related to fluid usage trends in consumers in the building;
and an inference unit that inputs the input data acquired by the input data acquisition unit into the learning model and infers a setting value when the inverter controls the motor. A fluid supply unit that is installed in a building and includes a motor whose rotation is controlled by an inverter and a pump that uses the motor as a drive source to transport a fluid;
An inference device that performs inference processing for settings when controlling a fluid supply device including a usage trend acquisition unit that acquires usage trends of a fluid in a consumer in the building,
The inference device includes an inference unit,
The inference unit is
an input data acquisition process for acquiring data relating to the fluid usage trend acquired by the usage trend acquisition unit as input data;
When the input data is acquired in the input data acquisition process, an inference process is performed to infer parameters when the inverter controls the motor. A fluid supply unit that is installed in a building and includes a motor whose rotation is controlled by an inverter and a pump that uses the motor as a drive source to transport a fluid;
A machine learning method for generating a learning model of settings for controlling a fluid supply device including a usage trend acquisition unit that acquires usage trends of a fluid in a consumer in the building,
a learning data storage step of storing a plurality of sets of learning data including at least data relating to the fluid usage trend acquired by the usage trend acquisition unit as input data;
a machine learning process of inputting a plurality of sets of the learning data into the learning model, thereby making the learning model learn a correlation between the input data and a parameter when the inverter controls the motor;
A machine learning method comprising: a trained model storage process for storing the learning model trained by the machine learning process in a trained model storage unit. A setting method for setting parameters when the inverter controls the motor, 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 data related to fluid usage trends of consumers in the building;
The input data acquired in the input data acquisition step is input into the learning model;
and an inference step of inferring a setting value when the inverter controls the motor. A fluid supply unit that is installed in a building and includes a motor whose rotation is controlled by an inverter and a pump that uses the motor as a drive source to transport a fluid;
An inference device that performs inference processing for settings when controlling a fluid supply device including a usage trend acquisition unit that acquires usage trends of a fluid in a consumer in the building,
an input data acquisition step of acquiring data relating to the fluid usage trend acquired by the usage trend acquisition unit as input data;
an inference step of inferring parameters when the inverter controls the motor when the input data is acquired in the input data acquisition step. A method for causing a computer to execute each step of the machine learning method according to claim 9.
Machine learning programs. A machine learning program for causing a computer to execute each step of the setting method according to claim 10. An inference program for causing a computer to execute each step of the inference method according to claim 11.

Images