Hot water system

Abstract

The present invention improves the accuracy of predicting total hot water demand in a hot water supply system having a plurality of hot water supply targets and a hot water storage tank that supplies hot water to the hot water supply targets. [Solution] The hot water supply system (1) includes an estimation unit (33) that predicts the total hot water demand based on time-series data of a first index that indicates the calorific value of water used by each of a plurality of hot water supply targets (4).

Description

[Technical field]

[0001] The present disclosure relates to a hot water system. [Background technology]

[0002] Patent Document 1 discloses a hot water supply system having a hot water generating means such as a fuel cell or a gas engine, and a hot water storage tank that utilizes the exhaust heat from the hot water generating means. This hot water supply system learns the demand pattern for hot water and provides an optimal operation plan for the fuel cell. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2006-183947 A Summary of the Invention [Problem to be solved by the invention]

[0004] The hot water supply system in Patent Document 1 learns the demand pattern of hot water mainly for filling and showering. Therefore, when supplying hot water to multiple hot water supply targets, the prediction accuracy of the required heat quantity of hot water is not necessarily high.

[0005] An object of the present disclosure is to improve the accuracy of predicting total hot water demand in a hot water supply system having a plurality of hot water supply targets and a storage tank that supplies hot water to the hot water supply targets. [Means for solving the problem]

[0006] The first aspect is A hot water supply system (10) including a heating device (20) for heating water, a tank (40) for storing the water heated by the heating device (20), and a water circuit (50) through which the water in the tank (40) circulates; a plurality of supply paths (5) connected to a plurality of hot water supply objects (4), respectively, for supplying water from the tank (40); The hot water supply system further includes an estimation unit (33) that estimates a total hot water demand based on time-series data of a first index that indicates a calorific value of water used in each of the plurality of hot water supply targets (4).

[0007] In the first aspect, the total hot water demand is estimated based on time-series data on the calorific value of water used in each hot water supply target (4), so that the prediction accuracy of the total hot water demand can be improved.

[0008] The second aspect is the first aspect, a first learning unit (32) that learns the time-series data in association with the total hot water demand, The estimation unit (33) predicts the total hot water demand based on a result of learning by the first learning unit (32).

[0009] In the second aspect, the total hot water demand can be predicted by the first learning unit (32).

[0010] The third aspect is the second aspect, The first learning unit (32) learns by machine learning.

[0011] In the third aspect, the first learning unit (32) can learn by machine learning. Based on the results of learning by machine learning, the total hot water demand can be predicted.

[0012] The fourth aspect is the first or third aspect, The estimation unit (33) includes an estimation model (M1) generated by machine learning so as to predict the total hot water demand based on the time series data, The total hot water demand is predicted using the estimation model (M1).

[0013] In the fourth aspect, the total hot water demand can be predicted using the estimation model (M1).

[0014] A fifth aspect is any one of the first to fourth aspects, The first index is the temperature and amount of water used in each of the plurality of hot water supply objects (4).

[0015] In the fifth aspect, the first index can be obtained based on the temperature and amount of water used in the hot water supply target (4).

[0016] A sixth aspect is any one of the first to fourth aspects, The first index is the temperature of water used in each of the plurality of hot water supply objects (4) and the pressure of water in each of the supply paths (5).

[0017] In the sixth aspect, the first index can be calculated based on the water pressure and temperature of water flowing through the supply passage (5) connected to each target of hot water supply (4) and the amount of water flowing out of the tank (40).

[0018] A seventh aspect is any one of the first to sixth aspects, The estimation unit (33) predicts the total hot water demand based on the time series data and at least one of the number of the hot water supply objects (4), the type of the hot water supply objects (4), and the specifications of the faucets of the hot water supply objects (4).

[0019] In the seventh aspect, the information for predicting the total hot water demand includes predetermined information for each hot water supply object in addition to time series data on the calorific value of the water used in each hot water supply object (4). In this way, since there are multiple pieces of significant information that can be used to predict the total hot water demand, the prediction accuracy of the total hot water demand can be improved compared to prediction using only the time series data on the calorific value of the water.

[0020] The eighth aspect is any one of the first to seventh aspects, The estimation unit (33) predicts the total hot water demand based on the time series data of the selected hot water supply target (4).

[0021] In the eighth aspect, calculation of hot water supply targets that have little effect on prediction of the total hot water demand can be omitted. By omitting noisy hot water supply targets that lower the prediction accuracy, the prediction accuracy of the total hot water demand can be improved.

[0022] A ninth aspect is any one of the first to eighth aspects, The system further includes a control unit (30) that controls the operation of the heating device (20) based on the total hot water demand.

[0023] In the ninth aspect, the operation control of the heating device (20) is predicted, so that the heating device (20) can be operated with high efficiency.

[0024] A tenth aspect of the present invention relates to the ninth aspect, a second learning unit (35) that learns the total hot water demand in association with an operating state of the heating device (20); The control unit (30) controls the operating state of the heating device (20) based on the result of learning by the second learning unit (35).

[0025] In the tenth aspect, the operation of the heating device (20) can be controlled based on the result of learning by the second learning unit (35).

[0026] The eleventh aspect is the method according to the tenth aspect, The second learning unit (35) learns by machine learning.

[0027] In an eleventh aspect, the second learning unit (35) can learn by machine learning. Based on the learning result by machine learning, the operation control of the heating device (20) can be predicted.

[0028] A twelfth aspect is any one of the ninth and eleventh aspects, the control unit (30) includes an operation prediction model (M2) generated by machine learning so as to predict operation control of the heating device (20) based on the total hot water demand, The operation prediction model (M2) is used to control the operation of the heating device (20).

[0029] In the twelfth aspect, the operation control of the heating device (20) can be predicted using an operation prediction model (M2).

[0030] A thirteenth aspect is any one of the first to twelfth aspects, The heating device (20) is of a heat pump type.

[0031] In the thirteenth aspect, even in a heat pump with a relatively low start-up capacity, the risk of running out of hot water can be reduced and the heating device (20) can be operated with high efficiency. [Brief description of the drawings]

[0032] [Figure 1] FIG. 1 is an overall configuration diagram of a hot water supply system according to an embodiment. [Diagram 2] FIG. 2 is a block diagram of the hot water supply system. [Diagram 3] FIG. 3 is a diagram showing the flow of the refrigerant when the heating device is in operation. [Figure 4] FIG. 4 is a flowchart showing the operation of the hot water supply system. [Diagram 5] FIG. 5 is a diagram showing the relationship between the total hot water demand and the operation control of the heating device. [Figure 6] FIG. 6 is a block diagram of a hot water supply system according to the first modification of the embodiment. [Figure 7] FIG. 7 is a block diagram of a hot water supply system according to the second modification of the embodiment. [Figure 8] FIG. 8 is a flowchart showing the operation of the hot water supply system. [Figure 9] FIG. 9 is a block diagram of a hot water supply system according to a third modification of the embodiment. [Figure 10] FIG. 10 is a block diagram of a hot water supply system according to the fourth modification of the embodiment. [Figure 11] FIG. 11 is a flowchart showing the operation of the hot water supply system. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0033] Hereinafter, embodiments will be described with reference to the drawings. Note that the following embodiments are essentially preferred examples and are not intended to limit the scope of the present invention, its applications, or its uses.

[0034] <<Embodiment>> As shown in Fig. 1, the present disclosure is a hot water supply system (1). The hot water supply system (1) heats water supplied from a water source (1) and stores the heated water in a tank (40). The hot water stored in the tank (40) is supplied to a plurality of hot water supply objects (4). The water source includes a water supply system. The hot water supply objects (4) include a bathtub, a shower, a faucet, and the like.

[0035] The hot water supply system (1) includes a heating device (20), a tank (40), a water circuit (50), a supply path (5), a first pipe (6), a pressure sensor (60), temperature sensors (61, 63), and a control unit (30).

[0036] <Heating device> The heating device (20) of this embodiment is of a heat pump type. The heating device (20) generates heat for heating water. The heating device (20) is of a vapor compression type. The heating device (20) has a refrigerant circuit (21). The refrigerant circuit (21) is filled with refrigerant. The refrigerant circuit (21) includes a compressor (22), a heat source heat exchanger (23), an expansion valve (24), and a utilization heat exchanger (25).

[0037] The compressor (22) compresses the sucked refrigerant and discharges the compressed refrigerant.

[0038] The heat-source heat exchanger (23) is an air-cooled heat exchanger. The heat-source heat exchanger (23) is disposed outside the room. The heating device (20) has an outdoor fan (27). The outdoor fan (27) is disposed near the heat-source heat exchanger (23). The heat-source heat exchanger (23) exchanges heat between the air delivered by the outdoor fan (27) and the refrigerant.

[0039] The expansion valve (24) is a pressure reducing mechanism for reducing the pressure of the refrigerant. The expansion valve (24) is provided between the liquid end of the utilization heat exchanger (25) and the liquid end of the heat source heat exchanger (23). The pressure reducing mechanism is not limited to the expansion valve, and may be a capillary tube, an expander, or the like. The expander recovers the energy of the refrigerant as power.

[0040] The utilization heat exchanger (25) is a liquid-cooled heat exchanger. The utilization heat exchanger (25) has a first flow path (25a) and a second flow path (25b). The second flow path (25b) is connected to the refrigerant circuit (21). The first flow path (25a) is connected to the water circuit (50). The utilization heat exchanger (25) exchanges heat between the water flowing through the first flow path (25a) and the refrigerant flowing through the second flow path (25b).

[0041] In the utilization heat exchanger (25), a first flow path (25a) is formed so as to extend along the second flow path (25b). In this embodiment, in a heating operation, the direction of the refrigerant flowing through the second flow path (25b) is substantially opposite to the direction of the water flowing through the first flow path (25a). In other words, the utilization heat exchanger (25) in the heating operation functions as a counterflow heat exchanger.

[0042] <Tank and water circuit> The tank (40) is a container for storing water. The tank (40) is formed in a vertically long cylindrical shape. The tank (40) has a cylindrical body (41), a bottom (42) that closes the lower end of the body (41), and a top (43) that closes the upper end of the body (41).

[0043] The water in the tank (40) circulates through the water circuit (50). The first flow path (25a) of the utilization heat exchanger (25) is connected to the water circuit (50). The water circuit (50) includes an upstream flow path (51) and a downstream flow path (52).

[0044] An inflow end of the upstream flow path (51) is connected to the bottom (42) of the tank (40). An inflow end of the upstream flow path (51) is connected to the bottom (42) of the tank (40). An outflow end of the upstream flow path (51) is connected to the inflow end of the first flow path (25a).

[0045] The inflow end of the downstream flow path (52) is connected to the outflow end of the first flow path (25a) and the outflow end of the downstream flow path (52) is connected to the top (43) of the tank (40).

[0046] The water circuit (50) includes a water pump (53). The water pump (53) circulates water in the water circuit (50). The water pump (53) transports water from the tank (40) to the first flow path (25a) of the utilization heat exchanger (25). The water pump (53) further transports water to the first flow path (25a) and sends it to the tank (40).

[0047] <First piping and supply line> An inflow end of the first pipe (6) is connected to the tank (40). An outflow end of the first pipe (6) is connected to each inflow end of the plurality of supply paths (5). An outflow end of each supply path (5) is connected to each hot water supply target (4).

[0048] <Pressure Sensor> A pressure sensor (60) is connected to each supply path (5). The pressure sensor (60) detects the pressure of water in the supply path (5). In other words, the pressure sensor (60) detects the pressure of water supplied to each hot water supply target (4).

[0049] <Temperature Sensor> The hot water supply system (1) has a first temperature sensor (61) and a second temperature sensor (63). The first temperature sensor (61) is provided in each hot water supply object (4). The first temperature sensor (61) detects the temperature of water used in the hot water supply object (4). The second temperature sensor (63) is provided at the inlet end of the first pipe (6). The second temperature sensor (63) detects the temperature of water flowing from the tank (40) into the first pipe (6).

[0050] <Flow sensor> A flow rate sensor (62) is provided at the inflow end of the first pipe (6). The flow rate sensor (62) detects the amount of water that has flowed from the tank (40) into the first pipe (6).

[0051] <Control Unit> The control unit (30) shown in Fig. 2 has a microcomputer and a memory device (specifically, a semiconductor memory) that stores software for operating the microcomputer. The control unit (30) is connected to various devices and sensors of the hot water supply apparatus (10) via wires or wirelessly. The control unit (30) controls the heating device (20) and the devices in the water circuit (50). The devices in the water circuit (50) include a water pump (53).

[0052] The control unit (30) includes a memory unit (31), a first learning unit (32), an estimation unit (33), and a second learning unit (35).

[0053] The memory unit (31) stores time series data of the first index. The first index indicates the calorific value of water used in each hot water supply target (4). Specifically, the first index is the temperature of water used in each hot water supply target (4) and the water pressure of each supply path (5). Hereinafter, the time series data of the first index is referred to as first time series data. The first time series data is the time series data of the present disclosure.

[0054] Here, the total hot water demand refers to the calorific value of water used by the entire hot water supply apparatus (10) in a given time period, and corresponds to the sum of the calorific values ​​of water used by each hot water supply target (4).

[0055] The memory unit (31) stores the actual total hot water demand as time-series data. The actual total hot water demand is determined by measuring the calorific value of the water flowing out from the tank (40) to the first pipe (6) using a detection means. Specifically, the total hot water demand is determined based on values ​​of the flow rate sensor (62) and the second temperature sensor (63). The time-series data of the total hot water demand stored in the memory unit (31) is referred to as second time-series data.

[0056] The first learning unit (32) learns by associating the first time series data for a predetermined period stored in the storage unit (31) with the second time series data for the same time period as the first time series data. The first learning unit (32) learns by machine learning.

[0057] The estimation unit (33) predicts the hot water demand based on the first time series data. Specifically, the estimation unit (33) predicts the total hot water demand based on the learning result learned by the first learning unit (32). More specifically, the estimation unit (33) predicts the total hot water demand using a learned estimation model (M1) that is learned by machine learning by associating the first time series data and the second time series data stored in the storage unit (31). The estimation unit (33) predicts the total hot water demand for one day, for example, the next day, which is a predetermined time. Here, the total hot water demand predicted by the estimation unit (33) may be time series data that changes every certain time (for example, every hour) as shown in FIG. 5.

[0058] The estimation model (M1) is included in the estimation unit (33). The estimation model (M1) is generated to predict the total hot water demand based on the first time series data. The estimation model (M1) is constructed as a multi-layer neural network that has acquired a measurement capability through machine learning. The estimation model (M1) of this embodiment is generated by "supervised learning". The neural network for generating the estimation model (M1) learns using training data and a discriminant function. The training data is a set of pairs of input data and teacher data corresponding to the input data.

[0059] The input data is first time series data for a predetermined period stored in the memory unit (31). Specifically, the input data is time series data of the water pressure in the supply line (5) for a predetermined period, and time series data of the temperature of water used in the hot water supply target (4) connected to the supply line (5). The teacher data is second time series data for the same period as the input data. By having the neural network perform "supervised learning" using the learning data as described above, a learned estimation model (M1) is generated as a learning result.

[0060] In this manner, the estimation unit (33) predicts the total hot water demand by using the trained estimation model (M1). The estimation unit (33) outputs the total hot water demand by inputting the first time-series data for a predetermined period (e.g., one week until the previous day) stored in the memory unit (31) into the trained estimation model (M1). In this manner, the estimation unit (33) predicts the total hot water demand.

[0061] The second learning unit (35) learns the total hot water demand and the operating state of the heating device (20) in association with each other. The second learning unit (35) learns by machine learning. Based on the result of learning by the second learning unit (35), the control unit (30) controls the operation of the heating device (20). Specifically, the control unit (30) controls the operation of the heating device (20) by using the operation prediction model (M2).

[0062] The control unit (30) includes an operation prediction model (M2). The operation prediction model (M2) is generated by machine learning so as to predict control of the operation of the heating device (20) based on the total hot water demand. The control unit (30) uses the trained operation prediction model (M2) to control the operation of the heating device (20).

[0063] The operation prediction model (M2) is generated by "reinforcement learning". Specifically, the second learning unit (35) sets the reward to the electricity charge for one day and the state variable to the operation state of the heating device (20). Here, the operation state of the heating device (20) refers to, for example, the ON state or the OFF state of the heating device (20). The second learning unit (35) inputs the second time-series data for a predetermined period as input data to the operation prediction model (M2). In this way, the second learning unit (35) learns so as to minimize the electricity charge for the operation of the heating device (20) for one day. The total hot water demand predicted by the estimation unit (33) is input to the learned operation prediction model (M2) generated in this way, and the operation of the heating device (20) is controlled so as to minimize the power.

[0064] -Heating operation- As shown in Fig. 3, the control unit (30) controls the heating device (20) to perform a heating operation. Specifically, the control unit (30) operates the compressor (22) and the outdoor fan (27). The control unit (30) appropriately adjusts the opening degree of the expansion valve (24). The control unit (30) operates the water pump (53).

[0065] The refrigerant compressed by the compressor (22) flows through the second flow path (25b) of the utilization heat exchanger (25). In the utilization heat exchanger (25), the refrigerant in the second flow path (25b) dissipates heat to the water in the first flow path (25a). The refrigerant that has dissipated heat or condensed in the second flow path (25b) is reduced in pressure by the expansion valve (24) and then flows through the heat source heat exchanger (23). In the heat source heat exchanger (23), the refrigerant absorbs heat from the outdoor air and evaporates. The refrigerant that has evaporated in the heat source heat exchanger (23) is sucked into the compressor (22).

[0066] In the water circuit (50), water in the tank (40) flows out into the upstream flow path (51). The water in the upstream flow path (51) flows through the first flow path (25a) of the utilization heat exchanger (25). The water in the first flow path (25a) is heated by the refrigerant of the heating device (20).

[0067] The heated water in the tank (40) flows through a predetermined supply passage (5) via a first pipe (6). The water flowing through the supply passage (5) flows out from a hot water supply target (4) connected to the supply passage (5) to the outside.

[0068] <Operation of hot water supply system> Next, an example of the operation of the hot water supply system (1) of this embodiment will be described with reference to FIG.

[0069] In step ST1, the control unit (30) inputs the first time series data for one week up to the previous day into the trained estimation model (M1).

[0070] In step ST2, the control unit (30) outputs the total hot water demand for the next day (future) from the trained estimation model (M1) in step ST2. The total hot water demand output here is time-series data that changes every hour in the next day.

[0071] In step ST3, the control unit (30) inputs the total hot water demand for the next day output in step ST2 into the learned operation prediction model (M2).

[0072] In step ST4, the control unit (30) outputs operation control for the heating device (20) for the next day from the learned operation prediction model (M2). The operation control output here is, for example, an operation plan for turning the heating device (20) ON or OFF every hour for the next day, as shown in FIG. 5. Based on this operation plan, the control unit (30) controls the heating operation of the heating device (20). The heating device (20) boils water in the tank (40) so that the amount of hot water required in each time period can be supplied to the hot water supply target (4) based on the predicted total hot water demand.

[0073] In step ST5, the control unit (30) controls the heating operation of the heating device (20) based on the operation control of the heating device (20) output in step ST4. For example, in the operation plan of FIG. 5, the control unit (30) controls the heating operation of the heating device (20) to boil the water in the tank (40) so that the required amount of hot water can be supplied to the hot water supply target (4) from 13:00 to 14:00. On the other hand, the control unit (30) controls the heating device (20) not to perform the heating operation from 14:00 to 15:00. This is because, in this operation plan, it is determined that it is more efficient not to perform the heating operation from 14:00 to 15:00 based on the required amount of hot water supply from 14:00 to 15:00 and the remaining amount of hot water stored in the tank (40). Thereafter, the control unit (30) controls the heating operation of the heating device (20) to boil the water in the tank (40) according to the required amount of hot water supply in each time period after 15:00.

[0074] Effects of the embodiment The hot water supply system (1) of this embodiment includes an estimation unit (33) that predicts a total hot water demand based on time series data (first time series data) of a first index indicating the calorific value of water used by each of the hot water supply objects (4). As a result, the total hot water demand is estimated based on the first index of each hot water supply object (4), and therefore the prediction accuracy of the total hot water demand can be improved compared to the case where the prediction accuracy is based only on hot water supply objects (4) that require a relatively large amount of hot water (e.g., a shower or a bathtub).

[0075] In addition, even in a house where multiple households reside in one building, such as an apartment building, the prediction accuracy of the total hot water demand of the house can be improved. To be more specific, in an apartment building, the amount of hot water used by each hot water supply object (4) varies from household to household. Therefore, for example, if the amount of hot water used for the shower or bathtub by household A is greater than the amount of hot water used by other households, the predicted value of the amount of hot water supply for household A may affect the predicted value of the entire apartment building, which may result in a lower prediction accuracy of the amount of hot water supply for other households. However, according to the hot water supply system (1) of this embodiment, the total hot water demand of the entire apartment building is estimated based on the time series data of the first index of each hot water supply object (4) of each household in the apartment building. This makes it possible to improve the prediction accuracy of the total hot water demand of the entire apartment building, and therefore the prediction accuracy of the total hot water demand of each household can also be improved.

[0076] The hot water supply system (1) of the embodiment has a first learning unit (32) that learns the first time series data and the second time series data (total hot water demand) in association with each other. The estimation unit (33) predicts the total hot water demand based on the learning result of the first learning unit (32). This makes it possible to predict the total hot water demand based on the learning result of the first learning unit (32).

[0077] In the hot water supply system (1) of the embodiment, the first learning unit (32) learns by machine learning, which makes it possible to predict the total hot water demand based on the learning result obtained by machine learning.

[0078] In the hot water supply system (1) of the embodiment, the estimation unit (33) includes an estimation model (M1) generated by machine learning to predict the total hot water demand based on the first time series data, and predicts the total hot water demand using the estimation model (M1). In this example, the trained estimation model (M1) is generated based on the first time series data for a predetermined period by supervised learning. By using such a trained estimation model (M1), it is possible to reliably improve the prediction accuracy of the total hot water demand.

[0079] In addition, the trained estimation model (M1) is updated by sequential learning. Therefore, the more frequently the hot water supply system (1) is used, the more frequently the trained estimation model (M1) is updated. As a result, the prediction accuracy of the total hot water demand output from the trained estimation model (M1) can be improved.

[0080] In the hot water supply system (1) of the embodiment, the first index is the temperature of water used in each of the multiple hot water supply objects (4) and the water pressure in each of the supply paths (5). By using time series data of the temperature of water used in the hot water supply objects (4) and time series data of the water pressure in the supply path (5) connected to the hot water supply objects (4) as input data and using the second time series data as teacher data, a learned estimation model (M1) can be generated by supervised learning.

[0081] The hot water supply system (1) of the embodiment includes a control unit (30) that controls the operation of the heating device (20) based on the total hot water demand. This allows the heating device (20) to operate in accordance with the amount of heat of water required by the hot water supply target (4) in each time slot of the next day, for example.

[0082] The hot water supply system (1) of the embodiment has a second learning unit (35) that learns the total hot water demand and the operating state of the heating device (20) in association with each other, and the control unit (30) controls the operation of the heating device (20) based on the learning result by the second learning unit (35). This makes it possible to control the operation of the heating device (20) based on the learning result by the second learning unit (35).

[0083] In the hot water supply system (1) of the embodiment, the second learning unit (35) learns by machine learning, which makes it possible to predict control of the operation of the heating device (20) based on the learning result obtained by machine learning.

[0084] In the hot water supply system (1) of the embodiment, the control unit (30) includes an operation prediction model (M2) generated by machine learning to predict operation control of the heating device (20) based on the total hot water demand, and controls the operation of the heating device (20) using the operation prediction model (M2). In this example, by using the operation prediction model (M2) learned by machine learning, the heating device (20) can be operated with high efficiency and the operation of the heating device (20) can be controlled so that the daily electricity bill is minimized compared to the past electricity bill. Since the operation of the heating device (20) can be controlled according to the total hot water demand, it is possible to prevent the tank (40) from running out of hot water during hot water supply.

[0085] In addition, the learned operation prediction model (M2) is updated by sequential learning. Therefore, the more frequently the hot water supply system (1) is used, the more frequently the learned operation prediction model (M2) is updated, and as a result, it is possible to predict operation control that reduces electricity costs. Therefore, the more frequently the hot water supply system (1) of this example is used, the lower the daily electricity costs can be kept.

[0086] In the hot water supply system (1) of the embodiment, the heating device (20) is a heat pump type, which reduces the risk of running out of hot water and enables highly efficient operation even in a heat pump with a relatively low start-up capacity.

[0087] Variation 1 In the hot water supply system (1) of this example, the estimation unit (33) includes a trained estimation model (M1) that has been trained in advance. Configurations that differ from the above embodiment will be described below.

[0088] 6, the control unit (30) does not have a first learning unit (32). The estimation model (M1) of this example is generated in advance before the hot water supply system (1) is used by a user (before the hot water supply system (1) is shipped) so as to predict the total hot water demand based on the first time-series data.

[0089] The estimation model (M1) in this example is also generated by "supervised learning." Predetermined learning data stored in a data server is input to the estimation model (M1). Specifically, the input data includes nationwide user information (number of people in a family, age, sex, and residential area), information on hot water supply objects (4) owned by each user (number, type, and faucet information of hot water supply objects), and time series data on the water pressure of the supply line (5) connected to the hot water supply object (4) of the hot water supply system (1) owned by each user, and time series data on the temperature of the water used in the hot water supply object. The teacher data is second time series data stored in the data server. Time series data from the past few years is stored in the data server.

[0090] Using such learning data, the neural network is subjected to “supervised learning.” As a result of the learning, a trained estimation model (M1) of this example is generated. The estimation unit (33) uses the trained estimation model (M1) to predict the total hot water demand.

[0091] User data is stored in the storage unit 31. The user data includes information such as family structure (number of family members, age, sex, etc.) and residential area.

[0092] In this way, in the hot water supply system (1) of this example, the total hot water demand is predicted based on the first time series data of each hot water supply target, so that the prediction accuracy of the total hot water demand can be improved. In particular, the trained estimation model (M1) of this example is generated using the first time series data of users nationwide stored in a data server. Therefore, by inputting information such as the user's residential area and family composition and the first time series data into the trained estimation model (M1), it is possible to obtain a total hot water demand with higher prediction accuracy according to the user than when only the first time series data is input.

[0093] In addition, since the hot water supply system (1) is already equipped with the trained estimation model (M1) at the time of shipment, the user can use the hot water supply system (1) with high prediction accuracy of the total hot water demand from the start of use.

[0094] Variation 2 The hot water supply system (1) of this embodiment predicts the total hot water demand by using a predetermined logical expression.

[0095] Here, the hot water demand refers to the calorific value of water used for each hot water supply target (4) in a given period of time. The hot water demand corresponds to the total calorific value of water used by each hot water supply target (4).

[0096] The hot water supply system (1) of this example predicts the hot water demand of each hot water supply object (4) based on time series data of the calorific value of water used in each hot water supply object (4) for a predetermined period. After that, the hot water supply system (1) predicts the total hot water demand based on the predicted hot water demand of each hot water supply object (4). Below, a description will be given of configurations that differ from the above embodiment and the above modification example 1.

[0097] 7, in the hot water supply system (1) of this example, the control unit (30) has a calculation unit (34). The control unit (30) of this example does not have an estimation model (M1). The first learning unit (32) does not perform machine learning.

[0098] The calculation section (34) determines time series data of the hot water demand of each hot water supply target (4) for a predetermined period, based on the first time series data stored in the storage section (31).

[0099] The estimation unit (33) predicts the hot water demand of each hot water supply object (4) based on the hot water demand of each hot water supply object (4) for a predetermined period calculated by the calculation unit (34). The prediction of the hot water demand of each hot water supply object (4) is performed by using a predetermined logical formula and a first coefficient. The total hot water demand predicted by the estimation unit (33) may be time-series data that changes every certain hour (e.g., every hour) of the next day. The estimation unit (33) predicts the total hot water demand based on the predicted hot water demand of each hot water supply object (4). The prediction of the total hot water demand is performed by using a predetermined logical formula and a second coefficient. The first coefficient and the second coefficient are input to the estimation unit (33) by the first learning unit (32).

[0100] The first learning unit (32) adjusts a first coefficient so as to reduce a residual between the hot water demand for a specified period predicted by the estimation unit (33) for each hot water supply target (4) and the hot water demand actually used in the specified period. The first learning unit (32) inputs the adjusted first coefficient to the estimation unit (33). The first learning unit (32) adjusts a second coefficient so as to reduce a residual between the total hot water demand for a specified period predicted by the estimation unit (33) and the total hot water demand actually used in the specified period. The first learning unit (32) inputs the second coefficient to the estimation unit (33).

[0101] <Operation of hot water supply system> Next, an example of the operation of the hot water supply system (1) of this embodiment will be described with reference to Fig. 8. Steps ST24 to ST26 are the same as steps ST3 to ST5 of the above embodiment, and therefore description thereof will be omitted.

[0102] In step ST21, the control unit (30) determines time series data of the hot water demand of each hot water supply target (4) for a predetermined period (e.g., one week up to the previous day) based on the first time series data for the period stored in the memory unit (31).

[0103] In step ST22, the control unit (30) predicts the hot water demand of each hot water supply target (4) for the next day from the time-series data of the hot water demand obtained in step ST21, using a predetermined logical expression and a first coefficient.

[0104] In step ST23, the control unit (30) predicts the total hot water supply demand for the next day from the hot water supply demand for each hot water supply target (4) for the next day predicted in step ST22.

[0105] In step ST27, the control unit (30) adjusts the first coefficient so as to reduce the prediction error based on the residual between the hot water demand of each hot water supply target (4) predicted in step ST22 and the hot water demand of each hot water supply target (4) actually used. The control unit (30) inputs the adjusted first coefficient to the estimation unit (33).

[0106] In step ST28, the control unit (30) adjusts a predetermined second coefficient based on the residual between the total hot water demand predicted in step ST23 and the total hot water demand actually used so as to reduce the prediction error. The control unit (30) inputs the adjusted second coefficient to the estimation unit (33).

[0107] According to this example, by repeatedly adjusting the input of the first coefficient for the hot water demand prediction for each hot water supply target (4), it is possible to reduce the residual between the predicted hot water demand for each hot water supply target (4) and the hot water demand actually used for each hot water supply target (4). Furthermore, by repeatedly adjusting the input of the second coefficient, it is possible to reduce the residual between the predicted hot water demand and the hot water demand actually used, thereby improving the prediction accuracy of the total hot water demand.

[0108] Variation 3 In the hot water supply system (1) of this example, the control unit (30) includes an operation prediction model (M2) that has been trained in advance. Configurations that are different from the above embodiment and the above modified example will be described below.

[0109] 9, the control unit (30) of this example does not have a second learning unit (35). The operation prediction model (M2) of this example is generated before the hot water supply system (1) is used by a user (before the hot water supply system (1) is shipped) so as to predict operation control based on the total hot water demand.

[0110] The operation prediction model (M2) of this example is also generated by "reinforcement learning." Input data (user information from across the country (number of people in a family, age, sex, residential area), information on hot water supply objects (4) owned by each user (number, type, and faucet information on hot water supply objects), and total hot water demand of each user, etc.) stored in the data server of the above embodiment are input to the operation prediction model (M2). By setting the reward as the daily electricity charge and the state variable as the operating state of the heating device (20), a trained operation prediction model (M2) is generated that has learned to minimize the electricity charge for the daily operation of the heating device (20). By inputting the total hot water demand into the trained operation prediction model (M2) in this way, operation control of the heating device (20) that minimizes power is output.

[0111] In this way, in this example, by using the operation prediction model (M2) learned by machine learning, the heating device (20) can be controlled so that the daily electricity bill is minimized compared to the past electricity bill. In particular, the learned operation prediction model (M2) in this example is generated using the total hot water demand of users nationwide stored in a data server. Therefore, by inputting information such as the user's residential area and family composition and the total hot water demand into the learned estimation model (M1), it is possible to predict operation control of the heating device (20) that reduces the electricity bill according to the user compared to the case where only the total hot water demand is input.

[0112] In addition, since the hot water supply system (1) is equipped with an operation prediction model (M2) that has already been trained at the time of shipment, the user can use the hot water supply system (1) that can control the operation of the heating device (20) so as to reduce the daily electricity bill from the start of use.

[0113] Variation 4 The hot water supply system (1) of this embodiment predicts operation control of the heating device (20) from the total hot water demand based on a predetermined control. Configurations different from the above embodiment and modified examples will be described below.

[0114] 10, in the hot water supply system (1) of this embodiment, the control unit (30) has a hot water shortage determination unit (36). The control unit (30) does not have an operation prediction model (M2).

[0115] The hot water shortage determiner (36) determines that there is no water in the tank (40) (hot water shortage) if the detection value of the flow rate sensor (62) becomes zero while the heating device (20) is ON.

[0116] The second learning unit (35) learns operational control of the heating device (20) that reduces the risk of running out of hot water, based on the total hot water demand. Specifically, if there is a time when hot water runs out as a result of controlling the operational state of the heating device (20) based on the total hot water demand predicted by the estimation unit (33), the second learning unit (35) corrects the operational control of the heating device (20) performed on that day. With such feedback, the second learning unit (35) corrects the operational control of the heating device (20) and predicts the operational control of the heating device (20) for the next day. <Operation of hot water supply system> Next, an example of the operation of the hot water supply system (1) of this embodiment will be described with reference to Fig. 11. Steps ST41 to ST42 are the same as steps ST1 to ST2 of the above embodiment, and therefore description thereof will be omitted.

[0117] In step ST43, the control unit (30) predicts the operation control of the heating device (20) based on the total hot water demand estimated by the estimating unit (33).

[0118] In step ST44, the control unit (30) controls the heating device (20) so as to execute the operation control predicted in step ST43.

[0119] In step ST45, the control unit (30) determines whether or not a hot water shortage has occurred when supplying hot water to the hot water supply target (4). If a hot water shortage has occurred (YES in step ST45), the control unit (30) controls the heating device (20) to boil the water supplied to the tank (40) (step ST44). If a hot water shortage has not occurred (NO in step ST45), step ST46 is executed.

[0120] In step ST46, the control unit (30) determines whether or not the operation of the heating device (20) for the day has ended. If the operation has ended (YES in step ST46), step ST47 is executed. If the operation has not ended (NO in step ST46), step ST44 is executed again.

[0121] In step ST47, the control unit (30) determines whether or not there has been a hot water shortage during a day's operation of the heating device (20). If there has been a hot water shortage (YES in step ST47), step ST48 is executed.

[0122] In step ST48, the control section (30) modifies the operation control based on the time when the hot water ran out, the amount of hot water that was boiled, the boiling temperature, and the like.

[0123] According to this example, the control unit (30) corrects the operation control of the heating device (20) whenever a hot water shortage occurs. By predicting the operation control based on such corrections, it is possible to reduce the risk of hot water shortage the next day.

[0124] Other Embodiments The above embodiment and modified examples may be configured as follows.

[0125] The first index may be the temperature of water used in the object of hot water supply (4) and the amount of water in the supply passage (5) connected to the object of hot water supply (4). In this case, a flow rate sensor (not shown) for measuring the flow rate of water is provided in each supply passage (5). The flow rate sensor is connected to the control unit (30) by wire or wirelessly.

[0126] The first index may be the calorific value of water used in each hot water supply target (4). Time series data of the calorific value of water used in each hot water supply target (4) may be obtained based on time series data of the temperature of the water used in the hot water supply target (4) and time series data of the water pressure in the supply line (5) connected to the hot water supply target (4). Time series data of the calorific value of water used in each hot water supply target (4) may be obtained based on time series data of the temperature of the water used in the hot water supply target (4) and time series data of the water volume in the supply line (5) connected to the hot water supply target (4). Time series data of the calorific value of water used in each hot water supply target (4) is input as input data to the estimation model (M1).

[0127] In the reinforcement learning of the operation prediction model (M2), the operating state of the heating device (20) set as a state variable may be the rotation speed of the compressor (22), the condensing temperature of the utilization heat exchanger (25), the opening of the expansion valve (24), etc. For example, the control unit (30) increases the rotation speed of the compressor (22) to heat the water in the tank (40), or reduces the opening of the expansion valve (24) to suppress an increase in the condensing temperature of the utilization heat exchanger (25), thereby suppressing heating of the water in the tank (40).

[0128] The hot water supply system (1) may include a predetermined sensor (not shown) capable of directly detecting the calorific value of the water used in each of the hot water supply objects (4). In this case, time-series data of the calorific value of the water used in each of the hot water supply objects (4), detected by the sensor, is input to the estimation model (M1) without using the calculation unit (34).

[0129] The input data input to the estimation model (M1) may include at least one of the number of hot water supply objects (4), the type of the hot water supply objects (4), and the specifications of the faucet of the hot water supply objects (4). In this way, by using information on each hot water supply object (4) as input data in addition to the first time series data, the total hot water demand can be predicted based on the hot water supply object (4) whose type is specified and the first time series data of the hot water supply object (4). For example, if the hot water supply objects (4) are a bathtub and a shower, the bathtub (filling with water) and the shower are often used at the same time or in relatively close time periods (such as before and during bathing), so the first learning unit (32) can learn the total hot water demand based on the first time series data of the bathtub, taking into account the hot water demand of the shower, and can also learn the total hot water demand based on the first time series data of the shower, taking into account the hot water demand of the bathtub. By using the learned estimation model (M1) learned in this way, the risk of running out of hot water can be reduced.

[0130] The total hot water demand predicted by the estimation unit (33) may be the amount of heat of water used in the hot water supply apparatus (10) only during a predetermined time period (for example, one hour from a certain time on the next day).

[0131] In the above embodiment, the second time-series data may be obtained based on a change in the hot water temperature of the tank (40). In this case, a hot water temperature sensor (not shown) is provided in the tank (40). The hot water temperature sensor is connected to the control unit (30) by wire or wirelessly. For example, when the temperature value of the hot water temperature sensor decreases, it is understood that hot water is supplied from the tank (40) and cold water is supplied to the tank (40). The calorific value of the water flowing out of the tank (40) can be obtained based on how much the temperature of the hot water temperature sensor has decreased.

[0132] The estimation unit (33) may estimate the total hot water demand based on the hot water consumption of each of the selected hot water supply targets (4), thereby improving the prediction accuracy of the total hot water demand by excluding noisy hot water supply targets that may reduce the prediction accuracy.

[0133] The estimation model (M1) can also be generated by "unsupervised learning". In this case, the neural network repeats a learning operation of grouping multiple input data into multiple categories by clustering so that similar input data (total hot water demand) are classified in the same category. This makes it possible to generate a trained estimation model (M1) without using teacher data. The estimation model (M1) may also be generated by "reinforcement learning".

[0134] The storage unit (31) does not have to be provided in the control unit (30). The storage unit (31) may be stored in a predetermined server that can communicate with the control unit (30).

[0135] In the above-mentioned second embodiment and the modified example, the driving prediction model (M2) can also be generated by "supervised learning" or "unsupervised learning".

[0136] The first learning unit (32) may learn using a learning method other than the above-described embodiment and modified examples. The second learning unit (35) may learn using a learning method other than the above-described embodiment and modified examples.

[0137] Although the embodiments and modifications have been described above, it will be understood that various modifications of form and details are possible without departing from the spirit and scope of the claims. Furthermore, the above embodiments and modifications may be appropriately combined or substituted as long as the functions of the subject of the present disclosure are not impaired. The descriptions "first" and "second" described above are used to distinguish the words to which these descriptions are attached, and do not limit the number or order of the words. [Industrial Applicability]

[0138] INDUSTRIAL APPLICABILITY As described above, the present disclosure is useful for hot water supply systems. [Explanation of symbols]

[0139] M1 Estimation Model M2 Driving Prediction Model 4. Hot water supply target 5 Supply route 10. Hot water supply equipment 20 Heating device 30 Control section 32 First Learning Section 33 Estimation part 35 Second Learning Section 40 Tank 50 water circuit

Claims

1. a hot water supply system (10) including a heating device (20) for heating water, a tank (40) for storing the water heated by the heating device (20), and a water circuit (50) through which the water in the tank (40) circulates; a plurality of supply paths (5) connected to a plurality of hot water supply objects (4), respectively, for supplying water from the tank (40); an estimation unit (33) that predicts a total hot water demand indicating a hot water demand for all of the hot water supply objects (4) in a predetermined time period based on time series data for each of the hot water supply objects (4) of a first index indicating a calorific value of water used in each of the hot water supply objects (4); a first learning unit (32) that performs machine learning by associating the time-series data with the total hot water demand, The estimation unit (33) predicts the total hot water demand based on a result of learning by the first learning unit (32). A hot water supply system comprising:

2. In the hot water supply system of claim 1, the estimation unit (33) includes an estimation model (M1) generated by machine learning so as to predict the total hot water demand based on the time-series data, The total hot water demand is predicted using the estimation model (M1). A hot water supply system comprising:

3. In any one of claims 1 and 2, The first index is a temperature and an amount of water used in each of the plurality of hot water supply objects (4). A hot water supply system comprising:

4. In any one of the hot water supply systems according to claims 1 to 3, The first indicator is a temperature of water used in each of the plurality of hot water supply objects (4) and a pressure of water in each of the supply paths (5). A hot water supply system comprising:

5. In any one of the hot water supply systems according to claims 1 to 4, The estimation unit (33) predicts the total hot water demand based on the time series data and at least one of the number of the hot water supply objects (4), the type of the hot water supply objects (4), and the specifications of the faucets of the hot water supply objects (4). A hot water supply system comprising:

6. In any one of the hot water supply systems according to claims 1 to 5, a control unit (30) that controls the operation of the heating device (20) based on the total hot water demand predicted by the estimation unit (33). A hot water supply system comprising:

7. The hot water supply system according to claim 6, a second learning unit (35) that learns the total hot water demand in association with an operating state of the heating device (20); The control unit (30) controls the operation of the heating device (20) based on the result of learning by the second learning unit (35). A hot water supply system comprising:

8. The hot water supply system according to claim 7, The second learning unit (35) learns by machine learning. A hot water supply system comprising:

9. In the hot water supply system according to claim 6 or 8, the control unit (30) includes an operation prediction model (M2) generated by machine learning so as to predict operation control of the heating device (20) based on the total hot water demand, The operation prediction model (M2) is used to control the operation of the heating device (20). A hot water supply system comprising:

10. In any one of the hot water supply systems according to claims 1 to 9, The heating device (20) is a heat pump type. A hot water supply system comprising:

Images