JP7310549B2 - air conditioning system - Google Patents

Description

The technology of the present disclosure relates to an air conditioning system.

2. Description of the Related Art There is known an air conditioning system that controls the operation of an air conditioner using a learning model generated by learning an operating state such as a set temperature of the air conditioner set by a user (Patent Document 1). Such an air conditioning system uses a learning model to predict the user's operation, outputs the prediction result to the air conditioner, and the air conditioner changes settings based on the prediction result. can provide a comfortable temperature environment.

JP 2015-117933 A

In the case of an air conditioner equipped with such a learning model, it is possible to notify the user of the prediction result by the learning model for the convenience of the user. On the other hand, as a learning model other than the above, for example, a learning model that predicts the temperature unevenness that may occur indoors due to changes in the outdoor environment (weather, temperature, etc.) and shows the prediction result of controlling the wind direction so that the temperature unevenness does not occur. I'll think about it. In such a learning model, if the prediction interval for predicting temperature unevenness is short, the wind direction may change frequently, and some users may feel uncomfortable. In this case, even if a change in wind direction is output as a prediction result from the learning model, if the prediction result is not beneficial to the user, the air conditioner may maintain the current wind direction without changing the wind direction. be done. In this way, the air conditioner determines whether or not the setting change indicated by the prediction result obtained using the learning model is beneficial to the user, and executes the setting change if it is determined not to be beneficial. It is conceivable not to. In an air conditioner equipped with such a learning model, if all the prediction results by the learning model are notified to the user, the user will be notified of unnecessary prediction results. In this case, the amount of setting change information to be notified to the user becomes large, and the user's convenience may be impaired.

The disclosed technique has been made in view of the above points, and aims to provide an air conditioning system that improves user convenience.

An air conditioning system in one aspect of an embodiment includes an air conditioning device and a communication terminal connected to the air conditioning device. The air conditioner includes a prediction unit that calculates a plurality of prediction results for executing a setting change instruction of the air conditioner using a learning model generated based on the operating state information of the air conditioner; a setting change unit for controlling an air conditioner; and a display object corresponding to an applied prediction result object among the plurality of prediction result objects calculated by the prediction unit, is transmitted to the communication terminal, and the plurality of prediction result objects. and a result transmitting unit that does not transmit to the communication terminal information about a non-applied prediction result object different from the applied prediction result object. The communication terminal displays the display object.

The disclosed air conditioning system can improve user convenience.

FIG. 1 is an explanatory diagram showing an air conditioning system of Embodiment 1. FIG. FIG. 2 is a block diagram showing an example of the hardware configuration of the adapter. FIG. 3 is a block diagram showing an air conditioner. FIG. 4 is a diagram showing a plurality of prediction result objects stored in the prediction result storage unit. FIG. 5 is a diagram showing a plurality of pieces of information possessed by one prediction result object out of a plurality of prediction result objects. FIG. 6 is a diagram showing a plurality of display objects stored in the history storage unit. FIG. 7 is a diagram showing one display object corresponding to a prediction result object among a plurality of display objects. FIG. 8 is a flow chart showing the operation of controlling an air conditioner using a plurality of learning models.

An air conditioning system according to embodiments disclosed in the present application will be described below with reference to the drawings. Note that the technology of the present disclosure is not limited by the following description. Also, in the following description, the same reference numerals are given to the same constituent elements, and overlapping explanations are omitted.

FIG. 1 is an explanatory diagram showing an air conditioning system 1 of Embodiment 1. FIG. The air conditioning system 1 includes an air conditioning device 10 , an access point 4 , a server device 5 , a relay device 6 , a communication terminal 7 (corresponding to the “communication terminal” of the present invention), and a communication network 8 . An air conditioner 10 includes an air conditioner 2 and an adapter 3 . The air conditioner 2 is a device that cools or heats the room. The communication terminal 7 is a terminal device such as a user's smart phone.

The adapter 3 has a communication function for wirelessly connecting the air conditioner 2 and the access point 4 and a control function for AI (Artificial Intelligence) control of the air conditioner 2 . The access point 4 is, for example, a device that connects the adapter 3 and the communication network 8 by wireless communication using WLAN (Wireless Local Area Network) or the like. The communication network 8 is, for example, a communication network such as the Internet. The server device 5 has a function of generating a learning model applied to the air conditioner 2 and a database storing operation history data and the like. The server device 5 is arranged in, for example, a data center. The relay device 6 is connected for communication with the communication network 8 and is connected for communication with the server device 5 . The relay device 6 transmits the driving history data used for generating or updating the learning model applied to the air conditioner 2 from the adapter 3 to the server device 5 via the communication network 8 . Also, the relay device 6 transmits the learning model generated or updated by the server device 5 to the adapter 3 via the communication network 8 . Incidentally, the relay device 6 is arranged, for example, in a data center or the like.

The relay device 6 has a first relay section 6A, a second relay section 6B, and a third relay section 6C. The first relay unit 6A relays various data related to AI control between the adapter 3 and the server device 5. FIG. The first relay unit 6A transmits the driving history data used for generating or updating the learning model received from the adapter 3 to the server device 5 via the communication network 8, and transmits the learning model generated or updated by the server device 5. It is transmitted to the adapter 3 via the communication network 8. The second relay unit 6B acquires, via the communication network 8, the operating conditions of the air conditioner 2 (operation modes such as cooling/heating, set temperature, etc.) set by the user using the communication terminal 7 from outside. , is transmitted to the air conditioner 2 via the communication network 8 . The third relay unit 6</b>C acquires external data such as a weather forecast from a communication network 8 such as the Internet, and transmits the acquired external data to the server device 5 . Also, the third relay unit 6C transmits external data to the adapter 3 via the communication network 8. FIG.

FIG. 2 is a block diagram showing an example of the hardware configuration of the adapter 3. As shown in FIG. The adapter 3 has a first communication section 11 , a second communication section 12 , a storage section 13 and a CPU (Central Processing Unit) 14 . The first communication unit 11 is a communication IF (Interface) such as a UART (Universal Asynchronous Receiver Transmitter) that is connected to the air conditioner 2 for communication. The second communication unit 12 is, for example, a communication IF such as WLAN, which is connected for communication with the access point 4 . The storage unit 13 has, for example, a ROM (Read Only Memory) and a RAM (Random Access Memory), and stores various information such as data and programs. The CPU 14 controls the adapter 3 as a whole by executing programs installed in the adapter 3 . The programs installed in the adapter 3 are composed of a plurality of programs for causing the adapter 3 to implement a plurality of functions. The multiple functions include a transmitter 16 , a receiver 17 , a predictor 18 and a discriminator 19 .

The transmission unit 16 transmits a plurality of set values (for example, an operation mode such as a cooling operation or a heating operation, a set temperature, a set air volume, etc.) set in the air conditioner 2 and a plurality of operation history data corresponding to a plurality of dates and times. and are collected from the air conditioner 2 . The operation history data corresponding to a certain date and time out of the plurality of operation history data includes the operation settings in which the air conditioner 2 is operating on that date and time, the operation settings input to the air conditioner 2 on that date and time, and the air Measured values detected by a plurality of sensors provided in the harmony machine 2 are shown. Operation settings include, for example, operation mode, set temperature, set air volume, and set air direction. Operation settings include, for example, an operation mode, set temperature, set air volume, set air direction, etc., which are changed by the user according to changes in outside temperature. Examples of measured values include indoor temperature, indoor humidity, indoor temperature distribution, indoor heat exchanger temperature, outdoor air temperature, outdoor air humidity, outdoor heat exchanger temperature, and the like. The transmitting unit 16 stores the collected set values and driving history data in the storage unit 13 . The transmission unit 16 creates driving state information based on the plurality of set values and the plurality of driving history data stored in the storage unit 13, and transmits the driving state information to the server device 5 via the communication network 8. do. The driving state information includes a plurality of setting values and a plurality of driving history data.

The receiving unit 17 receives the plurality of learning models transmitted from the server device 5 via the communication network 8 and stores the received plurality of learning models in the storage unit 13 . The prediction unit 18 generates a plurality of prediction result objects using a plurality of learning models stored in the storage unit 13 based on a plurality of measured values respectively detected by a plurality of sensors provided in the air conditioner 2. calculate. The prediction unit 18 outputs the calculated plurality of prediction result objects to the air conditioner 2 . The determination unit 19 transmits a request signal to the air conditioner 2 and acquires from the air conditioner 2 a plurality of thresholds corresponding to a plurality of learning models. The determination unit 19 determines whether each of the multiple prediction result objects calculated by the prediction unit 18 is valid or invalid based on multiple thresholds. Note that the prediction result object will be described later in detail with reference to FIGS. 4 and 5. FIG.

FIG. 3 is a block diagram showing the air conditioner 2. As shown in FIG. The air conditioner 2 includes an indoor unit 21, an outdoor unit 22, a plurality of sensors 23, a prediction result storage unit 24 (corresponding to the "first storage unit" of the present invention), and a history storage unit 25 ("second storage unit" of the present invention). ), a threshold storage unit 26 , a control device 27 and a remote controller 28 . The indoor unit 21 is arranged indoors and further includes an indoor fan, an indoor heat exchanger, and a wind direction plate (none of which are shown). The indoor fan generates an air flow in which the indoor air in which the indoor unit 21 is arranged passes through the indoor heat exchanger, and blows out the air that has passed through the indoor heat exchanger into the room. The indoor heat exchanger heats or cools the indoor air by exchanging heat between the refrigerant supplied from the outdoor unit 22 and the air whose flow is generated by the indoor fan. The wind direction plate changes the direction of the air blown from the indoor unit 21 into the room. The indoor unit 21 cools or heats the room by heating or cooling the indoor air with the indoor heat exchanger.

The outdoor unit 22 is arranged outdoors and includes an outdoor fan, an outdoor heat exchanger, a compressor, and an expansion valve (not shown). The outdoor fan generates an air flow in which outdoor air in which the outdoor unit 22 is arranged passes through the outdoor heat exchanger. The outdoor heat exchanger heats or cools the refrigerant by exchanging heat between the refrigerant and the outside air whose flow is generated by the outdoor fan. The compressor compresses the refrigerant and circulates the refrigerant between the indoor unit 21 and the outdoor unit 22 . The expansion valve expands and decompresses the refrigerant compressed by the compressor. The expansion valve further adjusts the amount of refrigerant flowing through the indoor unit 21 by adjusting the degree of opening thereof. The outdoor unit 22 supplies the indoor unit 21 with the refrigerant heated or cooled by the outdoor heat exchanger by causing the compressor to circulate the refrigerant between the indoor unit 21 and the outdoor unit 22 .

A plurality of sensors 23 are provided in the air conditioner 2 and output measured values of each sensor. The actual values measured by these sensors include the intake temperature and humidity of the air taken in by the indoor unit 21, the outdoor temperature where the outdoor unit 22 is arranged, the radiant temperature of the indoor floor, and the like.

The prediction result storage unit 24 stores a plurality of prediction result objects 42 in association with a plurality of learning models 41, as shown in FIG. FIG. 4 is a diagram showing a plurality of prediction result objects 42 stored in the prediction result storage unit 24. As shown in FIG. In this embodiment, the plurality of learning models 41 includes a sensory prediction learning model 43 , a summer temperature unevenness prediction learning model 44 , and a winter temperature unevenness prediction learning model 45 . Also, the plurality of prediction result objects 42 include a bodily sensation prediction result object 46 , a summer temperature unevenness prediction result object 47 , and a winter temperature unevenness prediction result object 48 . A bodily sensation prediction result object 46 is a prediction result object calculated using the bodily sensation prediction learning model 43 . The summer temperature unevenness prediction result object 47 is a prediction result object calculated using the summer temperature unevenness prediction learning model 44 . The winter temperature unevenness prediction result object 48 is a prediction result object calculated using the winter temperature unevenness prediction learning model 45 .

Each prediction result object 51 of the plurality of prediction result objects 42 has a plurality of pieces of information as shown in FIG. FIG. 5 is a diagram showing a plurality of pieces of information possessed by one prediction result object 51 out of the plurality of prediction result objects 42. As shown in FIG. A plurality of pieces of information that the prediction result object 51 has include a learning model type 53 , a flag 54 , date and time information 55 and a valid guess result 56 . The learning model type 53 indicates the type of learning model used to calculate the prediction result object 51 . Flag 54 is a value of "0" or "1". The date and time information 55 indicates the date and time when the prediction result object 51 was calculated, and indicates the year, month, day, hour, minute, and second of the date and time. A valid guess result 56 indicates a numerical value. A plurality of learning models 41 correspond to a plurality of events. The numerical value indicated by the valid guess result 56 corresponds to the probability of occurrence of an event corresponding to the learning model used to calculate the prediction result object 51 among a plurality of events, and the larger the value, the higher the probability of occurrence of the event. is shown.

For example, the bodily sensation prediction learning model 43 shown in FIG. 4 corresponds to an event in which the user changes the set temperature. A setting change corresponding to the bodily sensation prediction learning model 43 indicates a change in the set temperature. The valid guess result 56 of the bodily sensation prediction result object 46 corresponds to the probability that the user will change the set temperature.

The summer temperature nonuniformity prediction result object 47 and the winter temperature nonuniformity prediction result object 48 respectively correspond to a phenomenon in which temperature nonuniformity occurs in the room, and to a setting change to change the set wind direction. The effective estimation result 56 of the temperature unevenness prediction result object for summer 47 and the effective estimation result 56 of the temperature unevenness prediction result object for winter 48 correspond to the probability that temperature unevenness will occur in the room. The flag 54 of the summer temperature unevenness prediction result object 47 indicates whether the summer temperature unevenness prediction result object 47 has been determined as an applicable prediction result object or an unapplied prediction result object. The flag 54 of the winter temperature unevenness prediction result object 48 indicates whether the winter temperature unevenness prediction result object 48 has been determined as an applied prediction result object or an applied prediction result object. The flag 54 is "0" when it is determined as a non-applied prediction result object, and is "1" when it is determined as an applicable prediction result object.

As shown in FIG. 6, the history storage unit 25 stores a predetermined number (for example, eight) of display objects 61 . FIG. 6 is a diagram showing a plurality of display objects 61 stored in the history storage unit 25. As shown in FIG. A plurality of display objects 61 are displayed on the display section of the communication terminal 7 and are stored in the history storage section 25 in association with a plurality of dates and times. The date and time corresponding to a certain display object among the plurality of dates and times indicates the date and time when the prediction result object indicated by the display object was calculated by the prediction unit 18 . Each of the multiple display objects 61 corresponds to one prediction result object out of the multiple prediction result objects 42 .

One display object 62 corresponding to a prediction result object among the plurality of display objects 61 includes a learning model type 63, date and time information 64, and an inference result 65, as shown in FIG. It has multiple pieces of information. FIG. 7 is a diagram showing one display object 62 corresponding to a certain prediction result object among the plurality of display objects 61. As shown in FIG. The learning model type 63 indicates the learning model type 53 of the prediction result object. A guess result 65 indicates a valid guess result 56 for that predicted result object. The date/time information 64 indicates the date/time information 55 of the prediction result object, and indicates the date, hour, and minute of the date/time information 55 . The date/time information 64 is obtained by removing the second information from the date/time information 55 in FIG.

A threshold value corresponding to each of the plurality of learning models 41 is stored in the threshold storage unit 26 shown in FIG. That is, the thresholds stored in the threshold storage unit 26 are the sensory prediction threshold corresponding to the sensory prediction learning model 43, the summer temperature non-uniformity prediction threshold corresponding to the summer temperature non-uniformity prediction learning model 44, and the winter temperature non-uniformity prediction threshold. and a winter temperature unevenness prediction threshold corresponding to the learning model 45 .

The control device 27 controls each part of the air conditioner 2 and communicates with the adapter 3 . The control device 27 has multiple programs for implementing multiple functions. The plurality of functions are implemented by the control unit 31 , the determination unit 32 , the setting change unit 33 , the threshold value setting unit 34 and the result transmission unit 35 .

The control unit 31 changes a plurality of setting values of the air conditioner 2 by operating the remote controller 28, and changes a plurality of settings based on information acquired from the communication terminal 7 via the communication network 8 and the adapter 3. change the value. For example, the multiple set values include a set temperature and a set wind direction. The control unit 31 controls the indoor unit 21 and the outdoor unit 22 so that the indoor temperature approaches the changed set temperature when the set temperature is changed. The control unit 31 controls the wind direction plate of the indoor unit 21 so that air is blown out from the indoor unit 21 toward the changed set wind direction when the set wind direction is changed. Further, the control unit 31 controls the indoor unit 21 and the outdoor unit 22 so that the air conditioner 2 heats the room based on the set temperature when "heating" is selected by operating the remote controller 28. . The control unit 31 controls the indoor unit 21 and the outdoor unit 22 so that the air conditioner 2 cools the room based on the set temperature when "cooling" is selected by operating the remote controller 28 . The control device 27 controls the indoor unit 21 and the outdoor unit 22 so that the air conditioner 2 stops cooling or heating when "stop" is selected by operating the remote controller 28 .

The determination unit 32 determines whether a plurality of prediction result objects 42 stored in the prediction result storage unit 24 are applicable prediction result objects or not based on judgment criteria individually defined for each of the plurality of learning models 41. Determine if it is an object.

When the determination unit 32 determines that one prediction result object is the applied prediction result object, the setting change unit 33 changes settings corresponding to the learning model used to calculate the prediction result object. For example, when the learning model used to calculate the prediction result object identified as the applicable prediction result object corresponds to a change in the setting temperature, the setting change unit 33 changes the setting temperature based on the prediction result of the sensible temperature of the user. controls the air conditioner 2 so that When the learning model used to calculate the prediction result object determined to be the applied prediction result object corresponds to the change in the set wind direction, the setting change unit 33 directs the indoor unit 21 toward the changed set wind direction. The wind direction plate of the indoor unit 21 is controlled so that the air is blown out.

When the determination unit 32 determines that one prediction result object is the applied prediction result object, the setting change unit 33 stores the oldest display object among the plurality of display objects 61 in the history storage unit. Delete from 25. The setting change unit 33 further updates the history storage unit 25 so that a new display object indicating one prediction result object determined as the applied prediction result object is added to the plurality of display objects 61 .

The threshold setting unit 34 is an input terminal that can be communicably connected to the control device 27 , and when the threshold update input device 36 is connected to the control device 27 , the A plurality of thresholds stored in the threshold storage unit 26 are updated based on the input information. When the air conditioner 2 receives the AI control history request transmitted from the communication terminal 7, the result transmission unit 35 transmits the plurality of display objects 61 recorded in the history storage unit 25 to the communication terminal 7. Send via.

[Operation of air conditioning system 1]
The operation of the air conditioning system 1 includes the operation of continuously learning a plurality of learning models, the operation of controlling the air conditioner 2 using the plurality of learning models, and the control executed using the plurality of learning models. It comprises an operation of notifying history and an operation of updating a plurality of thresholds.

In the operation of continuously learning a plurality of learning models, the adapter 3 receives operation history data and a plurality of set values from the controller 27 of the air conditioner 2 at predetermined data acquisition intervals (for example, 5 minutes). to get The adapter 3 stores the driving history data and the plurality of setting values in the storage unit 13 in association with time. The adapter 3 creates driving state information at predetermined data transmission intervals (for example, 48 hours) and transmits the driving state information to the server device 5 via the communication network 8 . The driving state information indicates driving history data and a plurality of setting values stored in the storage unit 13 . When the server device 5 receives the driving state information transmitted from the adapter 3, the server device 5 continuously learns the plurality of learning models based on the driving state information, and obtains the latest plurality of learning models updated from the plurality of learning models. Generate. Server device 5 transmits the latest learning models to adapter 3 via relay device 6 . When adapter 3 receives the latest learning models transmitted from server device 5 , adapter 3 stores the received latest learning models in storage unit 13 .

FIG. 8 is a flow chart showing the operation of controlling the air conditioner 2 using a plurality of learning models. The adapter 3 acquires, from the air conditioner 2, a plurality of measured values respectively measured by the plurality of sensors 23 at predetermined prediction intervals (step S1). The prediction interval is roughly the same as the data acquisition interval, exemplified by 5 minutes. The adapter 3 uses a plurality of learning models stored in the storage unit 13 to calculate a plurality of prediction result objects based on a plurality of measured values (step S2). For example, the adapter 3 calculates a bodily sensation prediction result object 46 using the bodily sensation prediction learning model 43, calculates a summer temperature unevenness prediction result object 47 using the summer temperature unevenness prediction learning model 44, and calculates a winter temperature unevenness object 47. A winter temperature unevenness prediction result object 48 is calculated using the prediction learning model 45 .

The adapter 3 calculates a plurality of prediction result objects, and determines whether or not the prediction result objects are valid based on predetermined rules. Specifically, the adapter 3 transmits to the air conditioner 2 a request signal requesting a plurality of threshold values. Upon receiving the request signal transmitted from the adapter 3 , the controller 27 of the air conditioner 2 transmits a plurality of thresholds stored in the threshold storage unit 26 to the adapter 3 . Upon receiving the multiple thresholds, the adapter 3 determines whether each of the multiple prediction result objects is valid or invalid based on the multiple thresholds (step S3). Specifically, the adapter 3 determines the difference between the valid prediction result of the prediction result object corresponding to a certain learning model among the plurality of prediction result objects and the threshold corresponding to the learning model among the plurality of acquired thresholds. Determine size relationships. The adapter 3 determines that the prediction result object is valid when the valid guess result is greater than the threshold, and determines that the prediction result object is invalid when the valid guess result is less than the threshold. The adapter 3 outputs to the air conditioner 2 a plurality of valid prediction result objects determined to be valid among the plurality of prediction result objects. That is, the adapter 3 outputs a plurality of effective prediction result objects to the air conditioner 2 for each prediction interval (step S4).

Upon receiving the plurality of effective prediction result objects output from the adapter 3, the control device 27 of the air conditioner 2 causes the prediction result storage unit 24 to store the plurality of effective prediction result objects. The control device 27 determines whether one valid prediction result object of the plurality of prediction result objects 42 stored in the prediction result storage unit 24 is an applied prediction result object or a non-applied prediction result object based on a predetermined criterion. (step S5).

When it is determined that a certain prediction result object is an applied prediction result object (step S5, Yes), the control device 27 controls the indoor unit 21 and the outdoor unit 22 to calculate the prediction result object. A setting change corresponding to the used learning model is executed (step S6). When a certain prediction result object is determined to be an applied prediction result object (step S5, Yes), the control device 27 further creates a new display object indicating that prediction result object. Based on the plurality of dates associated with the plurality of display objects 61 stored in the history storage unit 25, the control device 27 stores the oldest display object among the plurality of display objects 61 in the history storage unit. Delete from 25. The control device 27 then updates the history storage unit 25 so that the new display object is added to the plurality of display objects 61 (step S7). At this time, the new display object is stored in the history storage unit 25 in association with the date and time when the new display object was stored in the history storage unit 25 . By updating the history storage unit 25 in this way, the history storage unit 25 does not store more than a predetermined number of display objects, and stores only the latest display objects related to each learning model. can do. Furthermore, the history storage unit 25 can store only a plurality of display objects 61 each representing a prediction result object determined to be an applied prediction result object.

The processing from step S5 to step S7 differs for each learning model. For example, the control device 27 determines whether or not the bodily sensation prediction result object 46 output from the adapter 3 immediately before is an application prediction result object at each predetermined set temperature change interval. The setting temperature change interval is equal to the prediction interval. That is, each time the sensory sensation prediction result object 46 is output by the adapter 3, it is determined whether or not it is an application prediction result object. The control device 27 changes the set temperature by controlling the indoor unit 21 and the outdoor unit 22 when the bodily sensation prediction result object 46 is the application prediction result object. That is, the air conditioner 2 can be controlled using the bodily sensation prediction learning model 43 to change the set temperature and appropriately perform cooling or heating so that the user feels comfortable.

When the air conditioner 2 is cooling, the control device 27 determines that the summer temperature unevenness prediction result object 47 output from the adapter 3 immediately before is the application prediction result object at each predetermined wind direction change interval. or not. The wind direction change interval is longer than the predicted interval, for example 40 minutes. The user may feel discomfort or discomfort due to frequent switching of the direction of the wind direction plate. Therefore, the wind direction change interval is set so that the operation of the wind direction plate is suppressed until a predetermined time elapses immediately after the direction of the wind direction plate is controlled, and is set longer than the predicted interval. The control device 27 regards the summer temperature unevenness prediction result object calculated immediately before the timing of every 40 minutes among the plurality of summer temperature unevenness prediction result objects calculated every 5 minutes as the applied prediction result object. discriminate. The control device 27 determines, among the plurality of summer temperature unevenness prediction result objects, other summer temperature unevenness prediction result objects that are not applicable prediction result objects to be non-application prediction result objects. When the summer temperature unevenness prediction result object 47 is the applied prediction result object, the control device 27 changes the wind direction corresponding to the summer temperature unevenness prediction learning model 44 by controlling the wind direction plate of the indoor unit 21. to change the direction of the air blown out from the indoor unit 21. - 特許庁

When the air conditioner 2 is heating, the control device 27 determines that the winter temperature unevenness prediction result object 48 output from the adapter 3 immediately before is the application prediction result object at each predetermined wind direction change interval. or not. The control device 27 regards the winter temperature unevenness prediction result object calculated immediately before the timing of every 40 minutes among the plurality of winter temperature unevenness prediction result objects calculated every 5 minutes as the applied prediction result object. discriminate. The control device 27 determines, among the plurality of winter temperature unevenness prediction result objects, other winter temperature unevenness prediction result objects that are not applicable prediction result objects to be non-applying prediction result objects. When the winter temperature unevenness prediction result object 48 is the applied prediction result object, the control device 27 changes the wind direction corresponding to the winter temperature unevenness prediction learning model 45 by controlling the wind direction plate of the indoor unit 21. to change the direction of the air blown out from the indoor unit 21. - 特許庁

That is, the air conditioner 2 is controlled by using the summer temperature unevenness prediction learning model 44 and the winter temperature unevenness prediction learning model 45, so that the temperature unevenness in the room is eliminated. Can be heated. Further, according to such an operation, the summer temperature unevenness prediction result object 47 and the winter temperature unevenness prediction result object 48 calculated by the adapter 3 are determined by the air conditioner because the wind direction change interval is longer than the prediction interval. 2 may be determined as a non-applied prediction result object without being used for the AI control. Therefore, the number of summer temperature unevenness prediction result objects 47 and winter temperature unevenness prediction result objects 48 determined as non-applied prediction result objects is greater than the bodily sensation prediction result objects 46 determined as non-applied prediction result objects.

In the operation of notifying the history of control executed using a plurality of learning models, the user executes a predetermined operation on the communication terminal 7 when he/she wishes to view the history of AI control. An example of the operation is tapping a browse button displayed on the display device of the communication terminal 7 by the application while the communication terminal 7 is running an application for controlling the air conditioner 10 . The communication terminal 7 transmits an AI control history request to the air conditioner 10 when the operation is executed. When the air conditioner 10 receives the AI control history request transmitted from the communication terminal 7 , the adapter 3 transfers the AI control history request to the control device 27 of the air conditioner 2 . Upon receiving the AI control history request, the control device 27 transmits a plurality of display objects 61 recorded in the history storage unit 25 to the communication terminal 7 via the adapter 3 . When receiving the plurality of display objects 61 transmitted via the adapter 3, the communication terminal 7 displays the plurality of display objects 61 on the display device.

According to such an operation, the air conditioning system 1 can notify the user of the history in which the air conditioner 2 has been controlled using a plurality of learning models. The display device of the communication terminal 7 may not be able to display all of the plurality of prediction result objects calculated by the adapter 3 on one screen. Moreover, when all of the plurality of prediction result objects calculated by the adapter 3 are displayed on the communication terminal 7, it may be difficult for the user to find the information that the user wants to see. In particular, the number of summer temperature unevenness prediction result objects 47 and winter temperature unevenness prediction result objects 48 that are determined as non-applied prediction result objects is greater than the bodily sensation prediction result objects 46 that are determined as non-applied prediction result objects. The user may feel annoyed when A plurality of display objects 61 show only prediction result objects determined to be applicable prediction result objects, and do not include information on prediction result objects determined to be non-applied prediction result objects. The air conditioning system 1 reduces the information amount of the plurality of display objects 61 by not including the information of the prediction result object determined as the non-applied prediction result object in the plurality of display objects 61. can be done. By displaying only the application prediction result object on the communication terminal 7, the air conditioning system 1 reduces the number of AI control histories displayed on the communication terminal 7, facilitates browsing of the user's history, and allows the user can improve the convenience of

The operation of updating a plurality of thresholds is performed by a service person, for example, when a user complains about the frequency of AI control. The serviceman connects a predetermined threshold update input device 36 to the control device 27 of the air conditioner 2 and inputs changes in the threshold to the control device 27 via the threshold update input device 36 . The control device 27 updates the plurality of thresholds stored in the threshold storage unit 26 based on the change input via the threshold update input device 36 .

A smaller threshold increases the number of valid guesses that exceed the threshold, and a larger threshold reduces the number of valid guesses that exceed the threshold. Therefore, in the air conditioner 10 of the present embodiment, by changing the threshold value, the case where the prediction result object is determined to be valid is changed, and the frequency with which the setting change corresponding to the prediction result object is executed can be adjusted. can be done. For example, the frequency of setting changes corresponding to a learning model is decreased by increasing the threshold of the learning model and increased by decreasing the threshold of the learning model. In this way, the air conditioning apparatus 10 can bring the frequency of AI control execution close to the frequency desired by the user, thereby improving convenience for the user.

[Effect of air conditioning system 1 of embodiment 1]
The air conditioning system 1 of Embodiment 1 includes an air conditioning device 10 and a communication terminal 7 connected to the air conditioning device 10 so as to be able to transmit information. The air conditioner 10 includes a prediction section 18 , a setting change section 33 and a result transmission section 35 . The prediction unit 18 calculates a plurality of prediction result objects using learning models generated based on the operating state information of the air conditioner 2 . The setting change unit 33 changes the settings corresponding to the non-applied prediction result objects among the plurality of prediction result objects so that the setting changes corresponding to the applicable prediction result objects among the plurality of prediction result objects are executed. The air conditioner 2 is controlled so as not to The result transmitting unit 35 transmits a plurality of display objects 61 corresponding to a plurality of applicable prediction result objects to the communication terminal 7 and does not send information indicating non-applying prediction result objects to the communication terminal 7 . Therefore, the communication terminal 7 can display only a plurality of display objects 61 indicating a plurality of applicable prediction result objects without displaying information about non-useful non-application prediction result objects. The air-conditioning system 1 can reduce the amount of information displayed on the communication terminal 7 by not transmitting the information about the non-application prediction result object to the communication terminal 7 and improve user convenience.

Further, the air conditioner 10 of the air conditioning system 1 of the first embodiment further includes a prediction result storage unit 24 that stores a plurality of prediction result objects, and a history storage unit 25 that stores a plurality of display objects 61. there is The air conditioning system 1 communicates the plurality of display objects 61 without extracting the plurality of applicable prediction result objects from the plurality of prediction result objects each time before transmitting the plurality of display objects to the communication terminal 7. It can be easily transmitted to the terminal 7 .

Also, the air conditioner 10 of the air conditioning system 1 of Example 1 includes the air conditioner 2 and the adapter 3 that is separate from the air conditioner 2 . The prediction unit 18 is provided in the adapter 3 . The result transmission unit 35 transmits a plurality of display objects 61 to the communication terminal 7 via the adapter 3 . A history storage unit 25 is provided in the air conditioner 2 . At this time, the adapter 3 does not need to store a plurality of display objects 61, and the storage capacity of the storage unit 13 provided in the adapter 3 can be reduced. It can be produced inexpensively.

Also, the prediction result storage unit 24 of the air conditioning system 1 of the first embodiment is provided in the air conditioner 2 . At this time, the adapter 3 does not need to store a plurality of prediction result objects, and the storage capacity of the storage unit 13 provided in the adapter 3 can be reduced. can be made to

Further, a display object corresponding to one application prediction result object among the plurality of display objects of the air-conditioning system 1 of the first embodiment includes setting changes corresponding to one application prediction result object and one application prediction result object. and the time when the result object was calculated. The air-conditioning system 1 allows the user to browse the content and time of the setting change that is AI-controlled, thereby improving convenience for the user.

Further, the air conditioner 10 of the air conditioning system 1 of the first embodiment further includes a determination unit 19 that determines whether each of the plurality of prediction result objects is an applied prediction result object or an unapplied prediction result object. there is The air conditioning system 1 can improve convenience for the user without requiring the user to determine whether each of the plurality of prediction result objects is an applied prediction result object or an unapplied prediction result object.

Further, the history storage unit 25 of the air-conditioning system 1 of the first embodiment stores a plurality of display objects 61, and when it is determined that a new prediction result object is an applied prediction result object, a plurality of display objects 61 are displayed. The oldest display object among the display objects 61 is deleted. The history storage unit 25 further adds a new display object representing the new prediction result object to the multiple display objects 61 . By storing a predetermined number of display objects, the air conditioning system 1 does not need to store more display objects than that number. Therefore, the air conditioning system 1 can reduce the storage capacity of the history storage unit 25, and by reducing the storage capacity of the history storage unit 25, the air conditioner 2 can be manufactured at low cost.

Further, when the air conditioning apparatus 10 receives the AI control history request transmitted from the communication terminal 7, the result transmission unit 35 of the air conditioning system 1 of the first embodiment transmits the plurality of display objects 61 to the communication terminal 7. Send. The air conditioning system 1 can display a plurality of display objects 61 on the communication terminal 7 when requested by the user, and can improve user convenience.

By the way, in response to the AI control history request transmitted from the communication terminal 7 being received by the air conditioning apparatus 10, the result transmitting unit 35 of the air conditioning system 1 of the first embodiment transmits the plurality of display objects 61. Although it is transmitted to the communication terminal 7, it may be transmitted at other timings. For example, the result transmission unit 35 may push a plurality of display objects 61 to the communication terminal 7 at predetermined timings. Even in this case, the air conditioning system 1 reduces the amount of information displayed on the communication terminal 7 by not transmitting information about non-applicable prediction result objects to the communication terminal 7, thereby improving user convenience. can be done.

By the way, in the air conditioner 2 of the air conditioning system 1 of the above-described first embodiment, the information about the prediction result object corresponding to the unexecuted setting change is not transmitted to the communication terminal 7, but the information is transmitted to the communication terminal You can send it to 7. The air conditioning system 1 transmits to the communication terminal 7 a prediction result object that is beneficial to the user, even if the prediction result object corresponds to a setting change that has not been executed, so that the user is informed about the prediction result object. Information can be notified, and user convenience can be improved.

By the way, the prediction result storage unit 24 of the air conditioning system 1 of the first embodiment described above is provided in the air conditioner 2 , but may be provided in the adapter 3 . Moreover, although the history storage unit 25 is provided in the air conditioner 2 , it may be provided in the adapter 3 . Even when the prediction result storage unit 24 or the history storage unit 25 is provided in the adapter 3, the air conditioning system 1 does not transmit the information about the non-applied prediction result object to the communication terminal 7, so that the information is displayed on the communication terminal 7. It is possible to reduce the amount of information to be stored and improve user convenience.

By the way, in the air conditioning system 1 of Embodiment 1 described above, the air conditioner 2 and the adapter 3 are formed separately, but the air conditioner 2 and the adapter 3 may be formed integrally. Even if the air conditioner 2 and the adapter 3 are integrally formed, the air conditioning system 1 does not transmit information about the non-applicable prediction result object to the communication terminal 7, so that the information displayed on the communication terminal 7 The amount can be reduced and user convenience can be improved.

By the way, although the threshold storage unit 26 of the air conditioning system 1 of the first embodiment is provided in the air conditioner 2 , it may be provided in the adapter 3 . Even when the adapter 3 is provided with the threshold storage unit 26, the air conditioning system 1 does not include the information of the non-application prediction result object in the plurality of display objects 61 transmitted to the communication terminal 7. It is possible to reduce the amount of information displayed on the screen and improve user convenience.

Although the embodiments have been described above, the embodiments are not limited by the contents described above. In addition, the components described above include those that can be easily assumed by those skilled in the art, those that are substantially the same, and those within the so-called equivalent range. Furthermore, the components described above can be combined as appropriate. Furthermore, at least one of various omissions, replacements, and modifications of components can be made without departing from the gist of the embodiments.

Reference Signs List 1: air conditioning system 2: air conditioner 3: adapter 5: server device 7: communication terminal 10: air conditioning device 18: prediction unit 19: determination unit 24: prediction result storage unit 25: history storage unit 26: threshold storage unit 27: control device 32: determination unit 33: setting change unit 34: threshold setting unit 35: result transmission unit

Claims (8)

an air conditioner;
A communication terminal connected to the air conditioner,
The air conditioner is
a prediction unit that calculates a plurality of prediction result objects for executing a setting change of the air conditioner using a learning model generated based on the operating state information of the air conditioner;
a setting change unit that controls the air conditioner;
transmitting a display object corresponding to an applied prediction result object among the plurality of prediction result objects to the communication terminal, and information about a non-applied prediction result object different from the applied prediction result object among the plurality of prediction result objects; and a result transmission unit that does not transmit to the communication terminal,
The air conditioning system, wherein the communication terminal displays the display object. The air conditioner is
a first storage unit that stores the plurality of prediction result objects;
The air conditioning system according to claim 1, further comprising a second storage section that stores the display object. The air conditioner is
an air conditioner in which the setting change is performed;
an adapter separate from the air conditioner;
The prediction unit is provided in the adapter,
The result transmission unit transmits the display object to the communication terminal via the adapter,
The air conditioning system according to claim 2, wherein the second storage unit is provided in the air conditioner. The air conditioning system according to claim 3, wherein the first storage unit is provided in the air conditioner. The display object is
a setting change corresponding to the application prediction result object;
The air conditioning system according to any one of claims 1 to 4, which indicates the date and time when the application prediction result object was calculated. The air conditioner according to any one of claims 1 to 5, further comprising a determination unit that determines whether each of the plurality of prediction result objects is an applied prediction result object or an unapplied prediction result object. The air conditioning system according to . The second storage unit stores a plurality of display objects, and stores the oldest display object among the plurality of display objects when a new prediction result object is determined to be an applied prediction result object. is deleted, and a new display object indicating the new prediction result object is added to the plurality of display objects. The result transmission unit according to any one of claims 1 to 7, wherein the result transmission unit transmits the display object to the communication terminal when the air conditioner receives a request transmitted from the communication terminal. air conditioning system.

Images