Inference device, learning device, and air conditioning device

Abstract

An inference device for obtaining an inferred value of the time required to melt frost in relation to the temperature of an outdoor heat exchanger which constitutes part of an outdoor unit of an air-conditioning device, or to defrosting temperature information which expresses the state of change in said temperature, the time required to melt frost being the period during which the temperature of the outdoor heat exchanger is stable within a first range, and the inference device being equipped with a first data acquisition unit for acquiring the defrosting temperature information for the air-conditioning device, and an inference unit for obtaining an inferred value of the time required to melt frost in relation to the defrosting temperature information, on the basis of the defrosting temperature information acquired by the first data acquisition unit, by using a learned model for inferring the time required to melt frost from the defrosting temperature information.

Description

Technical Field

[0001] The present disclosure relates to an inference device and a learning device for an air conditioner having a defrosting function. , and air conditioning equipment It relates to.

Background Art

[0002] When the air conditioner is operated for heating in winter, frost may form on the outdoor heat exchanger. For this reason, a defrosting operation (defrost operation) has been proposed that melts the frost on the outdoor heat exchanger by using the heat of the indoor heat exchanger mounted on the indoor unit (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In Patent Document 1, when a certain period of time has elapsed since the previous defrosting operation ended and the heating operation was restarted, and the detected temperature of the temperature sensor has become lower than a specified value, it is assumed that the condition for starting the defrosting operation is satisfied, and the defrosting operation is started.

[0005] However, in the control based on such a rule, there is a problem that the defrosting operation may be started even though there is almost no frost on the outdoor heat exchanger, resulting in inferior energy efficiency and comfort.

[0006] The present disclosure has been made to solve such problems, and an inference device capable of improving energy efficiency by appropriately determining the start timing of the defrosting operation 、 learning device and air conditioning systems is aimed at obtaining.

Means for Solving the Problems

[0007] The inference device relating to this disclosure is an inference device that determines an estimated value of the time required for frost thawing based on frost temperature information indicating the temperature of the outdoor heat exchanger of the outdoor unit of an air conditioning system or the state of change in said temperature, wherein the time required for frost thawing is During the period in which the refrigerant flowing through the refrigerant piping of the outdoor heat exchanger melts the frost adhering to the outdoor heat exchanger, The temperature of the outdoor heat exchanger is stable within a first range for a specified period. The inference device comprises a first data acquisition unit that acquires the defrosting temperature information of the air conditioner, and an inference unit that uses a trained model to infer the defrosting time from the defrosting temperature information, and calculates an inferred value of the defrosting time for the defrosting temperature information based on the defrosting temperature information acquired by the first data acquisition unit.

[0008] The learning device according to this disclosure is a learning device that generates a trained model for determining an estimated value of the time required for frost melting for defrost temperature information indicating the temperature of an outdoor heat exchanger of an outdoor unit of an air conditioner, wherein the time required for frost melting is the period during which the refrigerant flowing through the refrigerant piping of the outdoor heat exchanger melts the frost attached to the outdoor heat exchanger, and the period during which the temperature of the outdoor heat exchanger stabilizes within a first range, and the learning device comprises a second data acquisition unit that acquires learning data created based on a combination of the defrost temperature information and the measured value of the time required for frost melting, and a model generation unit that generates a trained model for determining an estimated value of the time required for frost melting from the defrost temperature information of the air conditioner by learning using the learning data so that the estimated value of the time required for frost melting for the defrost temperature information approaches the measured value of the time required for frost melting. The air conditioning system according to this disclosure comprises a refrigerant circuit including a compressor, a condenser, an expansion valve, and an evaporator. An outdoor heat exchanger that functions as the evaporator or condenser, and the outdoor heat exchanger The evaporator When it functionsThe system comprises a temperature sensor for detecting the evaporation temperature, a receiving unit for receiving an estimated value of the frost thawing time from the inference device, and a control unit. The control unit starts the defrosting operation of the evaporator when (a) the estimated value of the frost thawing time is equal to or greater than a threshold, (b) a certain amount of time has elapsed since the start of heating operation, and (c) the evaporation temperature detected by the temperature sensor is equal to or less than a specified value. [Effects of the Invention]

[0009] According to the inference device and learning device described herein, by inferring the time required for frost melting as the amount of frost on the outdoor heat exchanger, it is possible to appropriately determine the start time of defrosting operation and improve energy efficiency. [Brief explanation of the drawing]

[0010] [Figure 1] This is a schematic diagram showing an example of the configuration of the refrigerant circuit 100 of an air conditioning system 101 to which the inference device 20 and learning device 30 according to Embodiment 1 are applied. [Figure 2] This flowchart shows an example of defrost operation control for the air conditioning system 101 according to Embodiment 1. [Figure 3] This figure shows an example of the change over time of the defrosting temperature θ during defrosting operation. [Figure 4] This is a block diagram showing the configuration of the inference device 20 according to Embodiment 1. [Figure 5] This is a flowchart showing the processing flow of the inference device 20 according to Embodiment 1. [Figure 6] This is a flowchart showing the processing flow of the air conditioning system 101 according to Embodiment 1. [Figure 7] This is a block diagram showing the configuration of the learning device 30 according to Embodiment 1. [Figure 8] This figure schematically shows an example of a neural network 34 model provided by the model generation unit 32. [Figure 9] This is a flowchart showing the processing flow of the learning device 30 according to Embodiment 1.

Best Mode for Carrying Out the Invention

[0011] Hereinafter, embodiments of the inference device 20 and the learning device 30 according to the present disclosure will be described with reference to the drawings. The present disclosure is not limited to the following embodiments, and various modifications can be made without departing from the gist of the present disclosure. In addition, the present disclosure includes all possible combinations of the configurations shown in the following embodiments and their modifications. Also, in each figure, those with the same reference numerals are the same or corresponding ones, which is common throughout the entire specification. Note that in each drawing, the relative dimensional relationships or shapes of each component may be different from the actual ones.

[0012] Embodiment 1. Hereinafter, the inference device 20 and the learning device 30 according to Embodiment 1 will be described. The inference device 20 and the learning device 30 are mounted on or connected to the air conditioner 101 and used.

[0013] <Configuration of the air conditioner 101> FIG. 1 is a schematic diagram showing an example of the configuration of a refrigerant circuit 100 of an air conditioner 101 to which the inference device 20 and the learning device 30 according to Embodiment 1 are applied.

[0014] As shown in FIG. 1, the air conditioner 101 includes an indoor unit 1 disposed in an indoor space to be air-conditioned and an outdoor unit 2 disposed on the outdoor side. The environment where the outdoor unit 2 is installed is hereinafter referred to as the outside air environment. The indoor unit 1 includes an indoor heat exchanger 5. On the other hand, the outdoor unit 2 includes a compressor 3, a four-way valve 4, an outdoor heat exchanger 6, and an expansion valve 7. The compressor 3, the four-way valve 4, the outdoor heat exchanger 6, the expansion valve 7, and the indoor heat exchanger 5 are connected via refrigerant pipes to form a refrigerant circuit 100.

[0015] The indoor heat exchanger 5 performs heat exchange between the refrigerant flowing through the refrigerant pipes arranged inside and the indoor air. On the other hand, the outdoor heat exchanger 6 performs heat exchange between the refrigerant flowing through the refrigerant pipes arranged inside and the outside air. The indoor heat exchanger 5 and the outdoor heat exchanger 6 are, for example, fin-and-tube type heat exchangers. Note that the indoor heat exchanger 5 and the outdoor heat exchanger 6 may each be divided into a plurality of heat exchangers. In that case, those plurality of heat exchangers are connected in series or in parallel.

[0016] The compressor 3 sucks in the refrigerant flowing through the refrigerant circuit 100. The compressor 3 compresses and discharges the sucked-in refrigerant. The compressor 3 is, for example, an inverter compressor. The refrigerant discharged from the compressor 3 flows into the indoor heat exchanger 5 or the outdoor heat exchanger 6.

[0017] The four-way valve 4 is a flow path switching device configured to switch states between the case of a cooling operation for cooling the indoor space where the indoor unit 1 is arranged and the case of a heating operation for heating. FIG. 1 shows a state where the air conditioner 101 is performing a heating operation. As shown in FIG. 1, when the air conditioner 101 is performing a heating operation, the four-way valve 4 is in the state shown by the solid line in FIG. 1, and the refrigerant discharged from the compressor 3 flows into the indoor heat exchanger 5. At this time, the outdoor heat exchanger 6 acts as an evaporator, and the indoor heat exchanger 5 acts as a condenser. On the other hand, when the air conditioner 101 is performing a cooling operation, the four-way valve 4 is in the state shown by the broken line in FIG. 1, and the refrigerant discharged from the compressor 3 flows into the outdoor heat exchanger 6. At this time, the outdoor heat exchanger 6 acts as a condenser, and the indoor heat exchanger 5 acts as an evaporator. Note that instead of the four-way valve 4, other flow path switching devices having the same function may be used.

[0018] The expansion valve 7 is a pressure reducing device that reduces the pressure of the refrigerant and is, for example, composed of an electronic expansion valve. The expansion valve 7 is provided between the outdoor heat exchanger 6 and the indoor heat exchanger 5. Note that instead of the expansion valve 7, other pressure reducing devices having the same function may be used.

[0019] The refrigerant circuit 100 is filled with refrigerant. The type of refrigerant is not particularly limited, but may be R32 or R410A, etc.

[0020] The indoor unit 1 is further equipped with an indoor fan 8 for providing airflow to the indoor heat exchanger 5. The indoor fan 8 is positioned on the upwind side of the indoor heat exchanger 5. However, the indoor fan 8 may also be positioned on the downwind side of the indoor heat exchanger 5.

[0021] The outdoor unit 2 is further equipped with an outdoor fan 9 for providing airflow to the outdoor heat exchanger 6. The outdoor fan 9 is positioned downwind of the outdoor heat exchanger 6. Alternatively, the outdoor fan 9 may be positioned upwind of the outdoor heat exchanger 6.

[0022] The outdoor fan 9 is equipped with a device that detects or estimates the current value used by the outdoor fan 9 for airflow. This device is referred to as the current measuring device 16. The current measuring device 16 consists of, for example, a current sensor or a processor. The current information detected or estimated by the current measuring device 16 is output to the control unit 15 provided in the outdoor unit 2. The current measuring device 16 may also be provided in the control unit 15.

[0023] A temperature sensor 10 is attached to the outer casing of the compressor 3 of the outdoor unit 2. The temperature sensor 10 detects the temperature of the compressor 3. The temperature sensor 10 may be attached to any location that can detect the temperature of the compressor 3. For example, instead of the outer casing of the compressor 3, the temperature sensor 10 may be installed in the refrigerant piping from the compressor 3 to the four-way valve 4. The compressor temperature information detected by the temperature sensor 10 is output to the control unit 15.

[0024] A temperature sensor 11 is mounted on the upwind side of the indoor fan 8 of the indoor unit 1. The temperature sensor 11 detects the air temperature before it flows into the indoor heat exchanger 5, that is, the room temperature. Note that the position of the temperature sensor 11 is not limited to the location shown in Figure 1, as long as it can detect the room temperature. The room temperature information detected by the temperature sensor 11 is output to the control unit 15.

[0025] A temperature sensor 12 is attached to the wall of the refrigerant piping of the indoor heat exchanger 5. The temperature sensor 12 detects the temperature of the indoor heat exchanger 5 when it functions as a condenser during heating, that is, the condensation temperature. The temperature sensor 12 is not limited to the location shown in Figure 1, as long as it can detect the temperature of the indoor heat exchanger 5. The condensation temperature information detected by the temperature sensor 12 is output to the control unit 15.

[0026] The outdoor unit 2 is equipped with a temperature sensor 13 for measuring the temperature of the air blown to the outdoor heat exchanger 6 by the outdoor fan 9. The temperature sensor 13 is mounted on the windward side of the outdoor heat exchanger 6 to measure the temperature of the air before it passes through the outdoor heat exchanger 6, i.e., the outside temperature. Note that the position of the temperature sensor 13 is not limited to the location shown in Figure 1, as long as it can detect the temperature of the air before it passes through the outdoor heat exchanger 6. The outside temperature information detected by the temperature sensor 13 is output to the control unit 15.

[0027] A temperature sensor 14 is attached to the wall of the refrigerant piping of the outdoor heat exchanger 6. The temperature sensor 14 detects the temperature of the outdoor heat exchanger 6 when it functions as an evaporator during heating, that is, the evaporation temperature. The location of the temperature sensor 14 is not limited to the location shown in Figure 1, as long as it is a part from which the temperature of the outdoor heat exchanger 6 can be estimated. The evaporation temperature information detected by the temperature sensor 14 is output to the control unit 15.

[0028] In Figure 1, five temperature sensors 10-14 are provided, but the number of temperature sensors 10-14 is not limited to the number shown in Figure 1; there may be more or fewer. For example, if the outdoor heat exchanger 6 is divided into multiple heat exchangers, multiple temperature sensors 14 may be attached to the outdoor heat exchanger.

[0029] Furthermore, the types of sensors are not limited to those shown in Figure 1. For example, a humidity sensor 17 that measures the humidity of the outdoor environment in which the outdoor unit 2 is installed, or an illuminance sensor 18 that measures the illuminance of the outdoor environment, may be installed on the outdoor unit 2. In that case, for example, both temperature information and humidity information of the outdoor environment can be obtained using the temperature sensor 13 and the humidity sensor 17. It is desirable that the illuminance measured by the illuminance sensor 18 is a value that indicates the amount of solar radiation on the housing of the outdoor unit 6. The information detected by these sensors 10-14 and 16-18 is aggregated in the control unit 15 provided in the outdoor unit 2.

[0030] The control unit 15 consists of a control board. The control board of the control unit 15 is equipped with a control device, a storage device, and a drive circuit. The control device consists of, for example, a CPU (Central Processing Unit) or microprocessor that executes a program stored in dedicated hardware or memory. The storage device is a non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, or EPROM (Erasable Programmable ROM), or a disk such as a magnetic disk, flexible disk, or optical disk.

[0031] <Operation of air conditioning unit 101> Next, we will explain the general operation of the air conditioning system 101 shown in Figure 1.

[0032] As described above, refrigerant is sealed in the refrigerant circuit 100, and the refrigerant is compressed by the compressor 3. During cooling, a refrigeration cycle consisting of the following cooling operation circuit is configured. That is, the refrigerant compressed by the compressor 3 condenses and liquefies in the outdoor heat exchanger 6, expands in the expansion valve 7, and then evaporates in the indoor heat exchanger 5 before returning to the compressor 3. In addition, a refrigeration cycle consisting of the cooling operation circuit is configured during defrosting operation.

[0033] On the other hand, during heating, the refrigeration cycle consists of the following heating operation circuit. That is, the refrigerant compressed by the compressor 3 condenses and liquefies in the indoor heat exchanger 5, expands in the expansion valve 7, evaporates in the outdoor heat exchanger 6, and then returns to the compressor 3.

[0034] As described above, when the air conditioning system 101 in Figure 1 performs cooling or heating, it controls each component so that the temperature detected by the indoor temperature sensor 11, i.e., the room temperature, reaches the target value. In other words, the air conditioning system 101 controls the rotational speed of the compressor 3, the opening degree of the expansion valve 7, the airflow rate provided by the indoor fan 8, and the airflow rate provided by the outdoor fan 9.

[0035] This control is performed based on the temperature detected by temperature sensors 10-14, and controls the cooling or heating capacity of the air conditioning unit 101. This control is performed by the control unit 15 of the outdoor unit 2.

[0036] During heating operation, under conditions of low temperature and high humidity, frost may form on the outdoor heat exchanger 6, which functions as an evaporator. In this case, the airflow resistance of the outdoor fan 9 increases, reducing the amount of heat exchanged in the outdoor heat exchanger 6 and decreasing the heating capacity. Therefore, the air conditioning system 101 performs a defrosting operation to melt the frost on the outdoor heat exchanger 6.

[0037] <Defrosting operation> Figure 2 is a flowchart showing an example of defrost operation control for the air conditioning system 101 according to Embodiment 1. The defrost operation in Figure 2 is a general example and is not limited thereto. The outline of the defrost operation in Figure 2 will be described below. Here, the temperature of the refrigerant piping of the outdoor heat exchanger 6 detected by the temperature sensor 14 is referred to as the defrosting temperature θ. Therefore, the defrosting temperature θ is the temperature of the outdoor heat exchanger 6 when it functions as an evaporator during heating.

[0038] As shown in Figure 2, in step S1, the control unit 15 determines whether a certain amount of time has elapsed since the air conditioner 101 started or restarted heating operation. This certain amount of time is a preset value. If the control unit 15 determines that the certain amount of time has not elapsed, it terminates the process shown in Figure 2. On the other hand, if the control unit 15 determines that the certain amount of time has elapsed, it proceeds to step S2.

[0039] In step S2, the control unit 15 determines whether the defrost temperature θ detected by the temperature sensor 14 is less than or equal to a preset value. If the control unit 15 determines that the defrost temperature θ is greater than the preset value, it terminates the process shown in the flow of Figure 2. On the other hand, if the control unit 15 determines that the defrost temperature θ is less than or equal to the preset value, it proceeds to step S3.

[0040] In step S3, the control unit 15 determines that the conditions for starting defrosting operation have been met based on the determinations in steps S1 and S2, that is, that frost has formed on the outdoor heat exchanger 6. The control unit 15 then stops the compressor 3 to start the defrosting operation.

[0041] In step S4, the control unit 15 switches the four-way valve 4 to configure the cooling operation circuit described above, restarts the compressor 3, and starts the defrosting operation. In the defrosting operation, for example, the frost attached to the outdoor heat exchanger 6 is melted using a reverse defrosting method that circulates the refrigerant. In the defrosting operation, the refrigerant, which has become high temperature and high pressure in the compressor 3, flows through the refrigerant piping to the outdoor heat exchanger 6. The refrigerant heats the frost attached to the outdoor heat exchanger 6, causing the frost to melt and turn into water.

[0042] The defrosting operation continues until the defrosting temperature θ reaches or exceeds the specified value. Therefore, in step S5, the control unit 15 determines whether the defrosting temperature θ has reached or exceeds the specified value. If the control unit 15 determines in step S5 that the defrosting temperature θ has reached or exceeds the specified value, the process proceeds to step S6.

[0043] In step S6, the control unit 15 determines that the conditions for ending the defrosting operation have been met, terminates the defrosting operation, and proceeds to step S7. Upon termination of the defrosting operation, the control unit 15 first stops the compressor 3, switches the four-way valve 4, and returns to the heating operation circuit described above.

[0044] In step S7, the control unit 15 restarts the compressor 3 and resumes heating operation.

[0045] While conventional defrosting control systems have also implemented such control, in environments with low outside temperatures, such as mid-winter or cold regions, the defrosting operation may be initiated even if there is no frost. In such cases, heating operation is temporarily stopped during defrosting, leading to an unnecessary drop in indoor temperature. Furthermore, extra energy is required for the defrosting operation itself, or for restoring the indoor temperature to the set value after it has dropped due to defrosting.

[0046] Therefore, the inference device 20 according to Embodiment 1 uses a trained model to obtain an inferred value Pest for the "frost thawing time" described later. The control unit 15 of the air conditioner 101 determines whether or not to start the flow shown in Figure 2 based on the inferred value Pest for the "frost thawing time". This makes it possible to avoid performing unnecessary defrosting operations when there is no frost on the outdoor heat exchanger 6. This will be explained in detail below.

[0047] <Time required for frost thawing> Figure 3 shows an example of the change in defrost temperature θ over time during defrosting operation. In Figure 3, the horizontal axis represents time, and the vertical axis represents the defrost temperature θ. However, the pattern of change in defrost temperature θ over time is not limited to that shown in Figure 3, as it varies depending on the values ​​of various parameters such as the outside air temperature, the amount of frost attached to the outdoor heat exchanger 6, and the number of times the compressor 3 is driven during defrosting operation. In the following explanation, we will use the case shown in Figure 3 as an example.

[0048] As shown in Figure 3, when defrosting is started at time t0, the defrosting temperature θ rises after the start of defrosting. Subsequently, the refrigerant flowing through the refrigerant piping of the outdoor heat exchanger 6 heats the frost attached to the outdoor heat exchanger 6, melting it. As a result, the defrosting temperature θ temporarily stabilizes around 0°C and rises again after the frost on the outdoor heat exchanger 6 has melted. Then, as described above, the defrosting operation ends when the defrosting temperature θ exceeds the specified value.

[0049] Therefore, in Embodiment 1, as shown in Figure 3, two temperatures θ1 and θ2 are set near 0°C. In this case, temperature θ1 is set to a lower temperature than temperature θ2 (θ1 < θ2). In this case, temperatures θ1 and θ2 are set such that, for example, the relationship θ1 ≤ 0 < θ2 holds. Note that temperature θ1 is the temperature just before the defrost temperature θ stabilizes. Temperature θ2 is the temperature just before the defrost temperature θ rises again. Let t1 be the time when the defrost temperature θ becomes temperature θ1, and t2 be the time when the defrost temperature θ becomes temperature θ2. In this case, the period during which time t satisfies the relationship t1 ≤ t ≤ t2 is the period during which the defrost temperature θ is near the melting point, and specifically, it is the period during which θ1 ≤ θ ≤ θ2. This period is the period during which the refrigerant flowing through the refrigerant piping of the outdoor heat exchanger 6 heats the frost attached to the outdoor heat exchanger 6 and melts the frost. In the following, this period will be referred to as the "time required for frost thawing," and will be indicated by the symbol "P" in Figure 3.

[0050] Therefore, if we define the range between temperatures θ1 and θ2 as the "first range," the frost melting time P is the period during which the defrost temperature θ, which is the defrosting temperature information, stabilizes within the first range. During the frost melting time P, a state change occurs in which the frost attached to the outdoor heat exchanger 6 changes into water. Therefore, during the frost melting time P, the thermal energy of the refrigerant flowing through the refrigerant piping is consumed not as sensible heat due to the change in defrosting temperature θ, but as latent heat due to the state change in which the frost changes into water. Thus, the frost melting time P is the period during which the frost attached to the outdoor heat exchanger 6 melts while the defrosting temperature θ, which is the defrosting temperature information, fluctuates around its melting point. For this reason, during the frost melting time P, the defrosting temperature θ, which is the defrosting temperature information, stabilizes within the first range.

[0051] In the above explanation, temperatures θ1 and θ2 are set appropriately such that, for example, the relationship θ1 ≤ 0 < θ2 holds. However, since the melting point may fluctuate slightly from 0°C due to the influence of impurities such as dirt, they may also be set so that any one of the following relationships holds: θ1 < 0 ≤ θ2, θ1 < θ2 ≤ 0, or 0 ≤ θ1 < θ2. However, temperatures θ1 and θ2 are set within the range of -20°C to +20°C, preferably within the range of -5°C to +5°C, or -10°C to +10°C. In any case, the "first range" is the range that includes the melting point or 0°C.

[0052] The frost thawing time P varies depending on the amount of frost adhering to the outdoor heat exchanger 6. When there is a large amount of frost, it takes a long time to completely melt the frost, resulting in a longer frost thawing time P. On the other hand, when there is a small amount of frost, the frost thawing time P is shorter. In cases where there is a small amount of frost and the frost thawing time P is extremely short, the defrosting temperature θ may rise above θ2 after reaching a temperature of θ1 or higher, without temporarily stabilizing around 0°C as shown in Figure 3.

[0053] Depending on the setting of the specified value for the defrosting temperature θ and how temperatures θ1 and θ2 are determined, the defrosting temperature θ may already be θ1 ≤ θ < θ2 at time t0. In this case, the thawing time P is counted from time t0, i.e., t1 = t0. Furthermore, the stopping and starting of the compressor 3 during defrosting operation, and the resulting series of transient phenomena, may cause an unstable refrigerant state, resulting in the defrosting temperature θ temporarily changing from θ < θ1 to θ1 ≤ θ < θ2, and then back to θ < θ1. In this case, the starting point t1 for counting the defrosting time P is set to the second time θ1 ≤ θ < θ2. If this phenomenon occurs multiple times, the starting point t1 for counting is set to the last time θ1 ≤ θ < θ2. Furthermore, the stopping and starting of the compressor 3 during defrosting operation, and the resulting series of transient phenomena, may cause an unstable refrigerant state, resulting in the defrosting temperature θ temporarily changing from θ1≦θ<θ2 to θ2≦θ before returning to θ1≦θ<θ2. In this case, the point at which θ2≦θ occurs for the second time is set as the end point t2 for counting the defrosting time P. If this phenomenon occurs multiple times, the point at which θ2≦θ occurs last is set as the end point t2 for counting. Note that the method for determining the start and end points for counting the frost thawing time P is just one example and is not limited to the above.

[0054] The control unit 15 stores control values ​​such as the operating frequency of the compressor 3, the current value of the outdoor fan 9, detected values ​​from temperature sensors 10 to 14, and values ​​such as the time required for frost thawing in a memory device provided in the control unit 15.

[0055] <Inference Phase> Next, the inference device 20 according to Embodiment 1 will be described. The inference device 20 according to Embodiment 1 calculates the inference value Pest of the frost thawing time P shown in Figure 3 above, and outputs the inference value Pest of the frost thawing time P to the air conditioner 101. The air conditioner 101 according to Embodiment 1 then performs the processing shown in the flow of Figure 2 only if the inference value Pest of the frost thawing time P output from the inference device 20 is equal to or greater than a predetermined threshold Th. Therefore, in Embodiment 1, defrosting is performed only when all three of the following conditions (a) to (c) are met. This makes it possible to avoid unnecessary defrosting. (a): The inferred value Pest of the frost thawing time P is greater than or equal to a predetermined threshold Th. (b) A certain amount of time has elapsed since the air conditioning unit 101 started heating operation (step S1 in Figure 2 is YES). (c): The current defrost temperature θ detected by the temperature sensor 14 is less than or equal to a preset value (step S2 in Figure 2 is YES).

[0056] The configuration of the inference device 20 according to Embodiment 1 will be described below with reference to Figure 4. Figure 4 is a block diagram showing the configuration of the inference device 20 according to Embodiment 1. As shown in Figure 4, the inference device 20 comprises a first data acquisition unit 21 and an inference unit 22. A trained model storage unit 33 or an external device 40 is also connected to the inference device 20.

[0057] The inference device 20 may also be included as one component of the air conditioning system 101 shown in Figure 1. In that case, the inference device 20 may be built into, for example, the outdoor unit 2 of the air conditioning system 101. Alternatively, the inference device 20 may be provided separately from the air conditioning system 101. For example, the inference device 20 may reside on a cloud server. In that case, the control unit 15 of the air conditioning system 101 and the inference device 20 are connected so that they can communicate with each other.

[0058] The first data acquisition unit 21 acquires the defrost temperature θ detected by the temperature sensor 14 as defrost temperature information. As mentioned above, the defrost temperature θ is the temperature of the outdoor heat exchanger 6 detected by the temperature sensor 14. The first data acquisition unit 21 may acquire the defrost temperature θ directly from the temperature sensor 14, or it may acquire the defrost temperature θ from the temperature sensor 14 via the control unit 15.

[0059] The inference unit 22 uses the trained model to calculate the estimated value Pest of the frost thawing time. Specifically, the inference unit 22 inputs the frost removal temperature θ acquired by the first data acquisition unit 21 into the trained model to calculate the estimated value Pest of the frost thawing time P from the frost removal temperature θ.

[0060] The trained model is generated by the learning device 30 (described later) and stored in the trained model storage unit 33. The inference unit 22 retrieves the trained model from the trained model storage unit 33. Alternatively, the inference unit 22 retrieves the trained model from an external device 40 via a communication line such as the Internet. The external device 40 may be, for example, one or more other air conditioning units, a cloud server, or the homepage of the manufacturer of the air conditioning unit 101.

[0061] Here, the hardware configuration of the inference device 20 will be described. The inference device 20 consists of processing circuits that realize the functions of the first data acquisition unit 21 and the inference unit 22. The processing circuits consist of dedicated hardware or a processor. Dedicated hardware includes, for example, an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). The processor executes a program stored in memory. The inference device 20 also has a storage unit (not shown) that stores the program and calculation results. The storage unit consists of memory. Memory is a non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, or EPROM (Erasable Programmable ROM), or a disk such as a magnetic disk, flexible disk, or optical disk.

[0062] Next, the processing flow of the inference device 20 will be explained using Figure 5. Figure 5 is a flowchart showing the processing flow of the inference device 20 according to Embodiment 1. The processing in the flow of Figure 5 is performed repeatedly, for example, at a fixed period, when the air conditioner 101 is in heating operation.

[0063] As shown in Figure 5, first, in step S21, the first data acquisition unit 21 acquires the current defrost temperature θ detected by the temperature sensor 14. The defrost temperature θ at this time is the temperature of the outdoor heat exchanger 6 of the air conditioner 101 when it is functioning as an evaporator, and is the temperature before the defrost operation is started.

[0064] Next, in step S22, the inference unit 22 inputs the defrosting temperature θ obtained in step S21 to the trained model stored in the trained model storage unit 33, and obtains an inferred value Pest of the frost thawing time P. The trained model will be described later.

[0065] Next, in step S23, the inference unit 22 outputs the inferred value Pest of the frost thawing time P obtained by the trained model to the control unit 15 of the air conditioner 101.

[0066] The control unit 15 of the air conditioning system 101 receives the inferred value Pest of the frost thawing time P from the inference device 20 and performs the processing shown in Figure 6. Figure 6 is a flowchart showing the processing flow of the air conditioning system 101 according to Embodiment 1.

[0067] As shown in Figure 6, in step S31, the control unit 15 of the air conditioner 101 determines whether the inferred value Pest of the frost thawing time P received from the inference device 20 is smaller than a preset threshold Th. If the control unit 15 determines that the inferred value Pest of the frost thawing time P is smaller than the threshold Th, it proceeds to step S32. On the other hand, if the control unit 15 determines that the inferred value Pest of the frost thawing time P is greater than or equal to the threshold Th, it proceeds to step S33.

[0068] In step S32, the control unit 15 decides not to perform defrosting and terminates the process shown in Figure 6.

[0069] In step S33, the control unit 15 proceeds to the processing shown in the flow of Figure 2. In the flow of Figure 2, if the conditions of step S1 and step S2 are met, the control unit 15 performs defrosting operation.

[0070] As described above, in Embodiment 1, the air conditioner 101 compares the inferred value Pest of the frost thawing time P output from the inference device 20 with a threshold Th. If the inferred value Pest of the frost thawing time P is smaller than the threshold Th, even if a certain amount of time has elapsed since the start of heating operation (S1 in Figure 2 is YES) and the defrosting temperature θ has fallen below a specified value (S2 in Figure 2 is YES), the defrosting operation is not performed. This avoids unnecessary defrosting operations, reduces power consumption, and prevents a decrease in user comfort due to defrosting operations.

[0071] On the other hand, if the inferred value Pest for the frost thawing time P is greater than or equal to the threshold Th, a defrosting operation is performed when a certain amount of time has elapsed since the start of heating operation (S1 in Figure 2 is YES) and the defrosting temperature θ falls below a specified value (S2 in Figure 2 is YES). In other words, the defrosting operation is initiated when the conditions for starting defrosting are met by the conventional defrosting control of the air conditioner and when it is determined by inference that defrosting should begin. This allows defrosting to be performed at the appropriate timing when defrosting is necessary.

[0072] Thus, in Embodiment 1, the inference device 20 determines an inferred value Pest of the frost melting time P as the amount of frost, thereby appropriately determining the start time of the defrosting operation and improving energy efficiency.

[0073] In Embodiment 1, the defrosting temperature θ was used as the defrosting temperature information, but it does not have to be the value of the defrosting temperature θ itself. The defrosting temperature information may be a value that indicates the state of temperature change of the outdoor heat exchanger 6. That is, the defrosting temperature information may be, for example, at least one of the average value, cumulative value, integral value, maximum value, or minimum value of the defrosting temperature θ over the time interval from the end of the previous defrosting operation to the present. Alternatively, the defrosting temperature information may be the gradient α of the defrosting temperature θ with respect to time t, as shown in Figure 3. The gradient α indicates the ratio of the change in defrosting temperature θ to the change in time t. Incidentally, as shown in Figure 3, during the period of the frost thawing time P, the gradient α is 0 or approximately 0.

[0074] Furthermore, the inference device 20 may use humidity information in addition to defrosting temperature information to determine the inferred value Pest of the frost thawing time P. The higher the humidity of the outside environment, the greater the amount of frost on the outdoor heat exchanger 6. Therefore, the higher the humidity, the longer the frost thawing time P tends to be. In this case, the first data acquisition unit 21 acquires defrosting temperature information from the temperature sensor 14 and also acquires humidity information from the humidity sensor 17, which is the humidity of the outside environment where the outdoor unit 2 is installed. In this case, the trained model is a trained model for inferring the frost thawing time from the defrosting temperature information and humidity information. The inference unit 22 uses this trained model to determine the inferred value Pest of the frost thawing time P.

[0075] Alternatively, the inference unit 20 may use current information from the outdoor fan 9 of the outdoor unit 2 of the air conditioner 101, in addition to the defrosting temperature information, to infer the frost thawing time P. When the amount of frost on the outdoor heat exchanger 6 increases, the ventilation of the outdoor heat exchanger 6 deteriorates, increasing the load on the outdoor fan 9 and raising the current value. Therefore, the higher the current value rises relative to the rotation speed of the outdoor fan 9, the more likely it is that there is more frost on the outdoor heat exchanger 6, and the longer the frost thawing time P tends to be. In this case, the first data acquisition unit 21 acquires defrosting temperature information from the temperature sensor 14, and further acquires the current value used by the outdoor fan 9 for ventilation from the current measuring device 16 as current information. In this case, the learned model is a learned model for inferring the frost thawing time from the defrosting temperature information and humidity information. The inference unit 22 uses this learned model to obtain the inferred value Pest of the frost thawing time P.

[0076] Furthermore, the inference device 20 may use outside temperature information in addition to defrosting temperature information to infer the thawing time P. The higher the outside temperature, the shorter the thawing time P. In this case, the first data acquisition unit 21 acquires defrosting temperature information from the temperature sensor 14 and also acquires outside temperature information from the temperature sensor 13. In this case, the trained model is a trained model for inferring the thawing time from the defrosting temperature information and outside temperature information. The inference unit 22 uses this trained model to obtain an inferred value Pest of the thawing time P.

[0077] Furthermore, the inference device 20 may use illuminance information in addition to defrosting temperature information to infer the frost thawing time P. Generally, the greater the illuminance, the shorter the frost thawing time P tends to be. In this case, the first data acquisition unit 21 acquires defrosting temperature information from the temperature sensor 14 and also acquires illuminance information from the illuminance sensor 18, which represents the illuminance of the outdoor environment where the outdoor unit 2 is installed. In this case, the trained model is a trained model for inferring the frost thawing time from the defrosting temperature information and the illuminance information. The inference unit 22 uses this trained model to obtain an inferred value Pest for the frost thawing time P.

[0078] Thus, the inference device 20 can use at least one of the following to determine the inferred value Pest of the frost thawing time P using a trained model: frost temperature information, humidity information, current information, ambient temperature information, and illuminance information. The combination of these pieces of information can be changed as appropriate. The inference device 20 may acquire this information directly from each sensor, or it may acquire it via the control unit 15.

[0079] Furthermore, the inference device 20 outputs an inferred value Pest of the frost thawing time P using the trained model learned by the model generation unit 32 of the learning device 30. However, not limited to this case, as described above, the inference device 20 may also obtain a trained model from an external device 40 and output an inferred value Pest of the frost thawing time P based on that trained model.

[0080] <Learning Phase> The configuration of the learning device 30 according to Embodiment 1 will be described below with reference to Figure 7. Figure 7 is a block diagram showing the configuration of the learning device 30 according to Embodiment 1. As shown in Figure 7, the learning device 30 includes a second data acquisition unit 31, a model generation unit 32, and a trained model storage unit 33.

[0081] The second data acquisition unit 31 acquires the defrosting temperature θ as defrosting temperature information and the measured value Pm of the defrost thawing time P. The second data acquisition unit 31 also combines the defrosting temperature θ and the measured value Pm of the defrost thawing time P to generate training data. Here, the defrosting temperature θ is the surface temperature of the outdoor heat exchanger 6 of the air conditioner 101, and is the value detected by the temperature sensor 14. The measured value Pm of the defrost thawing time P is the period during which the temperature of the outdoor heat exchanger 6 was actually near the melting point during the defrosting operation of the air conditioner 101. Specifically, the measured value Pm of the defrost thawing time P is the length of time during which the defrosting temperature θ detected by the temperature sensor 14 satisfied the relationship θ1 ≤ θ ≤ θ2 for temperatures θ1 and θ2 shown in Figure 3. The measured value Pm of the frost thawing time P is measured by the control unit 15 of the air conditioning unit 101. That is, the control unit 15 calculates the measured value Pm of the frost thawing time P based on the defrosting temperature θ detected by the temperature sensor 14. The second data acquisition unit 31 may acquire the defrosting temperature θ and the measured value Pm of the frost thawing time P from the control unit 15, or it may acquire the defrosting temperature θ directly from the temperature sensor 14.

[0082] The model generation unit 32 learns the frost thawing time P based on training data created based on the combination of the defrosting temperature θ and the measured frost thawing time Pm output from the second data acquisition unit 31. In other words, the model generation unit 32 generates a trained model for inferring the optimal frost thawing time P from the defrosting temperature θ and the measured frost thawing time Pm of the air conditioner 101. Here, the training data is data that associates the defrosting temperature θ and the measured frost thawing time Pm with each other.

[0083] Here, we have used the example of using the defrost temperature θ as the defrost temperature information, but the explanation is not limited to this case. The defrost temperature information may be the current defrost temperature θ, or a value indicating the state of change in the defrost temperature θ. That is, the defrost temperature information may be the current defrost temperature θ, or at least one of the following: the average value, cumulative value, integral value, maximum value, or minimum value of the defrost temperature θ over the time interval from the end of the previous defrost operation to the present.

[0084] The learning device 30 is used to learn the frost thawing time P of the air conditioner 101. The learning device 30 may be provided as one component of the air conditioner 101, or it may be provided separately from the air conditioner 101. In that case, the learning device 30 is connected to the air conditioner 101 via a network, for example, together with the inference device 20. Furthermore, both the learning device 30 and the inference device 20 may be built into the air conditioner 101. In addition, the learning device 30 and the inference device 20 may reside on a cloud server.

[0085] The model generation unit 32 can use any known learning algorithm, such as supervised learning, unsupervised learning, or reinforcement learning, for training. As an example, the case where a neural network is applied will be described.

[0086] The model generation unit 32 learns the frost thawing time P by so-called supervised learning, for example, according to a neural network model. Here, supervised learning is a method in which a pair of input and result (label) data is provided to the learning device 30, which learns features in the training data and infers the result from the input.

[0087] A neural network consists of an input layer made up of multiple neurons, an intermediate layer (hidden layer) also made up of multiple neurons, and an output layer also made up of multiple neurons. The intermediate layer can consist of one or more layers.

[0088] Figure 8 is a schematic diagram showing an example of a neural network 34 model provided by the model generation unit 32. Here, we will explain using the example where the neural network 34 of the model generation unit 32 is a three-layer neural network as shown in Figure 8. The neural network 34 has three input layers X1, X2, and X3, two hidden layers Y1 and Y2, and three output layers Z1, Z2, and Z3.

[0089] At this time, when multiple inputs In1, In2, and In3 are input to input layers X1, X2, and X3 respectively, the inputs In1, In2, and In3 are multiplied by a predetermined first weight W1 in input layers X1, X2, and X3. In the example in Figure 8, the first weight W1 of input layer X1 is of two types: w11 and w12. Similarly, the first weight W1 of input layer X2 is of two types: w13 and w14, and the first weight W1 of input layer X3 is of two types: w15 and w16. The result of the multiplication by the first weight W1 is input to the hidden layers Y1 and Y2. Here, this multiplication result is called the first multiplication result.

[0090] In the hidden layers Y1 and Y2, the result of the first multiplication is multiplied by a predetermined second weight W2. In the example in Figure 8, the second weight W2 of hidden layer Y1 is one of three types: w21, w23, and w25. Similarly, the second weight W2 of hidden layer Y2 is one of three types: w22, w24, and w26. The multiplication result, after multiplying by the second weight W2, is input to the output layers Z1, Z2, and Z3. Here, this multiplication result is called the second multiplication result. The second multiplication result is output from the output layers Z1, Z2, and Z3 as output results Out1, Out2, and Out3. Output results Out1, Out2, and Out3 change depending on the values ​​of weights W1 and W2.

[0091] In Embodiment 1, the neural network 34 learns the frost thawing time P by so-called supervised learning, according to training data created based on the combination of frost removal temperature θ and frost thawing time Pm acquired by the second data acquisition unit 31.

[0092] In other words, the neural network 34 learns the frost thawing time P for a given frost removal temperature θ using the following procedure. First, the first weight W1 and the second weight W2 are adjusted so that the output results Out1, Out2, and Out3 output from the output layers Z1, Z2, and Z3 when the frost removal temperature θ is input to the input layers X1, X2, and X3 approach the measured value Pm of the frost thawing time P. In this way, the first weight W1 and the second weight W2 are set, and the frost thawing time P for a given frost removal temperature θ is learned.

[0093] The model generation unit 32 generates and outputs a trained model of the frost thawing time P for a given frost removal temperature θ by performing the learning described above.

[0094] The trained model storage unit 33 stores the trained model output from the model generation unit 32.

[0095] Here, the hardware configuration of the learning device 30 will be described. The learning device 30 consists of processing circuits that realize the functions of the second data acquisition unit 31 and the model generation unit 32. The processing circuits consist of dedicated hardware or a processor. Dedicated hardware includes, for example, an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). The processor executes a program stored in memory. The learning device 30 also has a storage device (not shown) that stores the program and calculation results. The storage device realizes the functions of the learned model storage unit 33. The storage device consists of memory. Memory is a non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, or EPROM (Erasable Programmable ROM), or a disk such as a magnetic disk, flexible disk, or optical disk.

[0096] Next, the processing flow of the learning device 30 will be explained using Figure 9. Figure 9 is a flowchart showing the processing flow of the learning device 30 according to Embodiment 1.

[0097] As shown in Figure 9, in step S41, the second data acquisition unit 31 acquires the defrosting temperature θ from the temperature sensor 14 and the measured value Pm of the thawing time P from the control unit 15. Although the defrosting temperature θ and the measured value Pm of the thawing time P are acquired simultaneously in this example, the system is not limited to this case. As long as the second data acquisition unit 31 can acquire the data for the defrosting temperature θ and the measured value Pm of the thawing time P in association with each other, the second data acquisition unit 31 may acquire the data for the defrosting temperature θ and the data for the measured value Pm of the thawing time P at different times.

[0098] Next, in step S42, the model generation unit 32 performs a learning process using the training data. The training data includes the defrosting temperature θ and the measured value Pm of the frost thawing time P. More specifically, the training data is created based on the combination of the defrosting temperature θ and the measured value Pm of the frost thawing time P acquired by the second data acquisition unit 31. Using this training data, the model generation unit 32 learns the frost thawing time P for the defrosting temperature θ through so-called supervised learning and generates a trained model.

[0099] Next, in step S43, the trained model storage unit 33 stores the trained model generated by the model generation unit 32.

[0100] In Embodiment 1, the case in which supervised learning is applied to the learning algorithm used by the model generation unit 32 was described, but this is not the only possible case. In addition to supervised learning, reinforcement learning, unsupervised learning, or semi-supervised learning can also be applied to the learning algorithm used by the model generation unit 32.

[0101] Furthermore, if multiple air conditioning units 101 exist in the same area or in different areas, the model generation unit 32 may learn the frost thawing time P according to the training data created for those multiple air conditioning units 101. That is, the second data acquisition unit 31 may acquire training data from multiple air conditioning units 101 used in the same area for the model generation unit 32 to learn the frost thawing time P. Alternatively, the second data acquisition unit 31 may acquire training data collected from multiple air conditioning units 101 operating independently in different areas. The second data acquisition unit 31 can also add or remove air conditioning units 101 from the list at any point during the process of collecting training data. In addition, the learning device 30 that has learned the frost thawing time P for a certain air conditioning unit 101 may be applied to another air conditioning unit 101, and the frost thawing time P for that other air conditioning unit 101 may be relearned and updated.

[0102] Furthermore, the learning algorithm used in the model generation unit 32 can be deep learning, which learns to extract the features themselves, or machine learning can be performed according to other known methods, such as genetic programming, functional logic programming, or support vector machines.

[0103] As described above, the inference device 20 according to Embodiment 1 includes a first data acquisition unit 21 that acquires defrost temperature information of the air conditioner 101. The inference unit 22 of the inference device 20 uses a trained model to determine an inferred value Pest of the frost thawing time P based on the frost temperature information acquired by the first data acquisition unit 21. The air conditioner 101 compares the inferred value Pest of the frost thawing time P determined by the inference device 20 with a threshold Th. If the inferred value Pest is smaller than the threshold Th, the air conditioner 101 does not start defrosting operation even if a certain amount of time has elapsed since the start of heating operation and the defrost temperature θ is below a specified value. This suppresses unnecessary defrosting operations. As a result, the power consumption of the air conditioner 101 is reduced and the reduction in user comfort due to unnecessary defrosting operations can be prevented.

[0104] On the other hand, the air conditioning unit 101 performs defrosting when the inferred value Pest of the frost thawing time P determined by the inference device 20 is greater than or equal to the threshold Th, a certain amount of time has elapsed since the start of heating operation (S1 in Figure 2 is YES), and the defrosting temperature θ falls below a specified value (S2 in Figure 2 is YES). This ensures that defrosting is performed at the appropriate time when it is necessary.

[0105] Furthermore, the learning device 30 according to Embodiment 1 includes a second data acquisition unit 31 that acquires learning data created based on a combination of defrosting temperature information and the measured value Pm of the frost thawing time P. The model generation unit 32 of the learning device 30 uses the learning data to learn so that the inferred value Pest of the frost thawing time P for the defrosting temperature information approaches the measured value Pm of the frost thawing time P. As a result, the model generation unit 32 generates a trained model that calculates the inferred value Pest of the frost thawing time P from the defrosting temperature information of the air conditioner 101. Since the learning device 30 generates a trained model using learning data that includes the measured value Pm of the frost thawing time P, it can accurately determine the inferred value Pest that is close to the measured value Pm of the frost thawing time P.

[0106] Thus, according to the inference device 20 and learning device 30 of Embodiment 1, by determining the inference value Pest of the frost melting time P as the amount of frost, it is possible to appropriately determine the start time of defrosting operation and improve energy efficiency. [Explanation of Symbols]

[0107] 1 Indoor unit, 2 Outdoor unit, 3 Compressor, 4 Four-way valve, 5 Indoor heat exchanger, 6 Outdoor heat exchanger, 7 Expansion valve, 8 Indoor fan, 9 Outdoor fan, 10 Temperature sensor, 11 Temperature sensor, 12 Temperature sensor, 13 Temperature sensor, 14 Temperature sensor, 15 Control unit, 16 Current measuring device, 17 Humidity sensor, 18 Illuminance sensor, 20 Inference device, 21 First data acquisition unit, 22 Inference unit, 30 Learning device, 31 Second data acquisition unit, 32 Model generation unit, 33 Learned model storage unit, 34 Neural network, 40 External device, 100 Refrigerant circuit, 101 Air conditioning system.

Claims

1. An inference device for determining an estimated time required for frost thawing based on frost temperature information indicating the temperature of the outdoor heat exchanger of an outdoor unit of an air conditioning system or the state of change in said temperature, The frost melting time is the period during which the refrigerant flowing through the refrigerant piping of the outdoor heat exchanger melts the frost adhering to the outdoor heat exchanger, and is the period during which the temperature of the outdoor heat exchanger stabilizes within the first range. The inference device is A first data acquisition unit that acquires the defrost temperature information of the air conditioner, An inference unit that uses a trained model to infer the time required for frost thawing from the frost temperature information, and obtains an inferred value of the time required for frost thawing based on the frost temperature information obtained by the first data acquisition unit. An inference device equipped with this.

2. The defrosting temperature information includes the defrosting temperature indicating the current temperature of the refrigerant piping of the outdoor heat exchanger, or at least one of the average value, cumulative value, integral value, maximum value, or minimum value of the defrosting temperature during the time interval from the end of the previous defrosting operation to the present. The inference device according to claim 1.

3. The trained model is a model that takes the defrost temperature information as input and outputs an inferred value of the time required for frost thawing based on the defrost temperature information. The trained model is a model trained to approximate the estimated frost thawing time for the frost removal temperature information to the measured frost thawing time, according to training data created based on a combination of the frost removal temperature information and the measured frost thawing time. The inference device according to claim 1 or 2.

4. The defrosting temperature information includes a gradient that shows the ratio of the change in defrosting temperature to the change in time. The inference device according to claim 2.

5. The first data acquisition unit acquires the defrosting temperature information and humidity information indicating the humidity of the outside air environment in which the outdoor unit of the air conditioner is installed. The inference unit uses the trained model that infers the frost thawing time from the frost temperature information and humidity information to determine an inferred value of the frost thawing time based on the frost temperature information and humidity information acquired by the first data acquisition unit. An inference device according to any one of claims 1 to 4.

6. The first data acquisition unit acquires the defrosting temperature information and current information indicating the current value used by the outdoor fan of the outdoor unit of the air conditioner, The inference unit uses the trained model, which infers the frost thawing time from the frost temperature information and the current information, to determine an inferred value of the frost thawing time based on the frost temperature information and the current information acquired by the first data acquisition unit. An inference device according to any one of claims 1 to 5.

7. The first data acquisition unit acquires the defrosting temperature information and the outside temperature information indicating the outside temperature of the outdoor environment where the outdoor unit of the air conditioner is installed. The inference unit uses the trained model that infers the frost thawing time from the frost temperature information and the ambient temperature information to determine an inferred value of the frost thawing time based on the frost temperature information and ambient temperature information acquired by the first data acquisition unit. An inference device according to any one of claims 1 to 6.

8. The first data acquisition unit acquires the defrosting temperature information and illuminance information indicating the illuminance of the outdoor air environment in which the outdoor unit of the air conditioner is installed. The inference unit uses the trained model that infers the frost thawing time from the frost temperature information and the illuminance information to determine an inferred value of the frost thawing time based on the frost temperature information and the illuminance information acquired by the first data acquisition unit. An inference device according to any one of claims 1 to 7.

9. A learning device that generates a trained model for calculating an estimated value of the time required for frost thawing based on frost temperature information indicating the temperature of the outdoor heat exchanger of an outdoor unit of an air conditioning system or the state of changes in said temperature, The frost melting time is the period during which the refrigerant flowing through the refrigerant piping of the outdoor heat exchanger melts the frost adhering to the outdoor heat exchanger, and is the period during which the temperature of the outdoor heat exchanger stabilizes within the first range. The learning device is A second data acquisition unit acquires training data created based on a combination of the aforementioned defrosting temperature information and the measured value of the defrosting time required, A model generation unit generates a trained model that calculates the estimated frost thawing time from the defrost temperature information of the air conditioner by training the model generation unit so that the estimated frost thawing time for the defrost temperature information approaches the measured frost thawing time, using the aforementioned training data. A learning device equipped with these features.

10. The aforementioned outdoor heat exchanger includes a plurality of outdoor heat exchangers, The aforementioned learning data includes the defrost temperature information for the plurality of outdoor heat exchangers. The learning device according to claim 9.

11. A refrigerant circuit comprising a compressor, a condenser, an expansion valve, and an evaporator, An outdoor heat exchanger that functions as the evaporator or condenser, A temperature sensor for detecting the evaporation temperature when the outdoor heat exchanger functions as the evaporator, A receiving unit that receives the estimated value of the frost thawing time from the inference device according to any one of claims 1 to 8, Control unit and Equipped with, The control unit, (a) If the estimated value of the time required for frost thawing is greater than or equal to the threshold, (b) A certain amount of time has elapsed since the start of heating operation, (c) When the evaporation temperature detected by the temperature sensor is less than or equal to a specified value, The defrosting operation of the evaporator is started. Air conditioning system.

Images