Heat load prediction system

Abstract

To solve the problem that it is difficult to predict a heat load for each air conditioning zone when a heat load in a building is learned by one outdoor unit.SOLUTION: A heat load prediction system 1 includes an environment data acquisition unit 191, an operation data acquisition unit 192, a storage unit 11, a learning unit 193, and a prediction unit 194. The environment data acquisition unit 191 acquires external environmental data D1 and internal environment data D2 on an air conditioning zone. The operation data acquisition unit 192 acquires operation data D3 on an air conditioner that executes air conditioning of the air conditioning zone. The storage unit 11 stores the data acquired by the environment data acquisition unit 191 and the operation data acquisition unit 192. The learning unit 193 executes learning of models M1 and M2 for predicting a heat load of the air conditioning zone using data LD1 and LD2 for learning acquired from the storage unit 11. The prediction unit 194 predicts the heat load of the air conditioning zone using the models M1 and M2.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] This relates to a heat load prediction system. [Background technology]

[0002] As shown in Patent Document 1 (Japanese Patent No. 5572799), there is a technology that learns and predicts the heat load within a building. Summary of the Invention [Problem to be solved by the invention]

[0003] When learning the heat load for each outdoor unit as in Patent Document 1, there is a problem in that it is difficult to predict the heat load for each air conditioning zone. [Means for solving the problem]

[0004] A heat load prediction system according to a first aspect predicts a heat load of an air conditioning zone in a building. The heat load prediction system includes an environmental data acquisition unit, an operation data acquisition unit, a storage unit, a learning unit, and a prediction unit. The environmental data acquisition unit acquires external environmental data related to an environment outside the building and / or internal environmental data related to an environment inside the building, which affect the heat load of the air conditioning zone. The operation data acquisition unit acquires operation data of an air conditioner that performs air conditioning of the air conditioning zone. The storage unit stores the data acquired by the environmental data acquisition unit and the operation data acquisition unit. The learning unit uses learning data acquired from the storage unit to learn a model that predicts the heat load of the air conditioning zone. The prediction unit predicts the heat load of the air conditioning zone using the model. Indoor units that constitute the air conditioner are installed in the air conditioning zone. The heat load of the air conditioning zone in the learning data is calculated based on the heat exchange amount in the indoor units installed in the air conditioning zone. The heat exchange amount in the indoor units is calculated based on the operation data.

[0005] In the heat load prediction system of the first aspect, the learning unit uses learning data obtained from the storage unit to learn a model that predicts the heat load of an air conditioning zone. The prediction unit predicts the heat load of an air conditioning zone using the model. The heat load of an air conditioning zone in the learning data is calculated based on the heat exchange amount in an indoor unit installed in the air conditioning zone. As a result, the heat load prediction system can learn and predict the heat load for each air conditioning zone.

[0006] A heat load prediction system according to a second aspect is the heat load prediction system according to the first aspect, wherein the external environment data includes at least an outside air temperature of the building, and the internal environment data includes at least an indoor temperature of an air-conditioning zone.

[0007] A heat load prediction system of a third aspect is a heat load prediction system of either the first aspect or the second aspect, in which the learning unit excludes records including missing values ​​or abnormal values ​​from the learning data before learning the model.

[0008] A heat load prediction system according to a fourth aspect is the heat load prediction system according to either the first or second aspect, in which the learning unit complements missing values ​​or abnormal values ​​contained in the learning data before learning the model.

[0009] A thermal load prediction system of a fifth aspect is a thermal load prediction system of any of the first aspect to the fourth aspect, in which the learning unit processes learning data based on the operating state of the air conditioner or the schedule of the air conditioner before learning the model.

[0010] The heat load prediction system of the fifth aspect can improve the accuracy of the model predicting the heat load of an air conditioning zone by excluding data that is weakly related to the heat load of an air conditioning zone based on the operating status of the air conditioner or the air conditioner schedule.

[0011] A heat load prediction system of a sixth aspect is a heat load prediction system of any of the first aspect to the fifth aspect, wherein the learning unit uses second learning data for a second air conditioning zone similar to the first air conditioning zone instead of the first learning data for the first air conditioning zone.

[0012] With such a configuration, the heat load prediction system according to the sixth aspect can share learning data between similar air-conditioning zones.

[0013] A heat load prediction system according to a seventh aspect is the heat load prediction system according to any one of the first aspect to the sixth aspect, in which the model is a machine learning model, a statistical model, a physical model, or a combination thereof.

[0014] A heat load prediction system according to an eighth aspect is the heat load prediction system according to any one of the first aspect to the seventh aspect, wherein the learning unit learns the model using learning data of a predetermined unit.

[0015] A heat load prediction system according to a ninth aspect is the heat load prediction system according to any one of the first aspect to the eighth aspect, wherein the learning unit updates the model at a predetermined timing.

[0016] A heat load prediction system of a tenth aspect is a heat load prediction system of any of the first to ninth aspects, wherein the heat exchange amount in the indoor unit is calculated based on the suction temperature, air volume, and refrigerant temperature in the indoor unit.

[0017] A heat load prediction system according to an eleventh aspect is the heat load prediction system according to any one of the first to ninth aspects, wherein the heat exchange amount in the indoor unit is calculated based on the intake temperature, the discharge temperature, and the air volume in the indoor unit. [Brief description of the drawings]

[0018] [Figure 1] FIG. 1 is a schematic configuration diagram of a heat load prediction system. [Diagram 2] FIG. 2 is a diagram showing a refrigerant circuit of the air conditioner. [Diagram 3] FIG. 2 is a control block diagram of the air conditioner. [Figure 4] FIG. 2 is a control block diagram of the heat load prediction device. [Diagram 5] 1 is a flowchart for explaining the processing of the heat load prediction system. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] (1) Overall structure The heat load prediction system 1 is a system that predicts the heat load of air conditioning zones Z1 and Z2 in a building BL. FIG. 1 is a schematic configuration diagram of the heat load prediction system 1. As shown in FIG. 1, the heat load prediction system 1 has an air conditioner 2 and a heat load prediction device 3. The air conditioner 2 and the heat load prediction device 3 are communicatively connected via a network NW. The network NW is, for example, the Internet.

[0020] The air conditioner 2 configures a vapor compression refrigeration cycle and performs air conditioning of air conditioning zones Z1 and Z2 in a building BL. In this embodiment, the air conditioner 2 is a so-called multi-type air conditioning system for buildings. As shown in FIG. 1, the air conditioner 2 has indoor units 20a to 20c and an outdoor unit 30. The indoor units 20a and 20b are installed in the air conditioning zone Z1. The indoor unit 20c is installed in the air conditioning zone Z2. The indoor units 20a to 20c and the outdoor unit 30 are connected to each other so as to be able to communicate with each other via a communication line 80. FIG. 2 is a diagram showing a refrigerant circuit 50 of the air conditioner 2. As shown in FIG. 2, the indoor units 20a to 20c and the outdoor unit 30 are connected by a liquid refrigerant connection pipe 51 and a gas refrigerant connection pipe 52 to configure the refrigerant circuit 50.

[0021] The heat load prediction device 3 predicts the heat loads of the air conditioning zones Z1 and Z2 based on various data acquired from the air conditioner 2 and the like.

[0022] 1, indoor temperature sensors 631, 632 that measure the indoor temperatures of the air-conditioning zones Z1, Z2 are installed in the air-conditioning zones Z1, Z2. In addition, an outdoor temperature sensor 65 that measures the outdoor air temperature of the building BL is installed outside the building BL. The heat load prediction device 3, the indoor temperature sensors 631, 632, and the outdoor temperature sensor 65 are communicatively connected via a network NW.

[0023] (2) Detailed configuration (2-1) Indoor unit In this embodiment, the indoor units 20a to 20c are ceiling-embedded units that are installed on the ceiling. As shown in Fig. 2, the indoor units 20a to 20c mainly include indoor heat exchangers 21a to 21c, indoor fans 22a to 22c, indoor expansion valves 23a to 23c, indoor control units 29a to 29c, indoor suction temperature sensors 61a to 61c, indoor heat exchanger temperature sensors 62a to 62c, gas-side temperature sensors 64a to 64c, and liquid-side temperature sensors 67a to 67c. Also, as shown in Fig. 2, the indoor units 20a to 20c include liquid refrigerant pipes 53a1 to 53c1 that connect the liquid side ends of the indoor heat exchangers 21a to 21c to the liquid refrigerant connection pipe 51. Furthermore, the indoor units 20a to 20c have gas refrigerant pipes 53a2 to 53c2 that connect the gas side ends of the indoor heat exchangers 21a to 21c to the gas refrigerant communication pipe 52.

[0024] (2-1-1) Indoor heat exchanger The indoor heat exchangers 21a to 21c exchange heat between the refrigerant flowing therethrough and the air in the air-conditioning zones Z1 and Z2. The indoor heat exchangers 21a to 21c are, for example, fin-and-tube type heat exchangers having a plurality of heat transfer fins and a plurality of heat transfer tubes. As shown in FIG. 2, one end of the indoor heat exchangers 21a to 21c is connected to the liquid refrigerant connection pipe 51 via the liquid refrigerant pipes 53a1 to 53c1. The other end of the indoor heat exchangers 21a to 21c is connected to the gas refrigerant connection pipe 52 via the gas refrigerant pipes 53a2 to 53c2. During cooling operation, the refrigerant flows into the indoor heat exchangers 21a to 21c from the liquid refrigerant pipes 53a1 to 53c1, and the indoor heat exchangers 21a to 21c function as refrigerant evaporators. During heating operation, refrigerant flows into the indoor heat exchangers 21a-21c from the gas refrigerant pipes 53a2-53c2, and the indoor heat exchangers 21a-21c function as refrigerant condensers. In this embodiment, the indoor heat exchangers 21a-21c are fin-and-tube heat exchangers, but are not limited to this and may be other types of heat exchangers.

[0025] (2-1-2) Indoor fan In this embodiment, the indoor units 20a to 20c have indoor fans 22a to 22c as blowers for drawing indoor air into the unit, exchanging heat with the refrigerant in the indoor heat exchangers 21a to 21c, and then supplying the air to the room as supply air. The indoor fans 22a to 22c are fans that supply air from the air conditioning zones Z1 and Z2 to the indoor heat exchangers 21a to 21c. The indoor fans 22a to 22c are centrifugal fans such as turbo fans and sirocco fans. As shown in FIG. 2, the indoor fans 22a to 22c are driven by indoor fan motors 22am to 22cm such as DC fan motors. The rotation speed of the indoor fan motors 22am to 22cm can be controlled by an inverter.

[0026] (2-1-3) Indoor expansion valve The indoor expansion valves 23a-23c are mechanisms for adjusting the pressure and flow rate of the refrigerant flowing through the liquid refrigerant pipes 53a1-53c1. The indoor expansion valves 23a-23c are provided in the liquid refrigerant pipes 53a1-53c1. In this embodiment, the indoor expansion valves 23a-23c are electronic expansion valves with adjustable opening degrees. Note that in this embodiment, the indoor expansion valves 23a-23c are provided in the indoor units 20a-20c, respectively, as expansion mechanisms, but the present invention is not limited to this. The expansion mechanisms (including the expansion valves) may be provided in the outdoor unit 30, or may be provided in a connection unit independent of the indoor units 20a-20c and the outdoor unit 30.

[0027] (2-1-4) Sensor The indoor intake temperature sensors 61a to 61c measure the temperature of air (indoor intake temperature) sucked into the indoor units 20a to 20c. The indoor intake temperature sensors 61a to 61c are provided near the air intake ports of the indoor units 20a to 20c.

[0028] The indoor heat exchanger temperature sensors 62a-62c measure the temperature (indoor saturation temperature) of the refrigerant flowing through the indoor heat exchangers 21a-21c. The indoor saturation temperature during cooling operation is the evaporation temperature (indoor evaporation temperature) of the refrigerant flowing through the indoor heat exchangers 21a-21c. The indoor saturation temperature during heating operation is the condensation temperature (indoor condensation temperature) of the refrigerant flowing through the indoor heat exchangers 21a-21c. The indoor heat exchanger temperature sensors 62a-62c are provided in the indoor heat exchangers 21a-21c.

[0029] The gas side temperature sensors 64a to 64c measure the temperature (indoor gas side temperature) of the refrigerant flowing through the gas refrigerant pipes 53a2 to 53c2. The gas side temperature sensors 64a to 64c are provided in the gas refrigerant pipes 53a2 to 53c2.

[0030] The liquid side temperature sensors 67a to 67c measure the temperature (indoor liquid side temperature) of the refrigerant flowing through the liquid refrigerant pipes 53a1 to 53c1. The liquid side temperature sensors 67a to 67c are provided in the liquid refrigerant pipes 53a1 to 53c1.

[0031] (2-1-5) Indoor control unit The indoor control sections 29a to 29c control the operation of each section constituting the indoor units 20a to 20c.

[0032] The indoor control units 29a-29c are electrically connected to various devices included in the indoor units 20a-20c, including the indoor expansion valves 23a-23c and the indoor fan motors 22am-22cm. The indoor control units 29a-29c are also connected to be able to communicate with various sensors provided in the indoor units 20a-20c, including the indoor suction temperature sensors 61a-61c, the indoor heat exchanger temperature sensors 62a-62c, the gas-side temperature sensors 64a-64c, and the liquid-side temperature sensors 67a-67c.

[0033] The indoor control units 29a to 29c each have a control arithmetic device and a storage device. The control arithmetic device is a processor such as a CPU or a GPU. The storage device is a storage medium such as a RAM, a ROM, or a flash memory. The control arithmetic device reads out a program stored in the storage device and performs predetermined arithmetic processing in accordance with the program, thereby controlling the operation of each part constituting the indoor units 20a to 20c. The control arithmetic device can also write arithmetic results to the storage device and read out information stored in the storage device in accordance with the program. The indoor control units 29a to 29c also have a timer.

[0034] The indoor control units 29a to 29c are configured to be able to receive various signals transmitted from an operating remote control (not shown). The various signals include, for example, signals instructing the start and stop of operation, and signals related to various settings. The signals related to various settings include, for example, signals related to the set temperature and air volume. The indoor control units 29a to 29c also exchange control signals, measurement signals, signals related to various settings, and the like with the outdoor control unit 39 of the outdoor unit 30 via a communication line 80. The indoor control units 29a to 29c and the outdoor control unit 39 work together to function as a controller 40. The functions of the controller 40 will be described later.

[0035] (2-2) Outdoor unit The outdoor unit 30 is installed outside the building BL, such as on the roof of the building BL. As shown in Fig. 2, the outdoor unit 30 mainly includes a compressor 31, a flow path switching valve 32, an outdoor heat exchanger 33, an outdoor expansion valve 34, an accumulator 35, an outdoor fan 36, a liquid side shutoff valve 37, a gas side shutoff valve 38, an outdoor control unit 39, an outdoor heat exchanger temperature sensor 66, a suction pressure sensor 68, and a discharge pressure sensor 69. The outdoor unit 30 also includes a suction pipe 54a, a discharge pipe 54b, gas refrigerant pipes 54c and 54e, and a liquid refrigerant pipe 54d.

[0036] The suction pipe 54a connects the flow path switching valve 32 and the suction side of the compressor 31. The accumulator 35 is provided in the suction pipe 54a. The discharge pipe 54b connects the discharge side of the compressor 31 and the flow path switching valve 32. The gas refrigerant piping 54c connects the flow path switching valve 32 and the gas side of the outdoor heat exchanger 33. The liquid refrigerant piping 54d connects the liquid side of the outdoor heat exchanger 33 and the liquid refrigerant connection piping 51. The liquid refrigerant piping 54d is provided with the outdoor expansion valve 34. The liquid side stop valve 37 is provided at the connection between the liquid refrigerant piping 54d and the liquid refrigerant connection piping 51. The gas refrigerant piping 54e connects the flow path switching valve 32 and the gas refrigerant connection piping 52. The gas side stop valve 38 is provided at the connection between the gas refrigerant piping 54e and the gas refrigerant connection piping 52. The liquid side shutoff valve 37 and the gas side shutoff valve 38 are valves that are manually opened and closed.

[0037] (2-2-1) Compressor As shown in FIG. 2, the compressor 31 draws in low-pressure refrigerant through a suction pipe 54a, compresses the refrigerant by a compression mechanism (not shown), and discharges the compressed refrigerant to a discharge pipe 54b.

[0038] The compressor 31 is, for example, a volumetric compressor of a rotary type, a scroll type, etc. The compression mechanism of the compressor 31 is driven by a compressor motor 31m. The rotation speed of the compressor motor 31m can be controlled by an inverter.

[0039] (2-2-2) Flow path switching valve The flow path switching valve 32 is a mechanism that switches the refrigerant flow path between a first state and a second state. In the first state, the flow path switching valve 32 connects the suction pipe 54a to the gas refrigerant pipe 54e and the discharge pipe 54b to the gas refrigerant pipe 54c, as shown by the solid lines in the flow path switching valve 32 in Fig. 2. In the second state, the flow path switching valve 32 connects the suction pipe 54a to the gas refrigerant pipe 54c and the discharge pipe 54b to the gas refrigerant pipe 53e, as shown by the dashed lines in the flow path switching valve 32 in Fig. 2.

[0040] During cooling operation, the flow path switching valve 32 sets the refrigerant flow path to the first state. At this time, the refrigerant discharged from the compressor 31 flows through the refrigerant circuit 50 in the order of the outdoor heat exchanger 33, the outdoor expansion valve 34, the indoor expansion valves 23a to 23c, and the indoor heat exchangers 21a to 21c, and returns to the compressor 31. In the first state, the outdoor heat exchanger 33 functions as a condenser, and the indoor heat exchangers 21a to 21c function as evaporators.

[0041] During heating operation, the flow path switching valve 32 sets the refrigerant flow path to the second state. At this time, the refrigerant discharged from the compressor 31 flows through the refrigerant circuit 50 in the order of the indoor heat exchangers 21a to 21c, the indoor expansion valves 23a to 23c, the outdoor expansion valve 34, and the outdoor heat exchanger 33, and then returns to the compressor 31. In the second state, the outdoor heat exchanger 33 functions as an evaporator, and the indoor heat exchangers 21a to 21c function as condensers.

[0042] (2-2-3) Outdoor heat exchanger The outdoor heat exchanger 33 exchanges heat between the refrigerant flowing through the outdoor heat exchanger 33 and the air outside the building BL. The outdoor heat exchanger 33 is, for example, a fin-and-tube type heat exchanger having a plurality of heat transfer fins and a plurality of heat transfer tubes. As shown in Fig. 2, one end of the outdoor heat exchanger 33 is connected to the liquid refrigerant connection pipe 51 via a liquid refrigerant pipe 54d. The other end of the outdoor heat exchanger 33 is connected to the flow path switching valve 32 via a gas refrigerant pipe 54c.

[0043] During cooling operation, refrigerant flows into the outdoor heat exchanger 33 from the gas refrigerant pipe 54c, and the outdoor heat exchanger 33 functions as a refrigerant condenser. During heating operation, refrigerant flows into the outdoor heat exchanger 33 from the liquid refrigerant pipe 54d, and the outdoor heat exchanger 33 functions as an evaporator of the refrigerant. In this embodiment, the outdoor heat exchanger 33 is a fin-and-tube type heat exchanger, but is not limited to this and may be another type of heat exchanger.

[0044] (2-2-4) Outdoor expansion valve The outdoor expansion valve 34 is a mechanism for adjusting the pressure and flow rate of the refrigerant flowing through the liquid refrigerant pipe 54d. As shown in Fig. 2, the outdoor expansion valve 34 is provided in the liquid refrigerant pipe 54d. In this embodiment, the outdoor expansion valve 34 is an electronic expansion valve whose opening degree can be adjusted.

[0045] (2-2-5) Accumulator The accumulator 35 is a container having a gas-liquid separation function that separates the refrigerant flowing in into a gas refrigerant and a liquid refrigerant. As shown in Fig. 2, the accumulator 35 is provided in the suction pipe 54a. The refrigerant flowing in the accumulator 35 is separated into a gas refrigerant and a liquid refrigerant, and the gas refrigerant that collects in the upper space flows into the compressor 31.

[0046] (2-2-6) Outdoor fan The outdoor fan 36 is a fan that supplies air outside the building BL to the outdoor heat exchanger 33. The outdoor fan 36 is, for example, an axial flow fan such as a propeller fan. As shown in Fig. 2, the outdoor fan 36 is driven by an outdoor fan motor 36m formed of a DC fan motor or the like. The rotation speed of the outdoor fan motor 36m can be controlled by an inverter.

[0047] (2-2-7) Sensors The outdoor heat exchanger temperature sensor 66 measures the temperature of the refrigerant flowing through the outdoor heat exchanger 33. The outdoor heat exchanger temperature sensor 66 is provided in the outdoor heat exchanger 33.

[0048] The suction pressure sensor 68 is a sensor that measures the suction pressure of the compressor 31. The suction pressure sensor 68 is provided in the suction pipe 54a. The suction pressure is a refrigerant pressure that corresponds to the evaporation pressure during cooling operation.

[0049] The discharge pressure sensor 69 is a sensor that measures the discharge pressure of the compressor 31. The discharge pressure sensor 69 is provided on the discharge pipe 54b. The discharge pressure is a refrigerant pressure that corresponds to the condensation pressure during heating operation.

[0050] (2-2-8) Outdoor control unit The outdoor control unit 39 controls the operation of each part that constitutes the outdoor unit 30.

[0051] The outdoor control unit 39 is electrically connected to various devices of the outdoor unit 30, including the compressor motor 31m, the flow path switching valve 32, the outdoor expansion valve 34, and the outdoor fan motor 36m. In addition, the outdoor control unit 39 is connected to be able to communicate with various sensors provided in the outdoor unit 30, including the outdoor heat exchanger temperature sensor 66, the suction pressure sensor 68, and the discharge pressure sensor 69.

[0052] The outdoor control unit 39 has a control arithmetic device and a storage device. The control arithmetic device is a processor such as a CPU or a GPU. The storage device is a storage medium such as a RAM, a ROM, or a flash memory. The control arithmetic device reads out a program stored in the storage device and performs predetermined arithmetic processing in accordance with the program, thereby controlling the operation of each part constituting the outdoor unit 30. The control arithmetic device can also write the results of calculations to the storage device and read out information stored in the storage device in accordance with the program. The outdoor control unit 39 also has a timer.

[0053] The outdoor control unit 39 exchanges control signals, measurement signals, signals related to various settings, and the like with the indoor control units 29a to 29c of the indoor units 20a to 20c via a communication line 80. The outdoor control unit 39 and the indoor control units 29a to 29c work together to function as a controller 40. The functions of the controller 40 will be described later.

[0054] (2-3) Controller The controller 40 is made up of indoor control units 29a-29c and an outdoor control unit 39. The controller 40 controls the operation of the entire air conditioner 2 by causing the control and arithmetic devices of the indoor control units 29a-29c and the outdoor control unit 39 to execute programs stored in the respective storage devices. Note that the air conditioner 2 may have, as the controller 40, a centralized controller (a so-called edge) that centrally controls the indoor units 20a-20c and the outdoor unit 30.

[0055] FIG. 3 is a control block diagram of the air conditioner 2. As shown in FIG. 3, the controller 40 is communicatively connected to the indoor intake temperature sensors 61a-61c, the indoor heat exchanger temperature sensors 62a-62c, the gas side temperature sensors 64a-64c, the liquid side temperature sensors 67a-67c, the outdoor heat exchanger temperature sensor 66, the intake pressure sensor 68, and the discharge pressure sensor 69. The controller 40 is also electrically connected to the indoor expansion valves 23a-23c, the indoor fan motors 22am-22cm, the compressor motor 31m, the flow path switching valve 32, the outdoor expansion valve 34, and the outdoor fan motor 36m. The controller 40 is also communicatively connected to the heat load prediction device 3 via the network NW. The controller 40 controls the operation of various devices of the air conditioner 2 based on control signals received from the operation remote controllers via the indoor units 20a-20c, control signals received from the heat load prediction device 3, measurement signals of various sensors, and the like.

[0056] The controller 40 mainly performs cooling operation and heating operation, and mainly has a data transmission function.

[0057] (2-3-1) Cooling operation Here, a case where the controller 40 causes the indoor unit 20a to perform cooling operation will be described.

[0058] When the controller 40 receives an instruction to perform cooling operation from, for example, the operation remote control via the indoor unit 20a, the flow path switching valve 32 is switched to the first state. Then, the controller 40 adjusts the rotation speed of the compressor 31, the temperature of the refrigerant flowing through the outdoor heat exchanger 33 (temperature measured by the outdoor heat exchanger temperature sensor 66), the opening degree of the outdoor expansion valve 34, the opening degree of the indoor expansion valve 23a, and the like so that the indoor refrigerant temperature becomes the target refrigerant temperature. Here, the indoor refrigerant temperature means the indoor saturation temperature (indoor evaporation temperature during cooling operation) and the degree of superheat. In this embodiment, the indoor evaporation temperature is a measurement value of the indoor heat exchanger temperature sensor 62a. However, the indoor evaporation temperature may be calculated from the suction pressure. The degree of superheat is calculated by subtracting the indoor evaporation temperature or the indoor liquid side temperature from the indoor gas side temperature. The target refrigerant temperature is set to a temperature according to the set temperature received from the operation remote control or to a target refrigerant temperature received from the heat load prediction device 3. The controller 40 may set the target refrigerant temperature to the temperature most recently received from the operation remote control or the heat load prediction device 3, for example.

[0059] As described above, the operations of the various devices are controlled, so that the refrigerant flows through the refrigerant circuit 50 during cooling operation as follows.

[0060] When the compressor 31 is started, low-pressure gas refrigerant is sucked into the compressor 31 and compressed by the compressor 31 to become high-pressure gas refrigerant. The high-pressure gas refrigerant is sent to the outdoor heat exchanger 33 via the flow path switching valve 32, and condenses through heat exchange with the air outside the building BL supplied by the outdoor fan 36, becoming high-pressure liquid refrigerant. The high-pressure liquid refrigerant flows through the liquid refrigerant pipe 54d and passes through the outdoor expansion valve 34. The high-pressure liquid refrigerant sent to the indoor unit 20a is decompressed to near the suction pressure of the compressor 31 in the indoor expansion valve 23a, becomes a gas-liquid two-phase refrigerant, and is sent to the indoor heat exchanger 21a. In the indoor heat exchanger 21a, the gas-liquid two-phase refrigerant exchanges heat with the air in the air-conditioning zone Z1 supplied to the indoor heat exchanger 21a by the indoor fan 22a, evaporates, and becomes a low-pressure gas refrigerant. The low-pressure gas refrigerant is sent to the outdoor unit 30 via the gas refrigerant communication piping 52, and flows into the accumulator 35 via the flow path switching valve 32. The low-pressure gas refrigerant that has flowed into the accumulator 35 is again sucked into the compressor 31. The temperature of the air supplied to the indoor heat exchanger 21a is reduced by heat exchange with the refrigerant flowing through the indoor heat exchanger 21a, and the air cooled by the indoor heat exchanger 21a is blown out into the air-conditioning zone Z1.

[0061] (2-3-2) Heating operation Here, a case where the controller 40 causes the indoor unit 20a to perform heating operation will be described.

[0062] For example, when the controller 40 receives an instruction to perform heating operation from the operation remote control via the indoor unit 20a, the controller 40 switches the flow path switching valve 32 to the second state. Then, the controller 40 adjusts the rotation speed of the compressor 31, the temperature of the refrigerant flowing through the outdoor heat exchanger 33, the opening degree of the outdoor expansion valve 34, the opening degree of the indoor expansion valve 23a, and the like so that the indoor refrigerant temperature becomes the target refrigerant temperature. Here, the indoor refrigerant temperature means the indoor condensing temperature (the indoor condensing temperature during heating operation) and the degree of subcooling. In this embodiment, the indoor condensing temperature is a measurement value of the indoor heat exchanger temperature sensor 62a. However, the indoor condensing temperature may be calculated from the discharge pressure. The degree of subcooling is calculated by subtracting the indoor liquid side temperature from the indoor condensing temperature or the indoor gas side temperature. The target refrigerant temperature is set to a temperature according to the set temperature received from the operation remote control or to a target refrigerant temperature received from the heat load prediction device 3. The controller 40 may set the target refrigerant temperature to the temperature most recently received from the operation remote control or the heat load prediction device 3, for example.

[0063] As described above, the operations of the various devices are controlled, so that the refrigerant flows through the refrigerant circuit 50 during heating operation as follows.

[0064] When the compressor 31 is started, low-pressure gas refrigerant is sucked into the compressor 31 and compressed by the compressor 31 to become high-pressure gas refrigerant. The high-pressure gas refrigerant is sent to the indoor heat exchanger 21a via the flow path switching valve 32, and condenses by heat exchange with the air in the air-conditioning zone Z1 supplied to the indoor heat exchanger 21a by the indoor fan 22a, becoming high-pressure liquid refrigerant. The temperature of the air supplied to the indoor heat exchanger 21a increases by heat exchange with the refrigerant flowing through the indoor heat exchanger 21a, and the air heated by the indoor heat exchanger 21a is blown out to the air-conditioning zone Z1. The high-pressure liquid refrigerant that has passed through the indoor heat exchanger 21a is depressurized by the indoor expansion valve 23a. The depressurized liquid refrigerant is sent to the outdoor unit 30 via the liquid refrigerant connection pipe 51 and flows into the liquid refrigerant pipe 54d. The refrigerant flowing through the liquid refrigerant pipe 54d is decompressed in the outdoor expansion valve 34 to nearly the suction pressure of the compressor 31, becomes a two-phase gas-liquid refrigerant, and flows into the outdoor heat exchanger 33. The low-pressure two-phase gas-liquid refrigerant that has flowed into the outdoor heat exchanger 33 exchanges heat with air outside the building BL supplied by the outdoor fan 36, and evaporates, becoming a low-pressure gas refrigerant. The low-pressure gas refrigerant flows into the accumulator 35 via the flow path switching valve 32. The low-pressure gas refrigerant that has flowed into the accumulator 35 is sucked into the compressor 31 again.

[0065] (2-3-3) Data transmission function The controller 40 transmits operation data D3 of the air conditioner 2 to the thermal load prediction device 3.

[0066] In this embodiment, the controller 40 transmits the indoor suction temperature, air volume, and indoor refrigerant temperature (indoor saturation temperature, and degree of superheating or subcooling) of the indoor units 20a to 20c as operation data D3 of the air conditioner 2 to the thermal load prediction device 3.

[0067] In this embodiment, the controller 40 obtains the indoor suction temperatures of the indoor units 20a to 20c from the indoor suction temperature sensors 61a to 61c. The controller 40 also obtains the air volumes of the indoor units 20a to 20c from the rotation speeds of the indoor fan motors 22am to 22cm. The controller 40 also obtains the indoor saturation temperatures of the indoor units 20a to 20c from the indoor heat exchanger temperature sensors 62a to 62c. The indoor saturation temperatures may be calculated from the suction pressure or the discharge pressure. The controller 40 also calculates the degree of superheat by subtracting the indoor evaporating temperature or the indoor liquid side temperature from the indoor gas side temperature. The controller 40 also calculates the degree of subcooling by subtracting the indoor liquid side temperature from the indoor condensing temperature or the indoor gas side temperature.

[0068] In this embodiment, the controller 40 acquires operation data D3 of the air conditioner 2 and transmits it to the thermal load prediction device 3 every 10 minutes.

[0069] (2-4) Heat load prediction device The heat load prediction device 3 is a computer installed on the cloud. Fig. 4 is a control block diagram of the heat load prediction device 3. As shown in Fig. 4, the heat load prediction device 3 mainly includes a storage unit 11, an input unit 12, a display unit 13, a communication unit 14, and a control unit 19.

[0070] (2-4-1) Storage section The storage unit 11 is a storage device such as a RAM, a ROM, and a HDD. The storage unit 11 stores programs executed by the control unit 19, data necessary for executing the programs, and the like.

[0071] In this embodiment, the storage unit 11 particularly stores external environment data D1 and internal environment data D2 acquired by an environment data acquisition unit 191, which will be described later, and driving data D3 acquired by a driving data acquisition unit 192.

[0072] (2-4-2) Input section The input unit 12 is a keyboard and a mouse. Various commands and various pieces of information for the heat load prediction device 3 can be input using the input unit 12.

[0073] (2-4-3)Display section The display unit 13 is a monitor. The display unit 13 can display various data stored in the storage unit 11.

[0074] (2-4-4) Communications Department The communication unit 14 is a network interface device for communicating with the air conditioner 2 and the like via the network NW.

[0075] (2-4-5) Control Unit The control unit 19 is a processor such as a CPU or a GPU. The control unit 19 reads and executes programs stored in the storage unit 11 to realize various functions of the heat load prediction device 3. The control unit 19 can also write calculation results to the storage unit 11 and read information stored in the storage unit 11 according to the programs.

[0076] As shown in FIG. 4, the control unit 19 has, as functional blocks, an environmental data acquisition unit 191, an operating data acquisition unit 192, a learning unit 193, a prediction unit 194, a determination unit 195, and a transmission unit 196.

[0077] (2-4-5-1) Environmental data acquisition section The environmental data acquisition unit 191 acquires external environmental data D1 related to the environment outside the building BL and internal environmental data D2 related to the environment inside the building BL, which affect the heat loads of the air-conditioning zones Z1 and Z2.

[0078] The external environment data D1 includes at least the outside air temperature of the building BL. In this embodiment, the external environment data D1 is the outside air temperature of the building BL. The environment data acquisition unit 191 acquires the outside air temperature of the building BL from the outdoor temperature sensor 65.

[0079] The internal environment data D2 includes at least the indoor temperatures of the air conditioning zones Z1 and Z2. In this embodiment, the internal environment data D2 is the indoor temperatures of the air conditioning zones Z1 and Z2. The environmental data acquisition unit 191 acquires the indoor temperatures of the air conditioning zones Z1 and Z2 from the indoor temperature sensors 631 and 632.

[0080] In this embodiment, the environmental data acquisition unit 191 acquires the external environmental data D1 and the internal environmental data D2 every 10 minutes.

[0081] The external environment data D1 and the internal environment data D2 are used as learning data LD1 and LD2, which will be described later.

[0082] (2-4-5-2) Operation data acquisition section The operation data acquisition unit 192 acquires operation data D3 of the air conditioner 2 from the air conditioner 2.

[0083] As described above, in this embodiment, the operating data D3 is the indoor suction temperature, air volume, and indoor refrigerant temperature of the indoor units 20a to 20c.

[0084] In this embodiment, the driving data acquisition unit 192 acquires the driving data D3 every 10 minutes at the same timing as when the environmental data acquisition unit 191 acquires the external environmental data D1 and the internal environmental data D2.

[0085] The driving data D3 is used as learning data LD1 and LD2, which will be described later.

[0086] (2-4-5-3) Learning Department The learning unit 193 uses the learning data LD1, LD2 obtained from the storage unit 11 to learn models M1, M2 that predict the heat loads of the air-conditioning zones Z1, Z2 for each of the air-conditioning zones Z1, Z2.

[0087] The learning data LD1, LD2 are composed of external environment data D1, internal environment data D2, and operation data D3. In this embodiment, the learning data LD1 of the model M1 is data that associates the indoor temperature of the air conditioning zone Z1, the outdoor temperature of the building BL, and the heat load of the air conditioning zone Z1. The learning data LD2 of the model M2 that predicts the heat load of the air conditioning zone Z2 is data that associates the indoor temperature of the air conditioning zone Z2, the outdoor temperature of the building BL, and the heat load of the air conditioning zone Z2. In other words, the learning unit 193 learns the models M1, M2 that predict the heat load of the air conditioning zones Z1, Z2 from the indoor temperatures of the air conditioning zones Z1, Z2 and the outdoor temperature of the building BL for each of the air conditioning zones Z1, Z2.

[0088] The learning unit 193 calculates the thermal loads of the air conditioning zones Z1, Z2 in the learning data LD1, LD2 based on the heat exchange amounts in each of the indoor units 20a-20c installed in the air conditioning zones Z1, Z2. The learning unit 193 calculates the heat exchange amounts in each of the indoor units 20a-20c based on the operating data D3.

[0089] Specifically, the learning unit 193 first calculates the heat exchange amount in each of the indoor units 20a to 20c from the indoor suction temperature, air volume, and indoor refrigerant temperature in each of the indoor units 20a to 20c using a predetermined relational expression that holds between the heat exchange amount, indoor suction temperature, air volume, and indoor refrigerant temperature in the indoor units 20a to 20c. Then, the learning unit 193 adds up the calculated heat exchange amounts of the indoor units 20a to 20c for each of the air conditioning zones Z1 and Z2 in which the indoor units 20a to 20c are installed, and sets these as the heat loads of the air conditioning zones Z1 and Z2. For example, the heat load of the air conditioning zone Z1 is the sum of the heat exchange amount of the indoor unit 20a and the heat exchange amount of the indoor unit 20b. Also, for example, the heat load of the air conditioning zone Z2 is the heat exchange amount of the indoor unit 20c.

[0090] The learning unit 193 may exclude records including missing values ​​or abnormal values ​​from the learning data LD1, LD2 as preprocessing before learning the models M1, M2. Furthermore, the learning unit 193 may complement missing values ​​or abnormal values ​​included in the learning data LD1, LD2 as preprocessing before learning the models M1, M2. Furthermore, the learning unit 193 may process the learning data LD1, LD2 based on the operating state of the air conditioner 2 or the schedule of the air conditioner 2 as preprocessing before learning the models M1, M2. The learning unit 193 may exclude, for example, data of operations (oil return operation, defrost operation, etc.) other than operations that control the indoor temperature (cooling operation, heating operation, etc.) from the learning data LD1, LD2 based on the operating state of the air conditioner 2. In addition, the learning unit 193 may exclude data, such as time periods when the air conditioner 2 is stopped or time periods when the air conditioner 2 is operating in fan mode, from the learning data LD1 and LD2 based on the schedule of the air conditioner 2.

[0091] In this embodiment, the models M1 and M2 are assumed to be deep learning models. The deep learning model is, for example, a fully connected neural network model. However, the models M1 and M2 may be other machine learning models, statistical models, physical models, or combinations of machine learning models, statistical models, and physical models. The models M1 and M2 may also be ensemble models or gray box models.

[0092] In this embodiment, the learning unit 193 learns (updates) the models M1 and M2 every week using the learning data LD1 and LD2 for the past week. The learning unit 193 may perform batch learning or mini-batch learning. Furthermore, the learning unit 193 may learn the models M1 and M2 (online learning) every time the learning data LD1 and LD2 are acquired.

[0093] (2-4-5-4) Prediction section The prediction unit 194 predicts the heat loads of the air conditioning zones Z1 and Z2 using the models M1 and M2.

[0094] In this embodiment, the prediction unit 194 predicts the heat loads of the air conditioning zones Z1 and Z2 10 minutes after the time point when the external environment data D1, the internal environment data D2, and the operation data D3 are acquired (hereinafter, may be referred to as the data acquisition time point) every time the external environment data D1, the internal environment data D2, and the operation data D3 are acquired. Specifically, the prediction unit 194 predicts the heat loads of the air conditioning zones Z1 and Z2 10 minutes after the data acquisition time point by inputting the indoor temperatures of the air conditioning zones Z1 and Z2 and the outdoor air temperature of the building BL 10 minutes after the data acquisition time point into the models M1 and M2.

[0095] The indoor temperature 10 minutes after the data acquisition time is predicted using, for example, a model trained on data correlating previously acquired indoor temperatures with the indoor temperature 10 minutes after that. Also, under the assumption that the indoor temperatures of air conditioning zones Z1 and Z2 do not change significantly over a 10-minute period, the indoor temperatures of air conditioning zones Z1 and Z2 most recently acquired may be used as the indoor temperatures of air conditioning zones Z1 and Z2 10 minutes after the data acquisition time.

[0096] The outdoor air temperature 10 minutes after the data acquisition time is predicted using, for example, a model trained on data correlating previously acquired outdoor air temperatures with the outdoor air temperature 10 minutes after that. Also, under the assumption that the outdoor air temperature of building BL does not change significantly over a 10-minute period, the outdoor air temperature of building BL most recently acquired may be used as the outdoor air temperature of building BL 10 minutes after the data acquisition time.

[0097] (2-4-5-5) Decision making section The determination unit 195 determines a control target value or a control command value for the air conditioner 2 based on the heat loads of the air conditioning zones Z1 and Z2 predicted by the prediction unit 194. In this embodiment, the control target value or the control command value is a target refrigerant temperature, which is a target value for the indoor refrigerant temperature. However, without being limited thereto, the control target value or the control command value may include the rotation speed of the compressor 31 for achieving the target refrigerant temperature.

[0098] First, the determination unit 195 calculates the heat load to be processed by each of the indoor units 20a to 20c based on the heat loads of the air conditioning zones Z1, Z2 predicted by the prediction unit 194 and information related to the indoor units 20a to 20c installed in the air conditioning zones Z1, Z2.

[0099] For example, the determination unit 195 apportions the heat load to be processed by each of the indoor units 20a, 20b based on the heat load of the air conditioning zone Z1 predicted by the prediction unit 194 and information about the indoor units 20a, 20b installed in the air conditioning zone Z1. The information about the indoor units 20a, 20b may be the positional relationship between the indoor units 20a, 20b, may be characteristic values ​​of the indoor units 20a, 20b, may be the operating status of the indoor units 20a, 20b, or may be a combination of these.

[0100] When the information regarding the indoor units 20a, 20b is the positional relationship between the indoor units 20a, 20b, the determination unit 195 assigns weights depending on, for example, whether the indoor units 20a, 20b are located in the perimeter zone or the interior zone, and apportions the predicted heat load of the air conditioning zone Z1 to the heat load processed by each of the indoor units 20a, 20b.

[0101] Furthermore, when the information on the indoor units 20a, 20b is characteristic values ​​of the indoor units 20a, 20b, the decision unit 195, for example, weights the indoor units 20a, 20b according to the magnitude of the rated capacity, and apportions the predicted heat load of the air conditioning zone Z1 to the heat load processed by each of the indoor units 20a, 20b.

[0102] Furthermore, when the information regarding the indoor units 20a, 20b is the operating status of the indoor units 20a, 20b, the decision unit 195, for example, weights the heat exchange amount of each of the indoor units 20a, 20b most recently calculated, and apportions the predicted heat load of the air conditioning zone Z1 to the heat load processed by each of the indoor units 20a, 20b.

[0103] Furthermore, when the information regarding the indoor units 20a, 20b is a combination of the operating status and characteristic values ​​of the indoor units 20a, 20b, the decision unit 195, for example, targets only indoor units whose on / off state on the operating remote control is ON, weights them according to the magnitude of the rated capacity, and apportions the predicted heat load of the air conditioning zone Z1 to the heat load to be processed by each of the indoor units 20a, 20b.

[0104] Furthermore, for example, the determination unit 195 calculates the heat load to be processed by the indoor unit 20c based on the heat load of air conditioning zone Z2 predicted by the prediction unit 194 and information about the indoor unit 20c installed in air conditioning zone Z2. For example, if the information about the indoor unit 20c is a characteristic value of the indoor unit 20c, the determination unit 195 sets the predicted heat load of air conditioning zone Z2 as the heat load to be processed by the indoor unit 20c within the range of the rated capacity of the indoor unit 20c.

[0105] Next, the determination unit 195 calculates the required refrigerant temperature in each of the indoor units 20a-20c that is required for each of the indoor units 20a-20c to process the calculated heat load. Specifically, the determination unit 195 calculates the required refrigerant temperature (indoor refrigerant temperature) in each of the indoor units 20a-20c from the calculated heat load (heat exchange amount), the indoor suction temperature 10 minutes after the data acquisition time, and the air volume, using a predetermined relational expression that holds between the heat exchange amount, indoor suction temperature, air volume, and indoor refrigerant temperature in the indoor units 20a-20c.

[0106] The indoor intake temperature 10 minutes after the data acquisition time point is predicted using a model learned from data associating a previously acquired indoor intake temperature with the indoor intake temperature 10 minutes after that. Also, under the assumption that the indoor intake temperature does not change significantly over 10 minutes, the most recently acquired indoor intake temperature may be used as the indoor intake temperature 10 minutes after the data acquisition time point. Also, when the indoor units 20a to 20c are in a steady state, under the assumption that the indoor intake temperature 10 minutes after the data acquisition time point is the most recently acquired set temperature, the indoor intake temperature 10 minutes after the data acquisition time point may be used as the indoor intake temperature 10 minutes after the data acquisition time point. At this time, the operation data acquisition unit 192 acquires the set temperatures of the indoor units 20a to 20c as the operation data D3.

[0107] The air volume 10 minutes after the data acquisition time is predicted using a model trained on data correlating previously acquired air volume with the air volume 10 minutes after that. In addition, under the assumption that the air volume setting will not be changed within 10 minutes, the most recently acquired air volume may be used as the air volume 10 minutes after the data acquisition time.

[0108] Next, the determination unit 195 determines a target refrigerant temperature for the air conditioner 2 that is required to bring the indoor refrigerant temperatures in each of the indoor units 20a to 20c closer to the required refrigerant temperature. The target refrigerant temperature for the air conditioner 2 here is a target refrigerant temperature common to the indoor units 20a to 20c that constitute the air conditioner 2. In other words, the determination unit 195 determines a target refrigerant temperature common to the indoor units that belong to the refrigerant system. In this embodiment, since the indoor units 20a to 20c belong to the same refrigerant system, the determination unit 195 determines the target refrigerant temperature for the air conditioner 2 by aggregating the required refrigerant temperatures in the indoor units 20a to 20c, for example, by maximum, minimum, or average value.

[0109] (2-4-5-6) Transmitter The transmitting unit 196 transmits control details including the target refrigerant temperature determined by the determining unit 195 to the air conditioner 2.

[0110] In this embodiment, when the air conditioner 2 receives a target refrigerant temperature from the heat load prediction device 3, it sets the received target refrigerant temperature as the target refrigerant temperature for the indoor units 20a to 20c. The air conditioner 2 maintains the set target refrigerant temperature for 10 minutes unless the set temperature or the like is changed by the operating remote control.

[0111] (3) Processing An example of the processing of the heat load prediction system 1 will be described with reference to the flowchart of FIG.

[0112] As shown in step S1, the thermal load prediction device 3 acquires the outside air temperature (external environment data D1) of the building BL, the indoor temperatures (internal environment data D2) of the air conditioning zones Z1 and Z2, and the indoor suction temperatures, air volumes, and indoor refrigerant temperatures (operating data D3) of the indoor units 20a to 20c.

[0113] After completing step S1, as shown in step S2, the heat load prediction device 3 predicts the heat load of air conditioning zones Z1 and Z2 10 minutes after the data acquisition time by inputting the indoor temperatures of air conditioning zones Z1 and Z2 and the outdoor temperature of building BL 10 minutes after the data acquisition time into models M1 and M2.

[0114] After completing step S2, as shown in step S3, the thermal load prediction device 3 calculates the thermal load to be processed by each of the indoor units 20a to 20c based on the predicted thermal loads of the air conditioning zones Z1, Z2 10 minutes from now and information on the indoor units 20a to 20c.

[0115] Upon completion of step S3, as shown in step S4, the thermal load predicting device 3 calculates the required refrigerant temperatures in the indoor units 20a to 20c that are required to process the calculated thermal loads.

[0116] Upon completion of step S4, as shown in step S5, the thermal load predicting device 3 determines the target refrigerant temperature of the air conditioner 2 that is necessary for bringing the indoor refrigerant temperatures in the indoor units 20a to 20c closer to the calculated required refrigerant temperature.

[0117] Upon completion of step S5, the thermal load predicting device 3 transmits control details including the determined target refrigerant temperature to the air conditioner 2, as shown in step S6.

[0118] After completing step S6, as shown in step S7, when the air conditioner 2 receives the target refrigerant temperature from the thermal load prediction device 3, it sets the received target refrigerant temperature as the target refrigerant temperature for the indoor units 20a to 20c. The air conditioner 2 maintains the set target refrigerant temperature for 10 minutes unless the set temperature or the like is changed by the operating remote control.

[0119] After completing step S7, as shown in step S8, the heat load prediction device 3 determines whether or not one week has passed since the previous learning (update) of the models M1 and M2. If one week has passed since the previous learning (update) of the models M1 and M2, the process proceeds to step S9. If one week has not passed since the previous learning (update) of the models M1 and M2, the process proceeds to step S10.

[0120] When proceeding from step S8 to step S9, the heat load prediction device 3 uses the learning data LD1 and LD2 to learn (update) models M1 and M2 that predict the heat load of air conditioning zones Z1 and Z2 for each air conditioning zone Z1 and Z2. In other words, the heat load prediction device 3 learns (updates) the models M1 and M2 every week.

[0121] When step S9 is completed, or when the process proceeds from step S8 to step S10, the thermal load predicting device 3 waits for 10 minutes.

[0122] After completing step S10, the process returns to step S1, and the thermal load prediction device 3 again acquires the external environment data D1, the internal environment data D2, and the operation data D3. In other words, the thermal load prediction device 3 acquires the external environment data D1, the internal environment data D2, and the operation data D3 every 10 minutes.

[0123] (4) Features (4-1) Conventionally, there is technology that learns and predicts the heat load in a building. However, there is an issue that it is difficult to predict the heat load for each air conditioning zone when learning the heat load for each outdoor unit as in the conventional method.

[0124] The heat load prediction system 1 of this embodiment predicts the heat loads of air conditioning zones Z1 and Z2 in a building BL. The heat load prediction system 1 includes an environmental data acquisition unit 191, an operation data acquisition unit 192, a storage unit 11, a learning unit 193, and a prediction unit 194. The environmental data acquisition unit 191 acquires external environmental data D1 related to the environment outside the building BL and internal environmental data D2 related to the environment inside the building BL, which affect the heat loads of the air conditioning zones Z1 and Z2. The operation data acquisition unit 192 acquires operation data D3 of the air conditioner 2 that performs air conditioning of the air conditioning zones Z1 and Z2. The storage unit 11 stores the data acquired by the environmental data acquisition unit 191 and the operation data acquisition unit 192. The learning unit 193 uses the learning data LD1 and LD2 obtained from the storage unit 11 to learn models M1 and M2 that predict the heat loads of the air conditioning zones Z1 and Z2. The prediction unit 194 predicts the heat loads of air conditioning zones Z1, Z2 using models M1, M2. Indoor units 20a, 20b and indoor unit 20c constituting the air conditioner 2 are installed in the air conditioning zones Z1, Z2, respectively. The heat loads of the air conditioning zones Z1, Z2 in the learning data LD1, LD2 are calculated based on the heat exchange amounts in the indoor units 20a-20c installed in the air conditioning zones Z1, Z2. The heat exchange amounts in the indoor units 20a-20c are calculated based on the operation data D3.

[0125] In the heat load prediction system 1, the learning unit 193 uses the learning data LD1, LD2 obtained from the storage unit 11 to learn models M1, M2 that predict the heat loads of the air conditioning zones Z1, Z2. The prediction unit 194 predicts the heat loads of the air conditioning zones Z1, Z2 using the models M1, M2. The heat loads of the air conditioning zones Z1, Z2 in the learning data LD1, LD2 are calculated based on the heat exchange amounts in the indoor units 20a-20c installed in the air conditioning zones Z1, Z2. As a result, the heat load prediction system 1 can learn and predict the heat load for each of the air conditioning zones Z1, Z2.

[0126] (4-2) In the heat load prediction system 1 of this embodiment, the external environment data D1 includes at least the outside air temperature of the building BL. The internal environment data D2 includes at least the indoor temperatures of the air conditioning zones Z1 and Z2.

[0127] (4-3) In the heat load prediction system 1 of this embodiment, the learning unit 193 removes records including missing values ​​or abnormal values ​​from the learning data LD1 and LD2 before learning the models M1 and M2.

[0128] (4-4) In the heat load prediction system 1 of this embodiment, the learning unit 193 complements missing values ​​or abnormal values ​​contained in the learning data LD1 and LD2 before learning the models M1 and M2.

[0129] (4-5) In the heat load prediction system 1 of this embodiment, the learning unit 193 processes the learning data LD1, LD2 based on the operating state of the air conditioner 2 or the schedule of the air conditioner 2 before learning the models M1, M2.

[0130] The thermal load prediction system 1 can improve the accuracy of models M1, M2 by excluding data that is weakly related to the thermal load of air conditioning zones Z1, Z2 based on the operating state of the air conditioner 2 or the schedule of the air conditioner 2.

[0131] (4-6) In the heat load prediction system 1 of this embodiment, the models M1 and M2 are a machine learning model, a statistical model, a physical model, or a combination thereof.

[0132] (4-7) In the heat load prediction system 1 of this embodiment, the learning unit 193 performs learning of the models M1 and M2 using learning data LD1 and LD2 for the past week (predetermined unit).

[0133] (4-8) In the heat load prediction system 1 of this embodiment, the learning unit 193 updates the models M1 and M2 once a week (at a predetermined timing).

[0134] (4-9) In the heat load prediction system 1 of the present embodiment, the heat exchange amount in each of the indoor units 20a to 20c is calculated based on the indoor suction temperature, air volume, and indoor refrigerant temperature in each of the indoor units 20a to 20c.

[0135] (4-10) The heat load prediction system 1 of the present embodiment further includes a determination unit 195 and a transmission unit 196. The determination unit 195 determines a target refrigerant temperature for the air conditioner 2 based on the heat loads of the air conditioning zones Z1 and Z2 predicted by the prediction unit 194. The transmission unit 196 transmits control content including the target refrigerant temperature determined by the determination unit 195 to the air conditioner 2.

[0136] As a result, the heat load prediction system 1 can perform feedforward control.

[0137] (4-11) In the heat load prediction system 1 of this embodiment, the determination unit 195 calculates the heat load to be processed by each of the indoor units 20a-20c based on the heat loads of the air conditioning zones Z1, Z2 predicted by the prediction unit 194 and information on the indoor units 20a-20c installed in the air conditioning zones Z1, Z2. The determination unit 195 calculates the required refrigerant temperature in each of the indoor units 20a-20c that is required for each of the indoor units 20a-20c to process the calculated heat load. The determination unit 195 determines a target refrigerant temperature for the air conditioner 2 that is required to bring the indoor refrigerant temperature in each of the indoor units 20a-20c closer to the required refrigerant temperature.

[0138] (4-12) In the heat load prediction system 1 of the present embodiment, the required refrigerant temperature is calculated based on the calculated heat load (heat exchange amount), indoor suction temperature, and air volume in each of the indoor units 20a to 20c.

[0139] (5) Variations (5-1) Variation 1A In this embodiment, the indoor temperatures of the air conditioning zones Z1 and Z2 are the measured values ​​of the indoor temperature sensors 631 and 632 installed in the air conditioning zones Z1 and Z2. However, the indoor temperatures of the air conditioning zones Z1 and Z2 may be calculated from the indoor intake temperatures of the indoor units 20a to 20c. For example, the indoor temperature of the air conditioning zone Z1 can be the average value of the indoor intake temperatures of the indoor units 20a and 20b. In other words, the indoor temperature of the air conditioning zone Z1 and the indoor intake temperatures of the indoor units 20a and 20b can be substituted for each other.

[0140] (5-2) Variation 1B In this embodiment, the outside air temperature of the building BL is a measurement value of the outdoor temperature sensor 65 installed outside the building BL. However, the outside air temperature of the building BL may be a measurement value of an outdoor intake temperature sensor that measures the temperature of air sucked into the outdoor unit 30. In this case, the outdoor intake temperature sensor is installed near the air intake port of the outdoor unit 30. In other words, the outside air temperature of the building BL and the measurement value of the outdoor intake temperature sensor can be substituted for each other.

[0141] (5-3) Variation 1C In this embodiment, the heat load prediction system 1 targets multiple air conditioning zones Z1 and Z2 and has a single refrigerant system. In other words, the heat load prediction system 1 targets a system having multiple air conditioning zones and a single refrigerant system. However, the heat load prediction system 1 may target a system having a single air conditioning zone and a single refrigerant system.

[0142] Furthermore, the thermal load prediction system 1 may target a system having a single air-conditioning zone and multiple refrigerant systems, or multiple air-conditioning zones and multiple refrigerant systems. In this case, the determination unit 195 calculates the thermal load to be processed by the indoor units based on the thermal load of one or multiple air-conditioning zones predicted by the prediction unit 194 and information on the indoor units installed in the air-conditioning zones. The determination unit 195 calculates the required refrigerant temperature required for the indoor units to process the calculated thermal load, and determines, for each of the multiple refrigerant systems, a target refrigerant temperature required to bring the indoor refrigerant temperature in the indoor unit closer to the required refrigerant temperature.

[0143] (5-4) Variation 1D In this embodiment, the air conditioner 2 includes the indoor units 20a to 20c and the outdoor unit 30. However, the air conditioner 2 may further include an air conditioner for processing outside air, such as a humidity regulator or a ventilator.

[0144] For example, the learning unit 193 calculates the heat exchange amount in the humidity regulator and the outdoor air processing air conditioner using a predetermined relational expression, similar to the indoor units 20a to 20c. The learning unit 193 adds up not only the heat exchange amount in the indoor units 20a to 20c but also the heat exchange amount in the humidity regulator and the outdoor air processing air conditioner for each air conditioning zone Z1, Z2, and calculates the heat load of the air conditioning zones Z1, Z2 in the learning data LD1, LD2. The determination unit 195 apportions the heat load of the air conditioning zones Z1, Z2 predicted by the prediction unit 194 not only to the heat load processed by the indoor units 20a to 20c, but also to the heat load processed by the humidity regulator and the outdoor air processing air conditioner, for example. The determination unit 195 determines the respective setting values ​​required for the humidity regulator and the outdoor air processing air conditioner to process the calculated heat load, similar to the indoor units 20a to 20c.

[0145] (5-5) Variation 1E In this embodiment, the external environment data D1 is the outside air temperature of the building BL. However, the external environment data D1 may further include the amount of solar radiation of the building BL, a weather forecast for the location of the building BL, cloud cover, wind speed, wind direction, and precipitation. For example, the environmental data acquisition unit 191 acquires the amount of solar radiation of the building BL from a solar radiation sensor installed outside the building BL. Also, for example, the prediction unit 194 predicts the amount of solar radiation 10 minutes after the data acquisition time point using a model learned from previously acquired amounts of solar radiation or representative amounts of solar radiation, weather forecasts, cloud cover, wind speed, wind direction, precipitation, and other information. Also, under the assumption that the amount of solar radiation of the building BL does not change significantly within 10 minutes, the most recently acquired amount of solar radiation of the building BL may be used as the amount of solar radiation of the building BL 10 minutes after the data acquisition time point.

[0146] In this embodiment, the internal environment data D2 is the indoor temperatures of the air-conditioning zones Z1 and Z2. However, the internal environment data D2 may further include the humidity, CO2 concentration, the amount of solar radiation reaching the windows and walls, the indoor temperatures of adjacent air-conditioning zones, the amount of outside air flowing in and out, and the amount of heat generated by office equipment, people, and lighting.

[0147] (5-6) Variation 1F In this embodiment, the learning data LD1, LD2 are composed of external environment data D1, internal environment data D2, and operation data D3. However, the heat load prediction device 3 may further use information on the building BL and information on the air conditioning zones Z1, Z2 as the learning data LD1, LD2.

[0148] The information on the building BL is, for example, the address, latitude and longitude, use, and heat transfer coefficient of the wall of the building BL. The information on the air-conditioning zones Z1 and Z2 is, for example, the use of the air-conditioning zones Z1 and Z2, the floor area, the window area ratio, the presence or absence of blinds, the ventilation type, the floor height, the wall length, the installed equipment (the number of ventilation devices installed, etc.), the window shading coefficient, the window heat transfer coefficient, the heat capacity of the furniture, etc., and the usage status (schedule, etc.).

[0149] (5-7) Variation 1G In this embodiment, the learning unit 193 learned the models M1 and M2 using the learning data LD1 and LD2 for each of the air conditioning zones Z1 and Z2. However, when the air conditioning zones Z1 and Z2 are similar in terms of heat load, etc., the learning unit 193 may learn the model M1 using the learning data LD2 for the air conditioning zone Z2 instead of the learning data LD1 for the air conditioning zone Z1, for example.

[0150] As a result, the thermal load prediction device 3 can share learning data between similar air conditioning zones.

[0151] (5-8) Variation 1H In this embodiment, the learning unit 193 uses a predetermined relational equation that holds between the heat exchange amount, indoor suction temperature, air volume, and indoor refrigerant temperature in the indoor units 20a to 20c to calculate the heat exchange amount in each of the indoor units 20a to 20c from the indoor suction temperature, air volume, and indoor refrigerant temperature in each of the indoor units 20a to 20c.

[0152] However, the heat exchange amount in each of the indoor units 20a-20c may be calculated from the indoor intake temperature, indoor discharge temperature, and air volume in each of the indoor units 20a-20c using a predetermined relational expression that holds between the heat exchange amount, indoor intake temperature, indoor discharge temperature, and air volume in the indoor units 20a-20c. In this case, the indoor discharge temperature in the indoor units 20a-20c is acquired from an indoor discharge temperature sensor that measures the temperature of the air discharged by the indoor units 20a-20c. The indoor discharge temperature sensor is provided near the air outlet of the indoor units 20a-20c.

[0153] Furthermore, the learning unit 193 may further add the indoor humidity of the air conditioning zones Z1, Z2 to calculate the heat exchange amount in each of the indoor units 20a-20c. The indoor humidity of the air conditioning zones Z1, Z2 is acquired, for example, from an indoor humidity sensor that measures the humidity of the air sucked into the indoor units 20a-20c. The indoor humidity sensor is provided near the air inlets of the indoor units 20a-20c. As a result, the learning unit 193 can more accurately calculate the heat exchange amount in each of the indoor units 20a-20c.

[0154] (5-9) Variation 1I In this embodiment, the determination unit 195 calculated the heat load to be processed by each of the indoor units 20a-20c based on the heat loads of the multiple air conditioning zones Z1, Z2 predicted by the prediction unit 194 and information related to the indoor units 20a-20c installed in the air conditioning zones Z1, Z2. The determination unit 195 calculated the required refrigerant temperature in each of the indoor units 20a-20c required for each of the indoor units 20a-20c to process the calculated heat load, and determined a target refrigerant temperature for the air conditioner 2 required to bring the indoor refrigerant temperature in each of the indoor units 20a-20c closer to the required refrigerant temperature.

[0155] However, when the heat load prediction system 1 targets a system having multiple air conditioning zones and a single refrigerant system (corresponding to this embodiment), or a single air conditioning zone and a single refrigerant system, the determination unit 195 may aggregate the heat loads of one or more air conditioning zones predicted by the prediction unit 194 by the maximum value (maximum heat load) and determine the target refrigerant temperature of the refrigerant system required to process the maximum heat load. The determination unit 195 determines the target refrigerant temperature using, for example, a predetermined relational expression that holds between the heat exchange amount (maximum heat load), the compressor rotation speed, and the target refrigerant temperature. At this time, the operation data acquisition unit 192 acquires the compressor rotation speed of the refrigerant system as the operation data D3.

[0156] Furthermore, when the thermal load prediction system 1 targets a system having a single air conditioning zone and multiple refrigerant systems, or multiple air conditioning zones and multiple refrigerant systems, the determination unit 195 may calculate the thermal load to be processed by the indoor units based on the thermal load of one or more air conditioning zones predicted by the prediction unit 194 and information on the indoor units installed in the air conditioning zone, aggregate the calculated thermal loads by the maximum value (maximum thermal load) for each of the multiple refrigerant systems, and determine a target refrigerant temperature for each refrigerant system required to process the maximum thermal load for each aggregated refrigerant system.

[0157] (5-10) Variation 1J In this embodiment, the learning unit 193 adds up the heat exchange amounts in the indoor units 20a-20c for each air conditioning zone Z1, Z2 to calculate the heat loads of the air conditioning zones Z1, Z2 in the learning data LD1, LD2. However, the heat loads of the air conditioning zones Z1, Z2 may be calculated based on the distribution of the heat exchange amounts in the indoor units 20a-20c or corrected values ​​of the heat exchange amounts in the indoor units 20a-20c.

[0158] (5-11) Variation 1K In this embodiment, the environmental data acquisition unit 191 acquires the external environmental data D1 and the internal environmental data D2. However, the environmental data acquisition unit 191 may acquire only one of the external environmental data D1 and the internal environmental data D2.

[0159] In this case, for example, the learning unit 193 uses only one of the external environment data D1 and the internal environment data D2 as the learning data LD1, LD2.

[0160] (5-12) Variation 1L In this embodiment, the air conditioner 2 acquires the indoor saturation temperatures of the indoor units 20a-20c from the indoor heat exchanger temperature sensors 62a-62c as the operating data D3, and transmits these to the heat load prediction device 3. However, the air conditioner 2 may acquire the suction pressure or the discharge pressure instead of the indoor saturation temperature, and transmit these to the heat load prediction device 3. The heat load prediction device 3 calculates the indoor saturation temperature from the acquired suction pressure or discharge pressure.

[0161] Furthermore, in this embodiment, the air conditioner 2 calculates the degree of superheat or subcooling as the operating data D3 and transmits these to the thermal load prediction device 3. However, the air conditioner 2 may obtain the indoor gas side temperature or the indoor liquid side temperature instead of the degree of superheat or subcooling, and transmit these to the thermal load prediction device 3. The thermal load prediction device 3 calculates the degree of superheat or subcooling from the obtained indoor gas side temperature or indoor liquid side temperature.

[0162] (5-13) Variation 1M In this embodiment, the heat load prediction device 3 learns by associating the indoor temperatures of the air conditioning zones Z1 and Z2 and the outdoor temperature of the building BL at the same time with the heat loads of the air conditioning zones Z1 and Z2. However, the heat load prediction device 3 may learn by associating the indoor temperatures of the air conditioning zones Z1 and Z2 and the outdoor temperature of the building BL at a certain point in time with the heat loads of the air conditioning zones Z1 and Z2 10 minutes after that point in time.

[0163] As a result, the thermal load prediction device 3 can predict the thermal load of air conditioning zones Z1 and Z2 10 minutes after the data acquisition time point using the indoor temperatures of air conditioning zones Z1 and Z2 and the outdoor air temperature of building BL at the time of data acquisition, without using the indoor temperatures of air conditioning zones Z1 and Z2 and the outdoor air temperature of building BL 10 minutes after the data acquisition time point.

[0164] (5-14) Although the embodiments of the present disclosure have been described above, it will be understood that various changes in form and details can be made without departing from the spirit and scope of the present disclosure described in the claims. [Explanation of symbols]

[0165] 1. Heat load prediction system 11 Storage section 20a~20c indoor unit 191 Environmental Data Acquisition Department 192 Operation data acquisition unit 193 Learning Department 194 Prediction Department BL Building D1 External environment data D2 Internal Environment Data D3 Operation Data LD1, LD2 learning data M1, M2 models Z1, Z2 air conditioning zone [Prior art documents] [Patent documents]

[0166] [Patent Document 1] Patent No. 5572799

Claims

1. A heat load prediction system (1) for predicting heat loads of a plurality of air conditioning zones (Z1, Z2) in a building (BL), An environmental data acquisition unit (191) that acquires external environmental data (D1) related to an environment outside the building and / or internal environmental data (D2) related to an environment inside the building that affects the heat load of the plurality of air conditioning zones; an operation data acquisition unit (192) that acquires operation data (D3) of an air conditioner (2) that performs air conditioning of the plurality of air conditioning zones; A storage unit (11) that stores data acquired by the environmental data acquisition unit and the driving data acquisition unit; A learning unit (193) that uses learning data (LD1, LD2) obtained from the memory unit to learn a model (M1, M2) that predicts the heat loads of the plurality of air conditioning zones; a prediction unit (194) for predicting heat loads of the plurality of air conditioning zones using the model; Equipped with The air conditioner has a plurality of indoor units (20a to 20c) connected to a same outdoor unit (30), The indoor units are distributed and installed in the air conditioning zones, the thermal loads of the plurality of air conditioning zones in the learning data are calculated based on heat exchange amounts in indoor units installed in the plurality of air conditioning zones; The heat exchange amount in the indoor unit is calculated based on the operating data. Heat load prediction system (1).

2. The external environment data includes at least an outside air temperature of the building, The internal environment data includes at least indoor temperatures of the plurality of air conditioning zones. A heat load forecasting system (1) according to claim 1.

3. The learning unit excludes records including missing values ​​or abnormal values ​​from the learning data before learning the model. A heat load forecasting system (1) according to claim 1 or 2.

4. The learning unit complements missing values ​​or abnormal values ​​contained in the learning data before learning the model. A heat load forecasting system (1) according to claim 1 or 2.

5. The learning unit processes the learning data based on an operating state of the air conditioner or a schedule of the air conditioner before learning the model. A heat load forecasting system (1) according to any one of claims 1 to 4.

6. The learning unit uses second learning data for a second air conditioning zone similar to the first air conditioning zone, instead of the first learning data for the first air conditioning zone. A heat load forecasting system (1) according to any one of claims 1 to 5.

7. The model is a machine learning model, a statistical model, a physical model, or a combination thereof. A heat load forecasting system (1) according to any one of claims 1 to 6.

8. The learning unit learns the model using a predetermined unit of the learning data. A thermal load forecasting system (1) according to any one of claims 1 to 7.

9. The learning unit updates the model at a predetermined timing. A thermal load forecasting system (1) according to any one of claims 1 to 8.

10. The heat exchange amount in the indoor unit is calculated based on the suction temperature, air volume, and refrigerant temperature in the indoor unit. A heat load forecasting system (1) according to any one of claims 1 to 9.

11. The heat exchange amount in the indoor unit is calculated based on the suction temperature, the blowing temperature, and the air volume in the indoor unit. A heat load forecasting system (1) according to any one of claims 1 to 9.

12. The air conditioner further includes an air conditioner for treating outdoor air, the outside air processing air conditioner is installed in a third air conditioning zone included in the plurality of air conditioning zones, the heat load of the third air-conditioning zone in the learning data is further calculated based on a heat exchange amount in the outside air processing air conditioner; The heat exchange amount in the outdoor air processing air conditioner is calculated based on the operating data. A thermal load forecasting system (1) according to any one of claims 1 to 11.

13. The outdoor air processing air conditioner includes a humidity regulator or a ventilation device. A thermal load forecasting system (1) according to claim 12.

Images