Air conditioning system

Abstract

An air conditioning system (100) according to the present disclosure includes a compressor (10), an outdoor unit (20), a first indoor unit (30A), a second indoor unit (30B), a refrigeration cycle (11), temperature sensors (35A, 35B), and a control device (50). The control device includes: a data acquisition unit that controls the operating frequency of the compressor when a default condition is satisfied, and that acquires, as status data, data indicating a period during which the default condition is satisfied, the period increased when each of the indoor units, which are determined as targets of the default condition, is excluded from the targets; and an inference unit that outputs the indoor unit to be excluded from the targets, by using a learned model for determining an indoor unit to be excluded.

Description

Technical Field

[0001] The present disclosure relates to an air conditioning system.

Background Art

[0002] Conventionally, an air conditioning system having a configuration in which a plurality of indoor units are connected to one outdoor unit is known. Such an air conditioning system is adopted in office buildings, commercial facilities, and the like.

[0003] International Publication No. 2018 / 220803 (Patent Document 1) describes that the target evaporation temperature of the refrigerant system is adjusted based on the detection values of the temperature and humidity detection means provided in at least one of the plurality of indoor units to achieve energy saving. Japanese Unexamined Patent Application Publication No. 2013-152071 (Patent Document 2) describes that, among a plurality of indoor units, the indoor unit in the state with the highest required capacity value is excluded from the selection target, and it is determined whether to adjust the evaporation temperature based on the required capacities of the indoor units included in the selection target.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] In Japanese Patent Application Laid-Open No. 2013-152071 (Patent Document 2), conditions for determining whether to increase the evaporation temperature are provided for each of a plurality of indoor units. When all the indoor units satisfy the conditions, control is performed to increase the evaporation temperature of the refrigerant circuit. When there is an indoor unit installed near a heating element, control to increase the evaporation temperature of the refrigerant circuit may not be performed because the indoor unit near the heating element does not satisfy the conditions. Therefore, in Japanese Patent Application Laid-Open No. 2013-152071 (Patent Document 2), an indoor unit that continues to be in a state with a high required capacity value is excluded from the indoor units subject to the conditions.

[0006] However, when the evaporation temperature is increased by excluding an indoor unit, the comfort of the room in which the excluded indoor unit is installed may be impaired. When the room in which the excluded indoor unit is installed is a room of high importance to the user, such as a president's office or a reception room, the user may desire comfort rather than energy savings.

[0007] The present disclosure has been made to solve such problems, and an object thereof is to provide an air conditioning system having a plurality of indoor units and performing control to increase the evaporation temperature, capable of achieving a balance between energy savings due to an increase in the evaporation temperature and indoor comfort.

Means for Solving the Problems

[0008] The air conditioning system in the present disclosure includes a compressor, an outdoor unit, a first indoor unit, a second indoor unit, a refrigeration cycle that connects the compressor, the outdoor unit, the first indoor unit, and the second indoor unit to circulate refrigerant, a first temperature sensor that detects a first indoor air temperature corresponding to the first indoor unit, a second temperature sensor that detects a second indoor air temperature corresponding to the second indoor unit, and a control device that controls the operating frequency of the compressor using a predetermined condition based on the indoor air temperature corresponding to the indoor unit. At least one of the first indoor unit and the second indoor unit is defined as an object of the predetermined condition. When all the indoor units defined as objects satisfy the predetermined condition, the control device controls the operating frequency of the compressor so that the evaporation temperature of the refrigeration cycle becomes the first evaporation temperature. When at least one of the indoor units defined as objects does not satisfy the predetermined condition, the control device controls the operating frequency of the compressor so that the evaporation temperature of the refrigeration cycle becomes a second evaporation temperature lower than the first evaporation temperature, and acquires, as state data, data indicating a period during which a predetermined condition that increases when the indoor unit defined as an object is excluded from the object is satisfied. The control device includes an inference unit that outputs an indoor unit to be excluded from the object based on the state data acquired by the data acquisition unit using a learned model for determining an indoor unit to be excluded from the object.

Effect of the Invention

[0009] According to the present disclosure, in an air conditioning system having a plurality of indoor units and performing control to increase the evaporation temperature, it is possible to achieve a balance between energy saving due to the increase in the evaporation temperature and indoor comfort.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

MODE FOR CARRYING OUT THE INVENTION

[0011] Hereinafter, embodiments of the technical idea according to the present disclosure will be described with reference to the drawings. In the following description, the same parts are denoted by the same reference numerals. Their names and functions are also the same. Therefore, detailed descriptions thereof will not be repeated.

[0012] <Overall Configuration of Air Conditioning System> FIG. 1 is an overall configuration diagram of an air conditioning system 100. The air conditioning system 100 in the present embodiment is installed, for example, in an office building, a commercial facility, or the like. Hereinafter, an outline of each component included in the air conditioning system 100 will be described. The air conditioning system 100 includes a control device 50, an outdoor unit 20, indoor units 30A to 30H, an outdoor unit 21, and indoor units 31A to 31H. In FIG. 1, for simplicity of explanation, the illustration of the indoor units 30C to 30G and the indoor units 31C to 31G is simplified. Hereinafter, the indoor units 30A to 30H may be collectively referred to simply as "indoor unit 30". Also, the indoor units 31A to 31H may be collectively referred to simply as "indoor unit 31".

[0013] The control device 50 is electrically connected to the outdoor unit 20 and the outdoor unit 21. The control device 50 includes, for example, a CPU (Central Processing Unit), a storage device (including, for example, a ROM (Read Only Memory) and a RAM (Random Access Memory)), an input / output buffer, and the like. The control device 50 can adjust the evaporation temperature of the refrigerant in the indoor units 30, 31 or the outdoor units 20, 21 by the CPU executing a program stored in the storage device. In the present embodiment, the control device 50 is installed, for example, in the management room of the office building where the air conditioning system 100 is installed.

[0014] Note that the control device 50 may be provided as a cloud server in the air conditioning system 100. That is, the control device 50 can be installed at a location different from the office building where the air conditioning system 100 is installed. The control device 50 includes a display unit 55 that displays various information. The display unit 55 is realized by, for example, a liquid crystal display or an organic EL (Electro Luminescence) display.

[0015] The office building where the air conditioning system 100 is installed includes a plurality of rooms. Each of the indoor units 30A to 30H and 31A to 31H is installed in a room within the office building where the air conditioning system 100 is installed. In other words, each of the indoor units 30A to 30H and 31A to 31H is associated with a room in the office building. Note that in some cases, a plurality of indoor units may be installed in one room.

[0016] The outdoor unit 20 and the indoor units 30A to 30H are connected by pipes that circulate refrigerant to form a refrigeration cycle 11. Also, the outdoor unit 21 and the indoor units 31A to 31H are connected by pipes that circulate refrigerant to form a refrigeration cycle 12. The refrigeration cycle 11 constituted by the outdoor unit 20 and the indoor units 30A to 30H is constituted by pipes different from the pipes of the refrigeration cycle 12 constituted by the outdoor unit 21 and the indoor units 31A to 31H. In other words, the pipes forming the refrigeration cycle 11 and the pipes forming the refrigeration cycle 12 do not communicate with each other.

[0017] In the air conditioning system 100 of the present embodiment, the indoor units 30A to 30H are installed, for example, in the rooms on the first floor of the office building, and the indoor units 31A to 31H are installed in the rooms on the second floor of the office building. The control device 50 includes an input device (not shown). The set temperature, which is the target room temperature of the rooms where the indoor units 30A to 30H and the indoor units 31A to 31H are installed, is input from the user to the input device. The input device is installed, for example, in the management room of the above-described office building. The control device 50 controls the compressors included in the refrigeration cycles 11 and 12 according to the set temperature.

[0018] Also, as will be described later, each of the indoor units 30A to 30H and 31A to 31H is connected to an input device installed indoors. The user can change the set temperature from indoors where the indoor units 30A to 30H and 31A to 31H are installed.

[0019] Thus, in the air conditioning system 100 according to the present embodiment, a plurality of indoor units 30A to 30H and 31A to 31H are respectively connected to one outdoor unit 20, 21, and the control device 50 comprehensively controls the compressors included in the respective refrigeration cycles 11, 12. The indoor unit 30A may correspond to the "first indoor unit" of the present disclosure. Also, the indoor unit 30B may correspond to the "second indoor unit" of the present disclosure.

[0020] <Configuration of Refrigeration Cycle and Circulation of Refrigerant> FIG. 2 is a diagram showing the configuration of the refrigeration cycle 11. The refrigeration cycle 11 is configured to circulate the refrigerant enclosed in the piping within the circulation flow path L1. In the refrigeration cycle 11, the compressor 10, the four-way valve 15, the outdoor unit 20, the expansion valves 25A to 25H, and the indoor units 30A to 30H are connected by piping. In FIG. 2, for simplicity of explanation, the illustration of the expansion valves 25C to 25G and the indoor units 30C to 30G is simplified.

[0021] The four-way valve 15 switches the flow path of the refrigerant flowing through the circulation flow path L1. In the present embodiment, the state of the four-way valve 15 switches between a first state and a second state, and FIG. 2 shows the circulation flow path L1 when the four-way valve 15 is in the first state. When the four-way valve 15 is in the first state, the refrigerant in the refrigeration cycle 11 circulates in the order of the compressor 10, the four-way valve 15, the outdoor unit 20, the expansion valves 25A to 25H, and the indoor units 30A to 30H. The type of refrigerant in the refrigeration cycle 11 is, for example, an HFC refrigerant or a natural refrigerant.

[0022] When the four-way valve 15 is switched from the first state to the second state, the refrigerant in the refrigeration cycle 11 circulates in the order of the compressor 10, the four-way valve 15, the indoor units 30A to 30H, the expansion valves 25A to 25H, and the outdoor unit 20. That is, depending on the state of the four-way valve 15, the flow direction of the refrigerant in the pipes connecting the four-way valve 15, the indoor units 30A to 30H, the expansion valves 25A to 25H, and the outdoor unit 20 is reversed.

[0023] Hereinafter, the configuration of the air conditioning system 100 will be described in the order of the refrigerant flow when the four-way valve 15 is in the first state (the state shown in FIG. 2). The compressor 10 is configured to compress the gaseous refrigerant in the circulation flow path L1. The refrigerant discharged from the compressor 10 becomes a superheated gas state of high temperature and high pressure. The direction D is the direction in which the compressor 10 discharges the refrigerant. Hereinafter, in the circulation flow path L1, the direction in which the refrigerant flows from an arbitrary position will be referred to as "downstream", and the direction opposite to "downstream" and in which the refrigerant flows will be referred to as "upstream" in some cases. For example, the compressor 10 is arranged upstream of the four-way valve 15, and the four-way valve 15 is arranged downstream of the compressor 10.

[0024] When the four-way valve 15 is in the first state, the outdoor unit 20 functions as a condenser. By the blowing of the fan F20, the gaseous refrigerant passing through the outdoor unit 20 exchanges heat with the air around the outdoor unit 20. As a result, the refrigerant passing through the outdoor unit 20 is condensed into a liquid refrigerant. The liquid refrigerant passes through the branch point BP1 and flows into each of the expansion valves 25A to 25H. The liquid refrigerant is depressurized by the expansion valves 25A to 25H and becomes a refrigerant in a gas-liquid two-phase state.

[0025] The refrigerant in the gas-liquid two-phase state flows into the indoor units 30A to 30H. When the four-way valve 15 is in the first state, the indoor units 30A to 30H function as evaporators. In the indoor units 30A to 30H, the refrigerant in the gas-liquid two-phase state exchanges heat with the air around the indoor units 30A to 30H by the blowing of the fans FA to FH. In FIG. 2, for simplicity of explanation, the illustration of the fans FC to FG is simplified. As a result, a part of the refrigerant in the gas-liquid two-phase state passing through the indoor units 30A to 30H evaporates and becomes a gaseous refrigerant. The gaseous refrigerant merges at the branch point BP2 and flows into the four-way valve 15.

[0026] Thereafter, the gas refrigerant returns to the compressor 10 and is compressed again by the compressor 10. An accumulator may be installed in the circulation flow path L1 downstream of the four-way valve 15 and upstream of the compressor 10. Thus, in the air conditioning system 100, a refrigeration cycle 11 in which the refrigerant circulates through each component shown in FIG. 2 is formed in the circulation flow path L1. When the state of the four-way valve 15 is the second state, the outdoor unit 20 functions as an evaporator, and the indoor units 30A to 30H function as condensers. The refrigeration cycle 12 has the same configuration as the refrigeration cycle 11. Therefore, the description of the refrigeration cycle 12 will not be repeated.

[0027] The control device 50 controls the amount of refrigerant discharged by the compressor 10 per unit time by adjusting the operating frequency of the compressor 10. The control device 50 can increase the energy saving effect in the air conditioning system 100 by decreasing the operating frequency of the compressor 10. By decreasing the operating frequency of the compressor 10, the evaporation temperature at which the refrigerant evaporates in the evaporator increases. The control of decreasing the operating frequency of the compressor 10 to increase the evaporation temperature is referred to as "high sensible heat control". Note that "high sensible heat control" may also be referred to as "evaporation temperature control". By executing the high sensible heat control, the energy consumed by the compressor 10 can be reduced. The evaporation temperature after the high sensible heat control is executed may correspond to the "first evaporation temperature" in the present disclosure. The evaporation temperature before the high sensible heat control is executed may correspond to the "second evaporation temperature" in the present disclosure.

[0028] The control device 50 determines whether to execute high latent heat control for each of the refrigeration cycles 11 and 12 each including a compressor. The control device 50 uses a predetermined condition to determine whether to execute high latent heat control. The control device 50 designates a plurality of indoor units among the indoor units 30 as the target of the predetermined condition, and determines whether the predetermined condition is satisfied for the target indoor units. When all of the indoor units that are the target of the predetermined condition satisfy the predetermined condition, the control device 50 executes high latent heat control, and when at least one of the indoor units that are the target of the predetermined condition does not satisfy the predetermined condition, the control device 50 does not execute high latent heat control.

[0029] Hereinafter, the predetermined condition regarding the high latent heat control for the refrigeration cycle 11 will be specifically described. In the initial setting, all of the indoor units 30A to 30H included in the refrigeration cycle 11 are set as the target of the predetermined condition. Hereinafter, the indoor units that are the target of the predetermined condition are referred to as "target units".

[0030] In the present embodiment, the predetermined condition is a condition as to whether the temperature difference between the set temperature set in the indoor unit and the temperature of the indoor air pressed against the indoor unit is within a predetermined range. The temperature of the indoor air is acquired by a sensor described later. More specifically, the temperature difference is the difference when the set temperature set in the indoor unit is subtracted from the temperature of the indoor air during the execution of the cooling operation. Note that during the execution of the heating operation, it is the difference when the temperature of the indoor air is subtracted from the set temperature set in the indoor unit.

[0031] When the temperature difference between the indoor air temperature and the set temperature set in each of the indoor units 30A to 30H that are the target units is within the predetermined range, the control device 50 executes the high latent heat control for the refrigeration cycle 11. When there is an indoor unit in the target units where the temperature difference between the air temperature and the set temperature is outside the predetermined range, the control device 50 does not execute the high latent heat control. The explanation of the execution condition of the high latent heat control will be described in more detail later.

[0032] <Configuration of the control device 50> FIG. 3 is a block diagram showing the configuration of the air conditioning system 100. As shown in FIG. 3, in addition to the indoor units 30A to 30H, 31A to 31H, the outdoor units 20, 21, and the control device 50, the air conditioning system 100 includes input devices 32A to 32H, 33A to 33H, sensors 35A to 35H, 36A to 36H, and an external terminal 60. Hereinafter, the sensors 35A to 35H may be collectively referred to simply as "sensor 35". Also, the sensors 36A to 36H may be collectively referred to simply as "sensor 36". Furthermore, the input devices 32A to 32H may be collectively referred to simply as "input device 32". Also, the input devices 33A to 33H may be collectively referred to simply as "input device 33".

[0033] The sensors 35A to 35H are respectively provided in the rooms where the indoor units 30A to 30H are installed. The sensors 35A to 35H detect the indoor air temperature of the rooms where the indoor units 30A to 30H are installed. In other words, the sensors 35A to 35H detect the air temperature of the air sucked in by the indoor units 30A to 30H. Similarly, the sensors 36A to 36H are respectively provided in the rooms where the indoor units 31A to 31H are installed. In other words, the sensors 36A to 36H detect the air temperature of the air sucked in by the indoor units 31A to 31H.

[0034] Note that the sensors 35A to 35H and the sensors 36A to 36H may be respectively arranged inside the indoor units 30A to 30H and the indoor units 31A to 31H. The sensors 35, 36 are, for example, temperature sensors such as thermistors. Also, each of the sensors 35, 36 may include a humidity sensor.

[0035] The sensor 35A may correspond to the "first temperature sensor" in the present disclosure. The sensor 35B may correspond to the "second temperature sensor" in the present disclosure. The air temperature detected by the sensor 35A may correspond to the "first air temperature" in the present disclosure. The air temperature detected by the sensor 35B may correspond to the "second air temperature" in the present disclosure.

[0036] Each of the sensors 35 and 36 transmits the detected air temperature to the control device 50. That is, the control device 50 acquires the detection values of each of the sensors 35 and 36 via the indoor units 30 and 31 and the outdoor units 20 and 21. Note that the control device 50 may directly acquire the detection values from the sensors 35 and 36.

[0037] The input devices 32A to 32H are respectively provided in the rooms where the indoor units 30A to 30H are installed. The input devices 33A to 33H are respectively provided in the rooms where the indoor units 31A to 31H are installed. The input devices 32A to 32H and 33A to 33H receive the set temperature from the users of the rooms where the indoor units 30A to 30H and 31A to 31H are installed. That is, the user can change the set temperature from the input devices 32A to 32H and 33A to 33H installed indoors when feeling that the indoor comfort has decreased. The input devices 32A to 32H and 33A to 33H are, for example, panels having buttons.

[0038] Each of the input devices 32A to 32H and 33A to 33H transmits the set temperature received from the indoor user to the control device 50. That is, the control device 50 acquires the detection values of each of the input devices 32 and 33 via the indoor units 30 and 31 and the outdoor units 20 and 21. Note that the control device 50 may directly acquire the set temperature input to the input devices 32 and 33 from the input devices 32 and 33.

[0039] As shown in FIG. 3, the control device 50 includes a totaling unit 51, a storage unit 52, a setting unit 53, a calculation unit 54, a display unit 55, a determination unit 56, and an output unit 57. The totaling unit 51 receives the detection values of each of the sensors 35 and 36 at predetermined intervals. For example, the totaling unit 51 acquires the detection values from each of the sensors 35 and 36 every 10 seconds and writes them into the storage unit 52. Note that the predetermined interval may be an interval other than 10 seconds. For example, it may be 1 second or 2 minutes. Further, the totaling unit 51 stores the set temperature input to the input devices 32 and 33 in the storage unit 52 as set temperature change history data together with the information indicating the input date and time.

[0040] The storage unit 52 is a non-volatile memory, such as a hard disk, SSD (Solid State Drive), CD-ROM, FD (Flexible Disk), magnetic tape, cassette tape, optical disk (MO (Magnetic Optical Disc) / MD (Mini Disc) / DVD (Digital Versatile Disc)), IC (Integrated Circuit) card (including memory card), optical card, mask ROM, EPROM (Electronically Programmable Read-Only Memory), EEPROM (Electronically Erasable Programmable Read-Only Memory), flash ROM, etc., which is a recording medium that fixedly carries data and programs.

[0041] The processes executed by the arithmetic unit 54, the determination unit 56, and the output unit 57 are realized by the cooperation of software executed by the CPU. The arithmetic unit 54 calculates the temperature difference between the air temperature and the set temperature for each of the indoor units 30A to 30H and the indoor units 31A to 31H. The arithmetic unit 54 transmits the temperature difference between the air temperature and the set temperature of each of the indoor units 30A to 30H and the indoor units 31A to 31H to the determination unit 56.

[0042] The setting unit 53 transmits to the arithmetic unit 54 the conditions related to the high latent heat control determined in advance. For example, the setting unit 53 transmits the range of the temperature difference that is the condition for whether to perform the high latent heat control. The range of the temperature difference is, for example, within +1°C. The control device 50 executes the high latent heat control of the refrigeration cycle 11 when the temperature difference between the air temperature and the set temperature is within the range of the temperature difference set by the setting unit 53 in all of the target machines.

[0043] Specifically, for example, when the set temperature of the indoor unit 30A is 26°C and the air temperature of the indoor unit 30A is 26.5°C, the temperature difference when subtracting the set temperature from the air temperature in the indoor unit 30A is +0.5°C. Since the temperature difference between the air temperature and the set temperature is within the range of the temperature difference set by the setting unit 53, the indoor unit 30A satisfies the conditions for executing the high latent heat control.

[0044] On the other hand, for example, when the set temperature of the indoor unit 30A is 26°C and the air temperature of the indoor unit 30A is 28°C, the temperature difference between the air temperature and the set temperature in the indoor unit 30A is +2°C. In this case, the indoor unit 30A does not satisfy the established conditions for executing the high latent heat control. When there is an indoor unit that does not satisfy the established conditions in the target indoor unit, the control device 50 does not execute the high latent heat control.

[0045] In this way, the calculation unit 54 calculates whether the conditions for executing the high latent heat control are satisfied for each indoor unit based on the air temperature acquired from the sensors 35 and 36, the set temperature set in the indoor units 30 and 31, and the range of the temperature difference acquired from the setting unit 53. Thus, the established conditions are the conditions for determining whether to execute the high latent heat control for each indoor unit, and are the conditions for whether the temperature difference between the air temperature and the set temperature is within the range of the temperature difference set by the setting unit 53. Hereinafter, the range of the temperature difference set by the setting unit 53 is referred to as the "setting range".

[0046] Note that the default condition may be other conditions. For example, the storage unit 52 stores the air temperature and the set temperature as a history. When the ratio of the period during which the temperature difference between the air temperature and the set temperature is within a predetermined range is longer than a predetermined ratio, the control device 50 may determine that the default condition is satisfied. For example, when the indoor unit 30A has a temperature difference between the air temperature and the set temperature within a predetermined range at a ratio of 90% in the most recent one hour from the history, the control device 50 determines that the indoor unit 30A satisfies the default condition. That is, if the period during which the temperature difference between the air temperature and the set temperature is within a predetermined range is long within the period stored as the history, it is highly likely that the temperature difference between the air temperature and the set temperature is also within the predetermined range at present.

[0047] The calculation unit 54 transmits the calculation result of the temperature difference calculated for each of the indoor units 30A to 30H and the indoor units 31A to 31H to the determination unit 56. The determination unit 56 determines whether all of the target units satisfy the conditions for executing the high latent heat control. As described above, in the initial setting, the target units are all the indoor units 30A to 30H included in the refrigeration cycle 11, and in the refrigeration cycle 12, the target units are all the indoor units 31A to 31H included in the refrigeration cycle 12. That is, the target units in the initial setting are all the indoor units included in the refrigeration cycle.

[0048] The storage unit 52 stores which indoor unit is set as the target unit, and it can be changed by accepting the setting from the user. Further, the storage unit 52 stores the setting range set by the setting unit 53. The user can use an input device (not shown) included in the control device 50 to set to exclude the indoor unit 30A from the target units. Hereinafter, such a setting to exclude an indoor unit from the target unit is referred to as an "exclusion setting".

[0049] When the difference between the set temperature and the air temperature in all of the target units is within a predetermined range, the determination unit 56 determines to execute the high latent heat control. When the difference between the set temperature and the air temperature is not within the predetermined range in at least one of the indoor units included in the target unit, the determination unit 56 determines not to execute the high latent heat control.

[0050] The output unit 57 outputs the determination result to the display unit 55, and the display unit 55 displays the determination result. Also, the output unit 57 outputs the determination result to the external terminal 60, and the external terminal 60 displays the determination result. The external terminal 60 is, for example, a general-purpose PC, a smartphone, or a tablet terminal. Thus, when the control device 50 determines whether each indoor unit satisfies the predetermined conditions and all of the target devices satisfy the predetermined conditions, it executes high latent heat control to increase the evaporation temperature of the refrigeration cycle.

[0051] <Specific example of temperature transition> FIG. 4 is a diagram showing a first example for explaining the details of the predetermined conditions indicating whether to execute high latent heat control. In FIG. 4, a coordinate system is shown in which the vertical axis represents the air temperature and the horizontal axis represents the time. Also, eight waveforms are shown as lines LnA to LnH in the coordinate system. The lines LnA to LnH respectively represent the detection values of the sensors 35A to 35H. That is, in FIG. 4, the transitions of the air temperatures of the indoor units 30A to 30H in the refrigeration cycle 11 are shown as waveforms. In the examples of FIGS. 4 and 5, all of the indoor units 30A to 30H are set as target devices.

[0052] In the example of FIG. 4, the set temperature of each of the indoor units 30A to 30H is uniformly set to 26°C. The setting unit 53 sets +1°C as the setting range as described above. For this reason, as shown in FIG. 4, the temperature serving as the threshold for determining whether to execute high latent heat control is 27°C.

[0053] In periods Dr11 and Dr13, the detection value of the sensor 35D shown as line LnD exceeds 27°C. Also, in period Dr15, the detection values of the sensors 35D and 35G shown as lines LnD and LnG respectively exceed 27°C. Further, in period Dr17, the detection values of the sensors 35D, 35G, and 35C shown as lines LnD, LnG, and LnC respectively exceed 27°C.

[0054] During periods Dr11, Dr13, Dr15, and Dr17, the target unit includes an indoor unit that does not meet the preset conditions. Therefore, the control device 50 does not execute the high latent heat control during periods Dr11, Dr13, Dr15, and Dr17 when there is an indoor unit that does not meet the preset conditions. On the other hand, during periods Dr12, Dr14, and Dr16, there is no indoor unit that does not meet the preset conditions among the target units, and all of the target units meet the preset conditions, so the control device 50 executes the high latent heat control. That is, during periods Dr12, Dr14, and Dr16, the compressor 10 is controlled to increase the evaporation temperature of the refrigeration cycle 11.

[0055] FIG. 5 is a diagram showing a second example for explaining the details of the preset conditions for determining whether to execute the high latent heat control. In FIG. 5, similar to FIG. 4, a coordinate system is shown in which the vertical axis represents the air temperature and the horizontal axis represents the time, and lines LnA to LnH indicating the detection values of sensors 35A to 35H are shown. Also in FIG. 5, the setting unit 53 sets +1°C as the range of the temperature difference that is the condition for performing the high latent heat control.

[0056] In FIG. 5, the set temperature of the indoor units 30A to 30C, 30E to 30H is set to 26°C. On the other hand, different from FIG. 4, the set temperature of the indoor unit 30D is set to 22°C. That is, the temperature that is the threshold for determining whether to execute the high latent heat control for the indoor units 30A to 30C, 30E to 30H is 27°C, and the temperature that is the threshold for determining whether to execute the high latent heat control for the indoor unit 30D is 23°C.

[0057] During periods Dr21, Dr23, Dr25, and Dr27, the detection value of the sensor 35D shown as the line LnD exceeds 23°C. Also, during period Dr26, the detection value of the sensor 35G shown as the line LnG exceeds 27°C. Further, during period Dr29, the detection values of the sensors 35G and 35C shown as the lines LnG and LnC respectively exceed 27°C.

[0058] That is, the control device 50 does not execute high latent heat control during periods Dr21, Dr23, Dr25, Dr26, Dr27, and Dr29. On the other hand, during periods Dr22, Dr24, and Dr28, since there is no indoor unit among the target units that does not satisfy the predetermined conditions and all of the target units satisfy the predetermined conditions, the control device 50 executes high latent heat control. That is, during periods Dr22, Dr24, and Dr28, the compressor 10 is controlled so that the evaporation temperature of the refrigeration cycle 11 increases.

[0059] FIG. 6 is a diagram showing a third example for explaining the details of the predetermined conditions for determining whether to execute high latent heat control. In FIG. 6, similar to FIGS. 4 and 5, a coordinate system is shown in which the vertical axis represents the air temperature and the horizontal axis represents time, and lines LnA to LnH indicating the detection values of sensors 35A to 35H are shown. Also in FIG. 6, the setting unit 53 sets +1°C as the range of the temperature difference that is the condition for performing high latent heat control. In FIG. 6, 26°C is set as the set temperature of the indoor units 30A to 30H, similar to FIG. 4. That is, the temperature that is the threshold for determining whether to execute high latent heat control for the indoor units 30A to 30H is 27°C.

[0060] During periods Dr32 and Dr34, the detection value of sensor 35D shown as line LnD exceeds 27°C. Also during period Dr36, the detection values of sensors 35G, 35C, and 35D shown as lines LnG, LnC, and LnD respectively exceed 27°C.

[0061] That is, the control device 50 does not execute high latent heat control during periods Dr32, Dr34, and Dr36. On the other hand, during periods Dr31, Dr33, and Dr35, since there is no indoor unit among the target units that does not satisfy the predetermined conditions and all of the target units satisfy the predetermined conditions, the control device 50 executes high latent heat control. That is, during periods Dr31, Dr33, and Dr35, the compressor 10 is controlled to increase the evaporation temperature of the refrigeration cycle 11.

[0062] In the example of FIG. 6, the sensor 35D is installed near the door. Therefore, every time the door is opened, it is exposed to the outside air, and the detected value of the sensor 35D sporadically increases and exceeds 30° C. during the periods Dr31, Dr33, and Dr35. In this way, the sensor 35D satisfies the default conditions while the door is closed and does not satisfy the default conditions while the door is open. That is, the increase in the detected value of the sensor 35D is not due to a constant decrease in the comfort level indoors, but is due to the temporary inflow of outside air into the room.

[0063] As shown in FIGS. 4, 5, and 6, among the target indoor units 30A to 30H, high-sensible heat control may not be executed because some of the indoor units do not satisfy the default conditions. The user can achieve an energy-saving effect in the refrigeration cycle by excluding the indoor units that do not satisfy the default conditions from the target units. In particular, when the failure to satisfy the default conditions as shown in FIG. 6 is not due to a constant decrease in the comfort level indoors but is due to being temporarily exposed to a heating element or the like, the user's comfort can be maintained while achieving an energy-saving effect in the refrigeration cycle by making the exclusion setting.

[0064] <History data related to high-sensible heat control> FIG. 7 is a diagram showing the history of the applicability of the conditions for each indoor unit for high-sensible heat control during a predetermined period. The control device 50 can store the detected values acquired from the sensors 35 and 36 as a history in the storage unit 52 and generate the data shown in FIGS. 7 to 10. In FIGS. 7 to 10, the history data is actually generated regardless of which indoor unit was the exclusion target. That is, the same data is generated in both the case where no exclusion setting was made during the predetermined period and the case where a predetermined indoor unit was excluded. Note that after generating the data shown in FIGS. 7 to 10, the control device 50 may display the generated data on the display unit 55.

[0065] Figure 7 shows in tabular form data indicating the applicability of the default conditions for each of the indoor units 30A to 30H from an arbitrary timing up to one hour before. The arbitrary timing is, for example, when a generation command for data indicating the applicability of the conditions for each indoor unit for high latent heat control is received from the user. The predetermined period is not limited to the most recent one hour, and may be, for example, one day before, one week before, or one month before from the timing specified by the user. The predetermined period can be specified by the user together with the generation command for data indicating the applicability of the conditions for each indoor unit for high latent heat control.

[0066] In the tabular data of Figure 7, the applicability of the default conditions for each of the indoor units 30 and 31 every two minutes within the most recent one hour is shown. More specifically, the column labeled "00" which is the leftmost column shows the applicability of the default conditions for indoor unit 30 in the two minutes starting from 60 minutes back from when the generation command was received. Also, the column labeled "20" shows the applicability of the default conditions for indoor unit 30 in the two minutes starting from 40 minutes back from when the generation command was received. Furthermore, the rightmost column labeled "58" shows the applicability of the default conditions for indoor unit 30 in the most recent two minutes from when the generation command was received.

[0067] After Figure 7, a state where indoor unit 30 does not satisfy the default conditions is illustrated as "OFF". That is, in Figure 7, a period during which the temperature difference between the indoor air temperature and the set temperature of the indoor unit is not within the predetermined range is represented as "OFF". Specifically, for indoor unit 30A, in the period from when 36 minutes have elapsed to when 42 minutes have elapsed, since the temperature difference between the indoor air temperature and the set temperature of the indoor unit is not within the predetermined range, "OFF" is displayed at the corresponding location in Figure 7. Also, for indoor unit 30D, in the period from when 10 minutes have elapsed to when 30 minutes have elapsed, since the temperature difference between the indoor air temperature and the set temperature of the indoor unit is not within the predetermined range, "OFF" is displayed at the corresponding location in Figure 7.

[0068] Below the lower part of FIG. 7, the periods during which the high latent heat control is not executed by the control device 50 are shown by being divided according to the presence or absence of the exclusion setting. The period during which the high latent heat control is not executed is represented as "OFF". As shown in FIG. 7, when no exclusion setting is made, the control device 50 executes the high latent heat control for 14 minutes within the most recent one hour.

[0069] The control device 50 can generate a table showing the execution period of the high latent heat control when an exclusion setting is made for the indoor unit 30D. In FIG. 7, when the indoor unit 30D is excluded from the target machines, that is, when all of the indoor units 30A to 30C, 30E to 30H satisfy the predetermined conditions regardless of the detection value of the sensor 35D, the period during which the high latent heat control is executed is displayed. When an exclusion setting is made for the indoor unit 30D, the control device 50 executes the high latent heat control for 36 minutes within the most recent one hour.

[0070] Furthermore, the control device 50 can generate a table showing the execution period of the high latent heat control when an exclusion setting is made for the indoor unit 30G in addition to the indoor unit 30D. When an exclusion setting is made for the indoor unit 30D and the indoor unit 30G, the control device 50 executes the high latent heat control for 44 minutes within the most recent one hour. In the example of FIG. 7, an example is shown in which the indoor unit 30D with the longest "OFF" period and the indoor unit 30G with the second longest "OFF" period after the indoor unit 30D are excluded, but data when other indoor units are excluded may also be generated. The data shown in FIG. 7 can be output by the output unit 57 and displayed by the display unit 55.

[0071] FIG. 8 is a diagram showing the execution period of the high latent heat control increased by the exclusion setting. The control device 50 may generate the table shown in FIG. 8 in addition to the table shown in FIG. 7 and display it on the display unit 55. As described with reference to FIG. 7, when no exclusion setting is made, the high latent heat control is executed for 14 minutes. When the indoor unit 30D is excluded, the high latent heat control is executed for 36 minutes. When the indoor unit 30D and the indoor unit 30G are excluded, the high latent heat control is executed for 44 minutes.

[0072] In addition to the execution period of the high latent heat control, the control device 50 causes the increased execution period of the high latent heat control due to the exclusion setting to be displayed as the increased period in the right column. When the indoor unit 30D is excluded, the execution period of the high latent heat control increases by 22 minutes compared to the case where no exclusion setting is made. When the indoor units 30D and 30G are excluded, the execution period of the high latent heat control increases by 30 minutes compared to the case where no exclusion setting is made. As described above, in FIGS. 7 and 8, the same table is generated regardless of which indoor unit is actually excluded. Thus, the control device 50 in the present embodiment calculates the execution period of the high latent heat control when no exclusion setting is made in the most recent one hour and the execution period of the high latent heat control when an exclusion setting is made.

[0073] FIG. 9 is a first figure that generates the total period of the periods that do not satisfy the predetermined conditions as data in a ranking format. FIG. 9 shows data for comparing the total periods of the periods that do not satisfy the predetermined conditions for the indoor units 30A to 30H included in the refrigeration cycle 11. The control device 50 may cause the data shown in FIG. 9 to be displayed on the display unit 55.

[0074] More specifically, as shown in FIG. 9, the control device 50 can display in a ranking format using a bar graph. The indoor units 30A to 30H are displayed in order from the indoor unit 30D having the longest period that does not satisfy the predetermined conditions. Thereby, in the air conditioning system 100, it is possible to easily make the user recognize which of the plurality of indoor units is an indoor unit that hinders the execution of the high latent heat control.

[0075] FIG. 10 is a second diagram for explaining the display in a ranking format of the total period of a period that does not satisfy the predetermined conditions. In FIG. 10, the total period of the periods that do not satisfy the predetermined conditions for the indoor units 31A to 31H included in the refrigeration cycle 12 is shown. Although not shown, in the present embodiment, the control device 50 generates table data corresponding to FIGS. 7 and 8 also for the refrigeration cycle 12. Thereby, the control device 50 can display, in a ranking format, the total period of the periods that do not satisfy the predetermined conditions also for the refrigeration cycle 12. The control device 50 can cause the display unit 55 to display the data corresponding to FIGS. 7 and 8 of the refrigeration cycle 12 and the data shown in FIG. 10.

[0076] Comparing FIG. 9 and FIG. 10, the user can grasp that the period in which the indoor unit 30D does not satisfy the predetermined conditions in FIG. 9 is more prominent than the other indoor units 30A to 30C, 30E to 30H. From FIG. 9, the user can grasp that the indoor unit 30D is an indoor unit in which the period in which the deviation between the set temperature and the air temperature is large is long. Thereby, the user who refers to FIG. 9 can recognize that the increased period of the high latent heat control becomes large when the exclusion setting is performed on the indoor unit 30D. Thus, in the air conditioning system 100, the user can be made to recognize an indoor unit in which the deviation between the set temperature and the air temperature is large.

[0077] Thus, in the air conditioning system 100, the execution period of the high latent heat control can be displayed by comparing the case where the exclusion setting is performed from the history with the case where it is not performed. The user can use the data shown in FIGS. 7 to 10 as a reference for which indoor unit the exclusion setting should be performed on. That is, the data in FIGS. 7 to 10 can be usefully utilized for making an appropriate exclusion setting. However, the user needs to consider which indoor unit the exclusion setting should be performed on with reference to the data in FIGS. 7 to 10, which may impose a burden on the user.

[0078] On the one hand, if the target of the exclusion setting is uniformly set for the indoor units that most interfere with the execution of the high-profile heat control without allowing the user to determine it in order to reduce the user's burden, if the excluded room is a room of high importance to the user, the comfort may decrease contrary to the user's intention. Rooms of high importance to the user are, for example, a living room, a president's office, etc. Further, as described with reference to FIG. 6, when a sensor is installed near the door, if the indoor unit corresponding to the sensor is not excluded from the setting, the execution of the high-profile heat control will be disturbed every time the door is opened.

[0079] Therefore, as described below, the air conditioning system 100 according to the present embodiment is configured to output an exclusion candidate based on the state data indicating the increasing period of the high-profile heat control in FIGS. 7 to 10 and perform an exclusion setting based on the exclusion candidate. The exclusion candidate includes the case where no indoor unit is excluded and the case where no exclusion setting is performed.

[0080] <Automatic control of exclusion setting> FIG. 11 is a diagram showing a setting screen W1 for high-profile heat control. The setting screen W1 for high-profile heat control shown in FIG. 11 is displayed by the display unit 55. That is, the control device 50 displays the setting screen W1 shown in FIG. 11 for the user in a management room of an office building or the like.

[0081] The setting screen W1 shows the enable / disable of automatic control, the target machines and setting ranges of the refrigeration cycle 11, and the target machines and setting ranges of the refrigeration cycle 12. Buttons Bt1 and Bt2 are shown in the enable / disable column of the automatic control. Based on the selection of the button Bt1 by the user, the control device 50 performs automatic control of the exclusion setting using the AI described later. Further, based on the selection of the button Bt2 by the user, the control device 50 cancels the automatic control of the exclusion setting.

[0082] In the example shown in FIG. 11, indoor units 30A to 30H are set as the target units of the refrigeration cycle 11, and the setting range of the refrigeration cycle 11 set by the setting unit 53 is set within a range of +1°C. Also, indoor units 31A to 31H are set as the target units of the refrigeration cycle 12, and the setting range of the refrigeration cycle 12 set by the setting unit 53 is set within a range of +1°C. The user can change the target units and the setting ranges in the refrigeration cycles 11 and 12 from the screen shown in FIG. 11.

[0083] FIG. 12 is a flowchart showing the processing procedure of the automatic control of the exclusion setting executed by the control device 50. In the air conditioning system 100 of the present embodiment, automatic control is performed using the data indicating the increasing period of the high latent heat control described with reference to FIGS. 7 to 10 and the data indicating the indoor comfort. More specifically, the control device 50 determines an exclusion candidate to be the target of the exclusion setting using the data indicating the increasing period of the high latent heat control and the data indicating the indoor comfort. Further, the control device 50 determines whether the determined exclusion candidate is appropriate based on the actual execution period of the high latent heat control that has actually increased after the exclusion setting and the data indicating the indoor comfort after the exclusion setting. In FIG. 12, "S" is used as an abbreviation for "STEP".

[0084] The control device 50 determines whether the automatic control of the high latent heat control is effective (step S10). That is, the control device 50 determines whether the button Bt1 in FIG. 11 has been selected. When the automatic control of the high latent heat control is not effective (NO in step S10), the control device 50 repeats the process of step S10.

[0085] When the automatic control of the high-precision heat control is effective (YES in step S10), the control device 50 calculates the periods that satisfy and do not satisfy the predetermined conditions within a predetermined period for each indoor unit (step S100). That is, the control device 50 refers to the storage unit 52 and generates the table described with reference to FIG. 7. Subsequently, the control device 50 generates data indicating in a ranking format the total period of the periods that do not satisfy the predetermined conditions for each indoor unit (step S110). That is, the control device 50 generates the data described with reference to FIGS. 9 and 10.

[0086] Subsequently, the control device 50 calculates the execution period of the high-precision heat control that increases when the exclusion setting is performed (step S120). That is, the control device 50 calculates the execution period of the high-precision heat control that increases by performing the exclusion setting as described with reference to FIG. 8. The control device 50 acquires data indicating the comfort of each room (step S130). The data indicating the comfort of each room will be described in detail later.

[0087] Using the AI, the control device 50 determines exclusion candidates by taking as inputs the data indicating the increased period of the high-precision heat control acquired in steps S100 to S120 and the data indicating the comfort of the room, and makes an exclusion setting based on the determined exclusion candidates (step S140). Thereafter, the control device 50 acquires the data indicating the execution period of the high-precision heat control increased by the exclusion setting and the data indicating the comfort of the room, and updates the function using reinforcement learning (step S150).

[0088] The control device 50 determines again whether the automatic control of the high-precision heat control is effective (step S160). That is, the control device 50 determines whether the button Bt2 in FIG. 11 is selected. When the state where the automatic control of the high-precision heat control is effective is maintained (YES in step S160), the control device 50 returns the process to step S100. When the automatic control of the high-precision heat control is not effective (NO in step S160), the control device 50 ends the automatic control process of the high-precision heat control.

[0089] In this way, the control device 50 obtains the increasing period of the high sensible heat control in steps S100 to S120, and obtains the data indicating the comfort level in the room in step S130. The control device 50 determines the exclusion candidates using the data indicating the increasing period of the high sensible heat control and the data indicating the comfort level in the room as inputs, and performs an exclusion setting for the determined exclusion candidates. The control device 50 determines whether the exclusion candidates were appropriate after the exclusion setting using the execution period of the high sensible heat control increased by the exclusion setting and the data indicating the comfort level in the room after the exclusion setting, and updates the function.

[0090] <Status data> As described above, the control device 50 determines the exclusion candidates using the data indicating the increasing period of the high sensible heat control and the data indicating the comfort level in the room. The data indicating the increasing period of the high sensible heat control is the detection history data of the sensors 35 and 36 at an arbitrary timing, and indicates the state between the indoor units 30 and 31 and the high sensible heat control at the arbitrary timing. The data indicating the comfort level in the room is, for example, the change history data of the set temperature at an arbitrary timing. Hereinafter, the data indicating the increasing period of the high sensible heat control and the data indicating the comfort level in the room are referred to as "status data".

[0091] Based on the status data shown in FIG. 7, the control device 50 can estimate which indoor unit should be excluded to increase the execution period of the high sensible heat control. Specifically, if the execution period of the high sensible heat control that would increase when it is assumed that a predetermined indoor unit is excluded within the most recent one hour is longer than the threshold value, the control device 50 estimates that the increasing period of the high sensible heat control will be long by excluding the predetermined indoor unit. On the other hand, if the execution period of the high sensible heat control that would increase when it is assumed that a predetermined indoor unit is excluded within the most recent one hour is shorter than the threshold value, the control device 50 estimates that the increasing period of the high sensible heat control will be short by excluding the predetermined indoor unit.

[0092] In addition, the control device 50 estimates the comfort level of each room based on the change history data of the set temperature by the user. Specifically, if the number of times the set temperature is changed by the user within a predetermined period is less than the threshold value, the control device 50 estimates that the comfort level of the corresponding room has not decreased. On the other hand, if the number of times the set temperature is changed by the user within a predetermined period is greater than the threshold value, the control device 50 estimates that the comfort level has decreased. Alternatively, the comfort level may be determined based on the images acquired by the camera that images the room. For example, the control device 50 determines the comfort level of the room from the actions, expressions, etc. of the user in the room by performing image recognition using the captured data. Note that for image recognition, image recognition technologies using convolutional neural networks such as CNN (Convolutional Neural Network) may be used.

[0093] Based on the data indicating the increasing period of the high sensible heat control shown in FIGS. 7 to 10, and the data indicating the comfort level such as the change history data of the set temperature and the imaging data of the user by the camera after the start of the high sensible heat control, the control device 50 infers exclusion candidates by AI (Artificial Intelligence) and outputs the inferred exclusion candidates. The control device 50 generates control data for performing the exclusion setting based on the output exclusion candidates.

[0094] In this way, the air conditioning system 100 performs the exclusion setting using the data indicating the increasing period of the high sensible heat control and the exclusion candidates inferred based on the data indicating the comfort level of the room. Thereby, the air conditioning system 100 can increase the execution period of the high sensible heat control while avoiding a decrease in comfort due to the execution of the high sensible heat control. That is, in the air conditioning system 100, it is possible to achieve a balance between energy saving due to an increase in the evaporation temperature and the comfort level of the room.

[0095] <Reinforcement learning> As described above, the control device 50 outputs exclusion candidates based on data indicating the increase period of the high sensible heat control after the execution of the high sensible heat control and data indicating the comfort level of the room, and performs exclusion setting based on the exclusion candidates to realize appropriate high sensible heat control. To realize such control, the control device 50 learns control data for performing exclusion setting by reinforcement learning using AI.

[0096] FIGS. 13 and 14 are diagrams for explaining the outline of reinforcement learning. As shown in FIG. 13, in reinforcement learning, an agent (acting entity) in a certain environment observes the current state (parameters of the environment) and determines the action to be taken based on a policy. A policy is a rule for determining an action, and by optimizing the policy, the selection of actions is optimized. The action of the agent dynamically changes the environment, and the agent is given a reward based on a reward criterion in response to the change in the environment. The agent repeats such actions and learns the action that can obtain the most reward through a series of actions. In FIG. 13, the current state is represented by B2 (state), the action of the agent is represented by B1 (action), and the reward criterion is represented by D (reward criterion).

[0097] As typical methods of reinforcement learning, Q-learning and TD-learning are known. For example, in the case of Q-learning, the general update formula for the action value function Q(s, a) is represented by Equation (1).

[0098]

Equation

[0099] In Equation (1), S t represents the state of the environment at time t, and a t represents the action at time t. By the action a t , the state changes to S t+1 . r t+1represents the reward that the agent can obtain due to the change in its state, γ represents the discount rate, and α represents the learning coefficient. Note that γ is in the range of 0 < γ ≤ 1, and α is in the range of 0 < α ≤ 1. When B1 (action) becomes action a t and B2 (state) becomes state S t and the best action a t at state S t at time t is learned.

[0100] The update formula represented by Equation (1) is such that if the action value function Q of the action a with the highest Q value at time t + 1 is greater than the action value function Q of the action a executed at time t, the action value function Q is increased; otherwise, the action value function Q is decreased. In other words, the action value function Q(s, a) is updated so that the action value function Q of the action a at time t approaches the best action value at time t + 1. As a result, the best action value in a certain environment is sequentially propagated to the action values in the previous environment.

[0101] As shown in FIG. 14, in the learning phase, the control device 50 generates (updates) the learned model 170 based on the learning data 180 including B1 (action) and B2 (state) and D (reward criterion) by executing the learning program 160.

[0102] In the utilization phase, the control device 50 obtains C (output) from B2 (state) using the learned model 170.

[0103] FIG. 15 is a diagram for explaining B1 (action), B2 (state), C (output), and D (reward criterion) in the reinforcement learning according to the embodiment.

[0104] As shown in FIG. 15, in the air conditioning system 100 according to the present embodiment, control data for performing an exclusion setting based on an exclusion candidate is applied to B1 (action). State data including data indicating an increase period of the high latent heat control and data indicating the comfort of the room is applied to B2 (state). As described above, the state data includes at least one of the table data shown in FIGS. 7 to 10, the change history data of the set temperature, and the captured image of the user. An exclusion candidate is applied to C (output). Data indicating the execution period of the high latent heat control increased by the exclusion setting and data indicating the comfort of the room after the exclusion setting are applied to D (reward criterion).

[0105] In addition to the exclusion candidate, C (output) may include that the setting range set by the setting unit 53 is changed. In the example shown in FIG. 11, +1° C. is set as the setting range, but C (output) may include, for example, changing the setting range from +1° C. to +2° C. Further, the change of the setting range may be applied to all the indoor units included in the refrigeration cycle, or may be applied only to a predetermined indoor unit.

[0106] FIG. 16 is a diagram for explaining a first example of reward calculation. As shown in FIG. 16, the control device 50 calculates a reward from the execution period of the high latent heat control increased based on excluding the exclusion candidate inferred by the AI. Specifically, the control device 50 acquires data indicating the execution period of the high latent heat control increased after the exclusion setting is made by B1 (action).

[0107] Specifically, when the control device 50 determines that the execution period of the high latent heat control executed after the exclusion setting by B1 (action) is shorter than the threshold value, the control device 50 sets a small reward, assuming that the exclusion setting should not have been performed. On the other hand, when the control device 50 determines that the execution period of the high latent heat control executed after the exclusion setting by B1 (action) is longer than the threshold value, the control device 50 sets a large reward, assuming that the exclusion setting should have been performed.

[0108] FIG. 17 is a diagram for explaining a second example of reward calculation. As shown in FIG. 17, the control device 50 calculates a reward based on the indoor comfort after the execution of the exclusion setting. Specifically, if the number of times the set temperature is changed by the user after the exclusion setting by B1 (action) is less than the threshold value, the control device 50 sets a large reward on the assumption that the corresponding indoor comfort has not been reduced by the execution of the high sensible heat control. On the other hand, if the number of times the set temperature is changed by the user after the exclusion setting by B1 (action) is more than the threshold value, the control device 50 sets a small reward on the assumption that the comfort has been reduced by the execution of the high sensible heat control. Alternatively, if the period during which the set temperature is not changed by the user after the exclusion setting by B1 (action) is longer than the threshold value, the control device 50 sets a large reward on the assumption that the comfort has not been reduced even when the high sensible heat control is executed. On the other hand, if the period during which the set temperature is not changed by the user after the exclusion setting by B1 (action) is shorter than the threshold value, the control device 50 sets a small reward on the assumption that the comfort has been reduced by the execution of the high sensible heat control. Also, as described above, the reduction in comfort may be estimated using the captured image of the user.

[0109] Note that the reward calculation shown in FIGS. 16 and 17 is an example, and the air conditioning system 100 may calculate the reward by other methods. For example, the air conditioning system 100 further includes a humidity sensor, and when the detected value of the humidity sensor is within a predetermined range after the exclusion setting by B1 (action), a large reward is set, and when the detected value of the humidity sensor is not within the predetermined range, a small reward may be set.

[0110] <Learning phase> FIG. 18 is a diagram showing the configuration of the learning device 301. The learning device 301 is realized by the arithmetic unit 54 of the control device 50. The learning device 301 can exchange data with each of the learning program storage unit 302 and the learned model storage unit 303. Each of the learning program storage unit 302 and the learned model storage unit 303 is realized by the storage unit 52 of the control device 50.

[0111] As shown in FIG. 18, the learning device 301 includes a data acquisition unit 310 and a model generation unit 320. By executing the learning program 160 stored in the learning program storage unit 302, the learning device 301 generates a learned model 170 based on the learning data 180 including B1 (action) and B2 (state), and D (reward criterion).

[0112] The data acquisition unit 310 acquires learning data 180 including B1 (action) and B2 (state). Specifically, the data acquisition unit 310 acquires an exclusion candidate as control data for performing exclusion setting as B1 (action). The data acquisition unit 310 acquires data indicating the increasing period of the high sensible heat control and data indicating the indoor comfort as B2 (state). Specifically, the data acquisition unit 310 acquires, as the state data of B2 (state), data indicating the increasing period of the high sensible heat control based on the indoor air temperature shown in FIG. 7, the change history data of the set temperature, and the captured image.

[0113] The model generation unit 320 generates a learned model 170 that infers C (output) from B2 (state) using the learning data 180 including B1 (action) and B2 (state) acquired by the data acquisition unit 310. For example, based on the data indicating the increasing period of the high sensible heat control and the data indicating the indoor comfort, the model generation unit 320 identifies an exclusion candidate that increases the execution period of the high sensible heat control and does not reduce the comfort by exclusion. In other words, the learning device 301 learns to select an appropriate exclusion candidate based on the data indicating the increasing period of the high sensible heat control and the data indicating the indoor comfort.

[0114] For example, when a predetermined indoor unit is excluded, if the execution period of the high latent heat control increases but the indoor comfort level decreases, the learning device 301 does not select it as an exclusion candidate. Alternatively, as shown in FIG. 10, even if the indoor comfort level does not decrease when a predetermined indoor unit is excluded, if the increase in the execution period of the high latent heat control is small, the learning device 301 does not select it as an exclusion candidate. Also, as shown in FIG. 6, when the indoor temperature of a predetermined indoor unit rises sporadically, the learning device 301 excludes the predetermined indoor unit. "Sporadically" means, for example, when the indoor air temperature rises and the number of times the established conditions are not satisfied by reaching a temperature equal to or higher than a predetermined temperature (30°C in the example of FIG. 6) within a certain period is more than a predetermined number of times. When the indoor temperature rises sporadically, since the sensor 35 is provided near the door, the learning device 301 determines it as an exclusion candidate.

[0115] That is, the learning device 301 balances energy savings by high latent heat control and indoor comfort by selecting an exclusion candidate that ensures both the increase in the execution period of the high latent heat control and indoor comfort. The model generation unit 320 generates the learned model 170. The model generation unit 320 stores the generated learned model 170 in the learned model storage unit 303.

[0116] The model generation unit 320 includes a reward calculation unit 321 and a function update unit 322. The reward calculation unit 321 calculates a reward based on D (reward criterion). Specifically, the reward calculation unit 321 calculates a reward based on the change history data of the set temperature, the captured image, and the increased execution period of the high latent heat control after the exclusion setting as D (reward criterion). That is, after the control data for performing the exclusion setting is output as B1 (action), the reward calculation unit 321 determines whether a sufficient effect has been obtained against the decrease in comfort due to the execution of the high latent heat control, and calculates a reward based on the determination result.

[0117] The function update unit 322 updates a function for determining the drying operation time based on the reward calculated by the reward calculation unit 321. For example, in the case of Q-learning, the function update unit 322 uses the action value function Q(St, at) represented by the above-described equation (1) as a function for calculating exclusion candidates. The learned model storage unit 303 stores the action value function Q(St, at) updated by the function update unit 322 as the learned model 170.

[0118] FIG. 19 is a flowchart relating to the processing executed by the learning device according to the present embodiment in the learning phase. The processing shown in FIG. 19 is executed by the processor of the control device 50 corresponding to the learning device. In FIG. 19, “S” is used as an abbreviation for “STEP”.

[0119] As shown in FIG. 19, the learning device acquires learning data 180 including B1 (action) and B2 (state) by the data acquisition unit 310 (S21). The learning device 301 calculates a reward based on the learning data 180 (S22). Specifically, the learning device 301 determines exclusion candidates based on the learning data 180, and if it can output both or either one of the result that the indoor comfort shown in FIG. 16 does not decrease and the result that the execution period of the increased high-sensible heat control shown in FIG. 17 is long as the determination result, the reward is increased (S23). The learning device 301 decreases the reward when it outputs both the result that the indoor comfort shown in FIG. 16 has decreased and the result that the execution period of the increased high-sensible heat control shown in FIG. 17 is short (S24). The learning device 301 updates the action value function Q(st, at) represented by equation (1) stored in the learned model storage unit 303 based on the calculated reward (S25).

[0120] The learning device 301 repeatedly executes the above-described processes of S21 to S25, stores the generated action value function Q(st, at) in the learned model storage unit 303 as the learned model 170 (S26), and ends this process.

[0121] <Utilization Phase> FIG. 20 is a diagram showing the configuration of the inference device 401. The inference device 401 is realized by the arithmetic unit 54 of the control device 50. The inference device 401 can transfer data with the learned model storage unit 303.

[0122] As shown in FIG. 20, the inference device 401 includes a data acquisition unit 410 and an inference unit 420. The inference device 401 uses the learned model 170 to obtain C (output) from B2 (state).

[0123] The data acquisition unit 410 acquires B2 (state). Specifically, the data acquisition unit 410 acquires, as B2 (state), state data indicating the indoor comfort after the exclusion setting and the execution period of the high sensible heat control increased after the exclusion setting. More specifically, the data acquisition unit 410 acquires either data indicating the increased period of the high sensible heat control or data indicating the indoor comfort as the state data of B2 (state).

[0124] The inference unit 420 uses the learned model 170 to obtain C (output) from B2 (state). Specifically, the inference unit 420 reads out the learned model 170 from the learned model storage unit 303. The inference unit 420 uses the learned model 170 to infer C (output) based on the data acquired by the data acquisition unit 410. That is, the inference unit 420 uses the learned model 170 to infer an exclusion candidate with the most increased execution period of the high sensible heat control and no decrease in comfort based on at least one of the increased period of the high sensible heat control, the change history data of the set temperature, and the user's captured image, and outputs the inferred exclusion candidate. Thereby, when performing the exclusion setting, the inference device 401 can infer a more optimal exclusion candidate according to the increased execution period of the high sensible heat control and the indoor comfort.

[0125] It should be considered that all the embodiments disclosed this time are illustrative in all respects and not restrictive. The scope of the present invention is shown not by the above description but by the claims, and it is intended that all modifications within the meaning and scope equivalent to the claims are included.

Explanation of Reference Numerals

[0126] 10 Compressor, 11, 12 Refrigeration cycle, 15 Four-way valve, 20, 21 Outdoor unit, 25A to 25H Expansion valve, 30A to 30H, 31A to 31H Indoor unit, 32A to 32H, 33, 33A to 33H Input device, 35A to 35H, 36A to 36H Sensor, 50, 80 Control device, 51 Aggregation unit, 52 Memory unit, 53 Setting unit, 54 Calculation unit, 55 Display unit, 56 Judgment unit, 57 Output unit, 60 External terminal, 100 Air conditioning system, 122, 170 Learned model, 160 Learning program, 180 Learning data, 301 Learning device, 302 Learning program storage unit, 303 Learned model storage unit, 310, 410 Data acquisition unit, 320 Model generation unit, 321 Reward calculation unit, 322 Function update unit, 401 Inference device, 420 Inference unit, BP1, BP2 Branch point, Bt1, Bt2 Button, D Direction, F20, FA, FH Fan, L1 Circulation flow path, LnA to LnH Line, W1 Setting screen.

Claims

1. An air conditioning system, comprising: a compressor; an outdoor unit; a first indoor unit; a second indoor unit; a refrigeration cycle that connects the compressor, the outdoor unit, the first indoor unit, and the second indoor unit and circulates a refrigerant; a first temperature sensor that detects a first indoor air temperature corresponding to the first indoor unit; a second temperature sensor that detects a second indoor air temperature corresponding to the second indoor unit; a control device that controls an operating frequency of the compressor using a predetermined condition based on an indoor air temperature corresponding to the indoor unit, wherein at least one of the first indoor unit and the second indoor unit is defined as a target of the predetermined condition, the control device: when all the indoor units defined as the target satisfy the predetermined condition, controls the operating frequency of the compressor so that an evaporation temperature of the refrigeration cycle becomes a first evaporation temperature; when at least one of the indoor units defined as the target does not satisfy the predetermined condition, controls the operating frequency of the compressor so that an evaporation temperature of the refrigeration cycle becomes a second evaporation temperature lower than the first evaporation temperature; a data acquisition unit that acquires, as state data, data indicating a period during which the predetermined condition that increases when the indoor unit defined as the target is excluded from the target is satisfied; an inference unit that outputs an indoor unit to be excluded from the target based on the state data acquired by the data acquisition unit using a learned model for determining an indoor unit to be excluded from the target. An air conditioning system.

2. The air conditioning system according to claim 1, wherein the control device excludes the indoor unit to be excluded from the target output by the inference unit when receiving an instruction from a user to enable automatic control.

3. The air conditioning system according to claim 1 or claim 2, wherein the state data includes, in addition to data indicating a period during which the predetermined condition that increases when the indoor unit defined as the target is excluded from the target is satisfied, data indicating comfort of the room corresponding to the first indoor unit and the second indoor unit.

4. The air conditioning system according to claim 3, wherein the data indicating the comfort of the room includes either change history data of a set temperature of the indoor unit or a photographed image of the room.

5. The control device further includes a model generation unit that generates the learned model. The model generation unit includes a reward calculation unit that calculates a reward based on the comfort level in the room after excluding the indoor unit defined as the target from the target, and a function update unit that updates a function for inferring the indoor unit to be excluded from the target based on the reward. The air conditioning system according to claim 1 or claim 2.

6. The reward calculation unit calculates a reward based on, in addition to the comfort level in the room after excluding the indoor unit defined as the target from the target, the period during which the predetermined condition that increases after excluding the indoor unit defined as the target from the target is satisfied. The air conditioning system according to claim 5.

7. When the air temperature in the room rises sporadically, the inference unit outputs the indoor unit corresponding to the room as an exclusion candidate. The air conditioning system according to claim 1 or claim 2.

8. The predetermined condition is a condition as to whether or not the temperature difference between the set temperature of the indoor unit and the air temperature in the room corresponding to the indoor unit is within a predetermined range. The air conditioning system according to claim 1 or claim 2.

9. The inference unit outputs a range after changing the predetermined range based on the state data acquired by the data acquisition unit using a learned model for changing the predetermined range. The air conditioning system according to claim 8.

Images