Power consumption prediction system, power consumption prediction method, learning device, and learning method

Abstract

This technology provides a way to more accurately predict future power consumption in equipment. [Solution] The equipment comprises an electric heat source and a combustion heat source. The power consumption prediction system comprises an acquisition unit that acquires energy consumption information relating to the energy consumption of the equipment, and a calculation unit that calculates the predicted power consumption for a first period after the present time based on the energy consumption information. The energy consumption information includes at least the power consumption of the equipment in a second period corresponding to the first period and prior to the present time, the amount of gas consumed by the combustion heat source in the second period, and climate information relating to the climate near the equipment in a third period including the first period. The calculation unit calculates the predicted power consumption by executing calculation processing of a trained learner that has undergone machine learning to calculate the predicted power consumption for the first period based on the energy consumption information.

Description

Technical Field

[0001] The technology disclosed in the present specification relates to a power consumption prediction system, a power consumption prediction method, a learning device, and a learning method.

Background Art

[0002] Patent Document 1 discloses a heat pump hot water supply system having a learning capability for predicting heat demand in a future heat demand prediction target period based on past hot water supply operation results.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Patent Document 1 does not disclose specific details of the hot water supply operation results. The present specification provides a technology capable of more accurately predicting the future power consumption of facility equipment compared to the prior art.

Means for Solving the Problems

[0005] In a first aspect of this technology, the power consumption prediction system may predict the power consumption of equipment. In this case, the equipment may include an electric heat source and a combustion heat source. The power consumption prediction system may include an acquisition unit that acquires energy consumption information relating to the energy consumption of the equipment, and a calculation unit that calculates the predicted power consumption for a first period after the present time based on the energy consumption information acquired by the acquisition unit. Furthermore, the energy consumption information may include at least the power consumption of the equipment in a second period corresponding to the first period and prior to the present time, the amount of gas consumed by the combustion heat source in the second period, and climate information relating to the climate near the equipment in a third period including the first period. The calculation unit may calculate the predicted power consumption by executing calculation processing of a trained learner that has undergone machine learning to calculate the predicted power consumption for the first period based on the energy consumption information.

[0006] Electric heat sources are heat sources that primarily heat water by consuming electricity, such as heat pumps and electric heaters. In contrast, combustion heat sources are heat sources that heat water by burning fuel gas, such as burners. The inventors have discovered that in equipment equipped with combustion heat sources and electric heat sources, not only the amount of power consumed by the equipment in the second period, but also the amount of gas consumed by the combustion heat source in the second period and climate information regarding the climate near the equipment in the third period, including the first period, correlate with the predicted power consumption of the equipment in the first period. The power consumption prediction system described above calculates the predicted power consumption of the equipment in the first period based on energy consumption information that includes, in addition to the power consumption of the equipment in the second period, at least the amount of gas consumed by the combustion heat source and climate information. Therefore, it is possible to predict power consumption more accurately than with conventional technology.

[0007] In a second aspect of this technology, the power consumption prediction system in the first aspect may further include a transmission unit that transmits the predicted power consumption calculated by the calculation unit to an external device.

[0008] According to this embodiment, the predicted power consumption calculated by the calculation unit can be transmitted to an external device.

[0009] In a third aspect of this technology, the power consumption prediction system in the first aspect may further include a server capable of communicating with equipment. In this case, the server may include an acquisition unit and a calculation unit.

[0010] According to this embodiment, for example, a server can calculate the predicted power consumption of multiple pieces of equipment.

[0011] In a fourth aspect of this technology, in the third aspect described above, the server may further include a transmission unit that transmits the predicted power consumption calculated by the arithmetic unit to an external device.

[0012] According to this embodiment, for example, the predicted power consumption of multiple pieces of equipment can be transmitted to an external device via a server.

[0013] In a fifth aspect of this technology, in the second or fourth aspect described above, the power consumption prediction system may further include an equipment control unit that controls equipment based on an external instruction received from an external device that has received the predicted power consumption.

[0014] According to this embodiment, based on external instructions received from an external device, the equipment can be controlled, for example, to reduce the power consumption of the equipment.

[0015] In a sixth aspect of this technology, in the fifth aspect described above, the equipment control unit may restrict the operation of the equipment during a reduction period if the external instruction includes a reduction instruction for reducing the power consumption of the equipment during a predetermined reduction period.

[0016] According to this embodiment, the power consumption of equipment can be reduced, for example, in response to a reduction instruction received from an external device.

[0017] In the seventh aspect of this technology, in the sixth aspect described above, the equipment may be capable of performing a bath filling operation in which high-temperature water heated by the equipment is supplied to the bathtub, and the equipment may be configured to specify a bath filling completion time before it starts the bath filling operation. In that case, the equipment control unit may change the time to start the bath filling operation so that the bath filling operation is completed before the reduction period if the period during which the equipment performs the bath filling operation in order to complete the bath filling operation by the specified bath filling completion time is included in the reduction period.

[0018] According to this embodiment, even if a time is specified for the equipment to perform a hot water filling operation during the reduction period, it is possible to prevent the equipment from performing the hot water filling operation during the reduction period. Therefore, it is possible to prevent an increase in the power consumption of the equipment during the reduction period. Furthermore, since the hot water filling operation is completed by the specified time, it is possible to prevent a decrease in user convenience.

[0019] In the eighth aspect of this technology, in the sixth or seventh aspect described above, the equipment may be capable of performing a bath-filling operation in which high-temperature water heated by the equipment is supplied to the bathtub. In that case, if the equipment control unit receives a bath-filling instruction to perform a bath-filling operation during the reduction period, it may stop the operation of the electric heat source and supply high-temperature water heated by the combustion heat source to the bathtub during the bath-filling operation.

[0020] Compared to combustion-type heat sources that heat water by burning fuel gas, electric heat sources that heat water mainly by consuming electricity tend to consume more power. According to this embodiment, even if a bath filling instruction is received during the reduction period, high-temperature water heated by the combustion-type heat source, rather than the electric heat source, is supplied to the bathtub. Since it is possible to prevent the electric heat source from heating the water during the reduction period, it is possible to prevent an increase in power consumption during the reduction period.

[0021] In the ninth aspect of the present technology, in any one of the sixth to eighth aspects described above, the facility equipment may be capable of performing a reheating operation for heating water in a bathtub. In that case, when receiving a reheating instruction for executing the reheating operation during the reduction period, the device control unit may stop the operation of the electric heat source device and heat the water in the bathtub by the combustion heat source device in the reheating operation.

[0022] According to this aspect, even when a heating instruction is received during the reduction period, the water in the bathtub is heated by the combustion heat source device instead of the electric heat source device. Since it is possible to prevent the electric heat source device from heating water during the reduction period, it is possible to prevent the power consumption during the reduction period from increasing.

[0023] In the tenth aspect of the present technology, in any one of the first to ninth aspects, the facility equipment may further include a remote controller communicably connected to the electric heat source device and the combustion heat source device. In that case, the electric heat source device may be configured to be capable of boiling water in a tank. The energy consumption information may further include period information indicating which unit period the first period is included in when one year is divided into a plurality of unit periods, boiling time information related to the time during which the electric heat source device in the tank boiled the water in the tank in the second period, and on-time information related to the time during which the power of the remote controller was turned on in the second period.

[0024] The operating time of the facility equipment may vary depending on which period information within one year the first period is included in. Further, when the boiling time in the second period increases, it is estimated that the operating time of the electric heat source device increases in the second period. Furthermore, when the on-time in the second period increases, it is estimated that the operating time of the facility equipment increases in the second period. According to this aspect, the predicted power consumption can be calculated more accurately further based on the period information, the boiling time information, and the on-time information.

[0025] Furthermore, the present specification also discloses a learning device capable of calculating a predicted power consumption amount for facility equipment. In the learning device of this aspect, the facility equipment may include an electric heat source machine and a combustion heat source machine. The learning device may include an acquisition unit that acquires energy consumption information related to the energy consumption of the facility equipment, and a learning unit that performs machine learning of a learning device so as to calculate the predicted power consumption amount of the facility equipment in a first period after the current point in time based on the energy consumption information acquired by the acquisition unit. Further, the energy consumption information may at least include the power consumption amount of the facility equipment in a second period that corresponds to the first period and precedes the current point in time, the gas consumption amount of the combustion heat source machine in the second period, and climate information related to the climate in the vicinity of the facility equipment in a third period including the first period.

[0026] The power consumption prediction method implemented by the above-described power consumption prediction system and the learning method executed by the above-described learning device are also novel and advantageous. [Brief Description of the Drawings]

[0027] [Figure 1] A configuration diagram of the power consumption prediction system 100 according to the embodiment is shown. [Figure 2] A flowchart of learning information generation processing is shown. [Figure 3] An example of each piece of information used in the learning processing is shown. [Figure 4] A flowchart of learning processing is shown. [Figure 5] A flowchart of power consumption amount prediction processing is shown. [Figure 6] An example of each piece of information used in the power consumption amount prediction processing is shown. [Figure 7] A flowchart of power consumption amount reduction processing is shown. [Mode for Carrying Out the Invention]

[0028] (Embodiment) As shown in Figure 1, the power consumption prediction system 100 of this embodiment comprises a lower-level management server 40 and a hot water heating system 10 installed in the house 2. The power consumption prediction system 100 is a system for predicting the amount of electricity consumed in the hot water heating system 10 based on various information obtained from the hot water heating system 10.

[0029] The subordinate management server 40 is installed on the Internet 8, for example, by the manufacturer of the hot water heating system 10. In a modified example, the subordinate management server 40 may be installed on the Internet 8 by a company other than the manufacturer of the hot water heating system 10. The subordinate management server 40 is connected to the superior management server 70 via the Internet 8.

[0030] The higher-level management server 70 is installed on the Internet 8 by the aggregator 72. The aggregator 72 is a business that controls the amount of electricity consumed between a power company (not shown) that generates or distributes electricity and a consumer (e.g., house 2) that consumes the power of the power company. For example, during a predetermined time when electricity consumption tends to increase (e.g., between 3pm and 6pm), the aggregator 72 sends reduction instructions via the higher-level management server 70 to house 2 and other houses (not shown) in the area surrounding house 2 so that the amount of electricity consumed at house 2 is below a predetermined amount. As a result, the amount of electricity consumed at house 2 during the predetermined time is reduced, and the maximum amount of electricity consumed in the area surrounding house 2 can be reduced. Therefore, it is possible to maintain a balance between the amount of electricity supplied by the power company and the amount of electricity consumed by consumers, so-called demand response can be achieved.

[0031] The hot water heating system 10 is a system for supplying hot water to the bathtub 4 or faucet 6 of the house 2, and also for heating the interior of the house 2. In the following, supplying hot water to the bathtub 4 will also be referred to as filling the bathtub. The hot water heating system 10 includes a kitchen remote control 12, a bathroom remote control 14, a dressing room remote control 16, an HP (abbreviation for heat pump) heat source unit 20, and a gas heat source unit 30.

[0032] The HP heat source unit 20 comprises an HP unit 22, a tank unit 24, and a first power sensor S1. The HP unit 22 is connected to the tank 25 of the tank unit 24 by an HP circulation circuit (not shown), and further connected to the heating circuit (not shown) of the gas heat source unit 30. The HP unit 22 can heat hot water (e.g., water) or heating water (e.g., antifreeze) by absorbing heat from the outside air through a heat pump cycle using a refrigerant (e.g., a fluorocarbon refrigerant). The HP heat source unit 20 can also heat the water in the tank 25 or heat the heating water circulating in the heating circuit using the HP unit 22.

[0033] The first power sensor S1 measures a first power consumption amount indicating the amount of power consumed by the HP heat source unit 20 and transmits it to the kitchen remote control 12. The first power consumption amount includes, for example, the power consumed by heating by the HP unit 22, i.e., the power consumed by the operation of the circulation pump, expansion valve, compressor, fan (not shown), etc. of the HP unit 22.

[0034] The gas heat source unit 30 includes a hot water burner 32S, a heating burner 32H, a panel heater 34, a bathroom heater 36, a floor heater 38, a second power sensor S2, and a gas sensor S3. The hot water burner 32S is installed downstream of the tank 25 of the tank unit 24. The hot water burner 32S heats the water supplied from the tank 25 by burning fuel gas (e.g., city gas) supplied from a fuel supply source (not shown). Downstream of the hot water burner 32S are a bathtub 4, a faucet 6, etc. High-temperature water heated by the hot water burner 32S is supplied to the bathtub 4 and the faucet 6. The bathtub 4 is connected to a reheating heat exchanger (not shown) installed in the heating circuit via a bathtub water circulation circuit (not shown). The bathtub water circulation circuit circulates the water in the bathtub 4 between the reheating heat exchanger (not shown) and the bathtub 4 using a bathtub water circulation pump (not shown) installed in the bathtub water circulation circuit. This allows the water in the bathtub 4 to be heated by the heat of the heating water circulating in the heating circuit. In the following, heating the water in the bathtub 4 as described above will also be referred to as reheating the water in the bathtub 4.

[0035] The heating burner 32H is installed in the heating circuit described above and heats the heating water circulating in the heating circuit by burning fuel gas supplied from the fuel source. The heating water is stored in a cistern (not shown), which is an open-topped container, and is circulated in the heating circuit by a heating water circulation pump (not shown). Panel heaters 34, bathroom heaters 36, floor heating 38, etc. are connected to the heating circuit. The panel heaters 34, bathroom heaters 36, and floor heating 38 heat the interior of the house 2 with the heat from the heating water heated by the heating burner 32H.

[0036] The second power sensor S2 measures a second power consumption amount, which indicates the amount of electricity consumed by the gas heat source unit 30, and transmits it to the kitchen remote control 12. The second power consumption amount includes, for example, the power consumed by igniting the hot water burner 32S and the heating burner 32H. The gas sensor S3 measures the amount of gas consumed, which indicates the amount of fuel gas consumed by the gas heat source unit 30, and transmits it to the kitchen remote control 12.

[0037] Thus, in the hot water supply and heating system 10 of this embodiment, both the HP heat source unit 20 and the gas heat source unit 30 are used to provide hot water and heating. The HP unit 22 consumes a certain amount of electricity when heating the hot water supply water and heating water. On the other hand, the gas heat source unit 30 consumes a certain amount of fuel gas when heating the hot water supply water and heating water, but does not consume much electricity.

[0038] The kitchen remote control 12 is connected to the internet 8 via a router 7 installed in the house 2. The kitchen remote control 12 is installed in the kitchen (not shown) of the house 2 and is equipped with a third power sensor S4. The kitchen remote control 12 is connected to each heat source unit 20, 30 in a communicative manner and displays various information including the operating status of each heat source unit 20, 30. For example, the kitchen remote control 12 displays the set temperature for filling the bathtub 4 and the set temperature for supplying hot water to the faucet 6. For example, the kitchen remote control 12 lights up a heat retention lamp when the water retention operation in the bathtub 4 is being performed. In the water retention operation in the bathtub 4 in this embodiment, after filling the bathtub 4, if the temperature of the water in the bathtub 4 drops, the water in the bathtub 4 is circulated in the bathtub water circulation circuit described above and heated to maintain the water in the bathtub 4 at the set temperature for filling the bathtub 4.

[0039] Furthermore, the kitchen remote control 12 receives various instructions from the user. For example, the kitchen remote control 12 receives a stop instruction from the user to stop the water heating operation of the bathtub 4 described above. When the kitchen remote control 12 receives a stop instruction, it stops the heating lamp from lighting up. The kitchen remote control 12 also receives a water filling instruction from the user to perform the water filling operation of the bathtub 4. Furthermore, the kitchen remote control 12 receives a reheating instruction from the user to perform the reheating operation of the water in the bathtub 4. The kitchen remote control 12 also receives instructions from the user to change the water filling temperature setting and the hot water supply temperature setting. Furthermore, the kitchen remote control 12 receives a timer instruction from the user to perform the water filling operation according to a timer time that has been entered in advance.

[0040] The third power sensor S4 measures a third power consumption, which indicates the amount of power consumed by the kitchen remote control 12. The third power consumption includes, for example, the power consumed by turning on the lighting (not shown) of the kitchen remote control 12.

[0041] The bathroom remote control 14 is installed in the bathroom (not shown) of the house 2 and is equipped with a fourth power sensor S5. The bathroom remote control 14 displays, for example, the set temperature for filling the bathtub, the set temperature for supplying hot water, etc. The bathroom remote control 14 receives instructions from the user for filling the bathtub, reheating, changing settings, timer instructions, etc. The fourth power sensor S5 measures a fourth power consumption amount, which indicates the amount of power consumed by the bathroom remote control 14. The fourth power consumption amount includes, for example, the power consumed by turning on the lighting (not shown) of the bathroom remote control 14.

[0042] The dressing room remote control 16 is installed in the dressing room (not shown) of the house 2 and is equipped with a fifth power sensor S6. The dressing room remote control 16 displays various information, including the on / off status of the bathroom heater 36 and the operating status of each operation of the bathroom heater 36. The dressing room remote control 16 receives, for example, an on / off instruction for the bathroom heater 36 from the user. The fifth power sensor S6 measures a fifth power consumption amount, which indicates the amount of power consumed by the dressing room remote control 16. The fifth power consumption amount includes, for example, the power consumed by turning on the lighting (not shown) of the dressing room remote control 16.

[0043] The kitchen remote control 12 is a computer equipped with a CPU and memory composed of volatile memory and non-volatile memory. The kitchen remote control 12 receives a first power consumption amount measured by a first power sensor S1 from the HP heat source unit 20. The kitchen remote control 12 integrates the first power consumption amounts received during a predetermined measurement period and calculates the first integrated power consumption amount consumed by the HP heat source unit 20 during that measurement period. Similarly, the kitchen remote control 12 receives a second power consumption amount measured by a second power sensor S2 and a gas consumption amount measured by a gas sensor S3 from the gas heat source unit 30. The kitchen remote control 12 integrates the second power consumption amounts received during the measurement period and calculates the second integrated power consumption amount consumed by the gas heat source unit 30 during that measurement period. Furthermore, the kitchen remote control 12 integrates the gas consumption amounts received during the measurement period and calculates the integrated gas consumption amount consumed by the gas heat source unit 30 during that measurement period.

[0044] The kitchen remote control 12 accumulates the third power consumption measured by the third power sensor S4 during the measurement period and calculates the third accumulated power consumption consumed by the kitchen remote control 12 during that measurement period. The kitchen remote control 12 receives the fourth power consumption measured by the fourth power sensor S5 from the bathroom remote control 14 and the fifth power consumption measured by the fifth power sensor S6 from the dressing room remote control 16. The kitchen remote control 12 accumulates the fourth power consumption received during the measurement period and calculates the fourth accumulated power consumption consumed by the bathroom remote control 14 during that measurement period, and accumulates the fifth power consumption received during the measurement period and calculates the fifth accumulated power consumption consumed by the dressing room remote control 16 during that measurement period.

[0045] Furthermore, the kitchen remote control 12 measures the on-time, which indicates the time during the measurement period when the kitchen remote control 12 is turned on. The kitchen remote control 12 also calculates the bath filling time, which indicates the time during the measurement period when the bath filling operation described above is being performed. For example, the kitchen remote control 12 calculates the bath filling time as the time from when it receives the bath filling instruction described above until it receives a bath filling completion signal from the gas heat source unit 30. The kitchen remote control 12 further measures the warming time, which indicates the time during the measurement period when the warming operation described above is set. For example, the kitchen remote control 12 identifies the warming time as the time when the warming lamp described above is lit. Furthermore, the kitchen remote control 12 calculates the combustion time, which indicates the time during the measurement period when either the hot water burner 32S or the heating burner 32H of the gas heat source unit 30 is burning. For example, the kitchen remote control 12 calculates the combustion time by receiving a combustion signal from the gas heat source unit 30 indicating that either the hot water burner 32S or the heating burner 32H is burning. In a modified example, the kitchen remote control 12 may calculate the time during which only the hot water burner 32S is burning as the burning time. The kitchen remote control 12 also calculates the boiling time during which the HP heat source unit 20 is heating the water in the tank 25. The kitchen remote control 12 calculates the boiling time, for example, by receiving a boiling signal from the HP heat source unit 20 indicating that it is boiling the water in the tank 25. Furthermore, the kitchen remote control 12 calculates the total power consumption by summing the first cumulative power consumption of the HP heat source unit 20, the second cumulative power consumption of the gas heat source unit 30, the third cumulative power consumption of the kitchen remote control 12, the fourth cumulative power consumption of the bathroom remote control 14, and the fifth cumulative power consumption of the dressing room remote control 16.

[0046] The kitchen remote control 12 transmits the cumulative gas consumption, total power consumption, on-time, bath filling time, heat retention time, combustion time, and boiling time to the lower management server 40 at predetermined intervals. Hereafter, this information (i.e., cumulative gas consumption, total power consumption, on-time, bath filling time, heat retention time, combustion time, and boiling time) may be collectively referred to as operation information H1. The measurement period is a unit of time corresponding to the prediction period described later, and is a unit of time that divides one day. The measurement period is specified by the aggregator 72. In this embodiment, the measurement period is 30 minutes, so one day is divided into 48 measurement periods. The kitchen remote control 12 updates the operation information H1 every 30 minutes.

[0047] The lower-level management server 40 comprises a communication interface 42, a display unit 46, an operation unit 48, and a control unit 50. The communication interface 42 is connected to the kitchen remote control 12 via the internet 8. The communication interface 42 is further connected to the higher-level management server 70 via the internet 8. The display unit 46 displays, for example, information about the hot water heating system 10 of the house 2. The operation unit 48 receives various inputs from the administrator of the lower-level management server 40.

[0048] The control unit 50 is a computer comprising a CPU 52 and a memory 54 composed of volatile memory and non-volatile memory. The memory 54 stores operational information H1 received from the kitchen remote control 12, a measurement period of "30 minutes" received from the higher-level management server 70, installation equipment information P1 (described later), performance information T1, learning information E1, learning result data D1, and the learner 60. The installation equipment information P1 is information about the equipment of the hot water heating system 10 installed in the house 2. The installation equipment information P1 includes, for example, the model name and number of panel heaters 34, the model name and number of bathroom heaters 36, and the model name and size of floor heating 38. When the hot water heating system 10 is installed in the house 2, the lower-level management server 40 obtains the installation equipment information P1 along with house identification information that identifies the house 2 via the kitchen remote control 12, and stores it in the memory 54 in association with the house identification information. The performance information T1 and learning information E1 are generated by the learning information generation process described in Figure 2. The learning unit 60 is equipped with a neural network. The learning unit 60 has already performed the learning process described later, referring to Figure 4. In the modified example, the learning unit 60 may be a computer with a different hardware configuration from the subordinate management server 40.

[0049] Referring to Figure 2, the learning information generation process for generating learning information E1 will be explained. Learning information E1 contains various information regarding the energy consumption of the hot water heating system 10 during the actual period described later, and is used in the learning process shown in Figure 4. In the following, the entity that executes the learning information generation process and the power consumption prediction process (see Figure 5) will be referred to as the subordinate management server 40 instead of the CPU 52. The subordinate management server 40 executes the learning information generation process by accepting an operation from the administrator of the subordinate management server 40 after the predetermined actual period (for example, three years from January 1, 2022 to December 31, 2024) has ended.

[0050] In S10, the subordinate management server 40 generates performance information T1. As shown in Figure 2, the performance information T1 includes, for example, date information indicating the dates included in the performance period for house 2, and for each day indicated by the date information, weekly information, the predicted maximum temperature TH1, the predicted minimum temperature TL1, the installed equipment information P1 mentioned above, and operation information H1. The weekly information indicates the week of the date indicated by the date information. That is, the weekly information indicates which week of the year the date indicated by the date information falls into, when the year is divided into one-week intervals. The predicted maximum temperature TH1 and predicted minimum temperature TL1 are obtained from a weather forecast website for the area surrounding the hot water heating system 10. In a modified example, the CPU 52 of the subordinate management server 40 may obtain the maximum and minimum temperatures by measuring the actual outside temperature near the hot water heating system 10 using, for example, an outside air temperature sensor (not shown) installed in the tank unit 24. As described above, the operational information H1 includes the cumulative amount of gas consumed for each measurement period (e.g., CG1, CG2), total power consumption (e.g., SE1, SE2), on-time (e.g., OT1, OT2), bath filling time (e.g., HT1, HT2), heat retention time (e.g., KT1, KT2), ​​combustion time (e.g., FT1, FT2), and boiling time (e.g., BT1, BT2). Although not shown in the diagram, as described above, in this embodiment the measurement period is 30 minutes, so the actual information T1 includes 48 measurement periods and corresponding information for each day in the actual period. The actual information T1 further includes similar information about other houses, etc., located in the area surrounding house 2. The subordinate management server 40 stores the generated actual information T1 in memory 54 (see Figure 1).

[0051] In S12, the lower-level management server 40 identifies a specific measurement period from among the multiple measurement periods included in the actual data T1 generated in S10 as the target measurement period. The target measurement period is the measurement period for which the predicted power consumption is calculated during the learning process.

[0052] In S14, the subordinate management server 40 calculates measurement condition information E2. Measurement condition information E2 is information input to the learner 60 during the learning process, and is information about various conditions during the target measurement period. As shown in Figure 3, measurement condition information E2 includes date information for the target measurement day (e.g., January 31, 2022), target measurement period (e.g., 9:00~9:29), weekly information for the target measurement day (e.g., 5), installed equipment information P1, predicted maximum temperature TH2, predicted minimum temperature TL2, daily temperature difference TD1, maximum temperature difference TH3, and minimum temperature difference TL3. From the actual information T1 generated in S10, the subordinate management server 40 identifies the date information, weekly information, installed equipment information P1, predicted maximum temperature TH2, and predicted minimum temperature TL2. Furthermore, the subordinate management server 40 calculates the daily temperature difference TD1 from the difference between the predicted maximum temperature TH2 and the predicted minimum temperature TL2. Furthermore, the subordinate management server 40 calculates the maximum temperature difference TH3 from the difference between the predicted maximum temperature TH2 and the predicted maximum temperature on the day before the target measurement date. Similarly, the subordinate management server 40 calculates the minimum temperature difference TL3 from the difference between the predicted minimum temperature TL2 and the predicted minimum temperature on the day before the target measurement date. In this way, the subordinate management server 40 calculates the measurement condition information E2.

[0053] In S16, the subordinate management server 40 calculates the input period information E3. The input period information E3 is information input to the learner 60 in the learning process, and is information regarding the operation information H1 of the hot water heating system 10 during the input period immediately preceding the target measurement date. In this embodiment, the input period is the 30 days immediately preceding the target measurement date (for example, January 1st to January 30th, 2022). As shown in Figure 3, the input period information E3 includes the operation information H1 for the same measurement period as the target measurement period in the input period (i.e., 9:00 to 9:29), and the period average corresponding to each piece of information included in the operation information H1. The period average is the average value over three measurement periods, including the same measurement period as the target measurement period and the measurement periods before and after that measurement period. In the example shown in Figure 3, the period average AG1 is calculated by averaging the cumulative gas consumption CG3 for the same measurement period as the target measurement period (e.g., 9:00-9:29), the cumulative gas consumption for the measurement period before the target measurement period (e.g., 8:30-8:59), and the cumulative gas consumption for the measurement period after the target measurement period (e.g., 9:30-09:59). Similarly, the other period averages AE1, AO1, AH1, AK1, AF1, and AB1 are calculated by averaging the corresponding values ​​for three consecutive measurement periods. The subordinate management server 40 identifies the operational information H1 for all measurement periods that are the same as the target measurement period in the input period and calculates the period average. In this embodiment, the input period information E3 includes the operational information H1 and period averages corresponding to 30 measurement periods.

[0054] In S18, the subordinate management server 40 calculates statistical information E4. Statistical information E4 is information input to the learner 60 during the learning process, and includes the item mean and item variance of the operational information H1 during the input period (e.g., 30 days) described above. As shown in Figure 3, the item mean represents the average value of each value of the operational information H1 during the input period. For example, the item mean MG1 is calculated by averaging the cumulative gas consumption for the same measurement period as the target measurement period, which is 30 days from January 1 to January 30, 2022. Similarly, the other item means ME1, MO1, MH1, MK1, MF1, and MB1 are also calculated by averaging the values ​​for the same measurement period as the target measurement period, which is 30 days from January 1 to January 30, 2022. The variance represents the variance of each value of the operational information H1 during the input period. For example, item variance VG1 is the variance value for the cumulative gas consumption over the same measurement period as the target measurement period, from January 1st to January 30th, 2022, and is calculated based on the cumulative gas consumption CG3 and the item average MG1. Similarly, the other item variances VE1, VO1, VH1, VK1, VF1, and VB1 are also the variance values ​​for their respective values.

[0055] In S20, the subordinate management server 40 identifies all measurement periods included in the performance information T1 generated in S10 as target measurement periods and determines whether or not it has calculated the above-mentioned information E2 to E4. The subordinate management server 40 repeats the processes in S12 to S18 until all measurement periods are identified as target measurement periods (NO in S20), and proceeds to S22 if all measurement periods are identified as target measurement periods (YES in S20). Note that if, for example, January 2, 2022, 9:00 to 9:29 is identified as a target measurement period among the measurement periods included in the performance information T1, the input period has not reached 30 days. In such a case, the subordinate management server 40 may proceed to S22 without identifying that measurement period as a target measurement period.

[0056] In S22, the subordinate management server 40 stores the measurement condition information E2 calculated in S14, the input period information E3 calculated in S16, and the statistical information E4 calculated in S18 as learning information E1 in memory 54 (see Figure 1). With this, the subordinate management server 40 terminates the process shown in Figure 2.

[0057] Referring to Figure 4, the learning process performed by the learning unit 60 will be described. The learning process is machine learning to calculate, for example, the predicted amount of power consumed by the hot water heating system 10 in the house 2, based on the learning information E1 generated by the process in Figure 2. After the learning information generation process in Figure 2 is completed, the learning unit 60 executes the learning process by accepting an operation from the administrator of the subordinate management server 40.

[0058] In S30, the learning device 60 acquires the performance information T1 stored in the memory 54. As described above, the performance information T1 includes multiple measurement periods included in the performance period and corresponding operational information H1.

[0059] In S32, the learner 60 acquires the learning information E1 stored in memory 54. As described above, the learning information E1 includes measurement condition information E2, input period information E3, and statistical information E4 for each measurement period included in the performance period.

[0060] In S34, the learner 60 uses the actual data T1 acquired in S30 and the learning data E1 acquired in S30 to perform a neural network training process to calculate the predicted power consumption from the prediction condition information E12, the previous period information E13, and the prediction statistics information E14, which will be described later. At this time, the learner 60 uses the measurement condition information E2, the input period information E3, and the statistics information E4 contained in the learning data E1 as input data, and uses the total power consumption for the target measurement period contained in the actual data T1 as training data (ground truth data). Using the measurement condition information E2, the input period information E3, and the statistics information E4 contained in the learning data E1 as input data, the learner 60 adjusts, for example, the weights of the connections between each neuron in the neural network and the thresholds of each neuron until the value calculated from the neural network (i.e., the predicted power consumption) matches the value corresponding to the total power consumption for the target measurement period contained in the actual data T1. This allows the learner 60 to construct a neural network for calculating predicted power consumption from prediction condition information E12, previous period information E13, and prediction statistical information E14.

[0061] In S36, the learner 60 stores the weights of the connections between each neuron and the threshold values ​​of each neuron in the neural network constructed in the learning process of S34 as learning result data D1 in memory 54 (see Figure 1). After that, the learner 60 terminates the process shown in Figure 4.

[0062] Referring to Figure 5, the power consumption prediction process performed by the lower-level management server 40 will be explained. The power consumption prediction process is a process for calculating the predicted power consumption expected to be consumed by, for example, the hot water heating system 10 of house 2. The lower-level management server 40 receives, for example, the prediction period (for example, 9:00 to 9:29 on January 31, 2025) to be used for calculating the predicted power consumption from the upper-level management server 70. The lower-level management server 40 executes the process shown in Figure 5, for example, a predetermined period before the end of the day before the prediction date (January 31, 2025) that includes the prediction period (for example, 10 minutes before). In a modified example, the lower-level management server 40 may execute the process shown in Figure 5 at the end of the day two days before the prediction date (for example, January 29, 2025), or one week before the prediction date.

[0063] In S40, the lower-level management server 40 calculates forecast condition information E12 regarding various conditions on the forecast day. Forecast condition information E12 is information that corresponds to the measurement condition information E2 described above. As shown in Figure 6, forecast condition information E12 includes the forecast period received from the upper-level management server 70, weekly information for the forecast day, installed equipment information P1 for the forecast day, the expected maximum temperature TH12 and expected minimum temperature TL12 for the forecast day, the temperature difference TD11 for the day, the maximum temperature difference TH13, and the minimum temperature difference TL13. Similar to the measurement condition information E2 described above, the lower-level management server 40 identifies the date information, weekly information, and installed equipment information P1 from the actual information T1 acquired in S30. Furthermore, the lower-level management server 40 obtains the expected maximum temperature TH12 and expected minimum temperature TL12 for the forecast day from the weather forecast website described above. The lower-level management server 40 calculates the temperature difference TD11 for the day from the difference between the expected maximum temperature TH12 and the expected minimum temperature TL12. Furthermore, the lower-level management server 40 calculates the maximum temperature difference TH13 from the difference between the predicted maximum temperature TH12 and the previous day's predicted maximum temperature, and calculates the minimum temperature difference TL13 from the difference between the predicted minimum temperature TL12 and the previous day's predicted minimum temperature. In this way, the lower-level management server 40 obtains the forecast condition information E12.

[0064] In S42, the subordinate management server 40 calculates the preceding period information E13 for the period immediately preceding the forecast period. The preceding period information E13 is the same information as the input period information E3 described above. The preceding period information E13 includes, for example, information from the operational information H1 in the actual period for the same measurement period (e.g., 9:00 to 9:29) as the forecast period for the period immediately preceding the forecast date (e.g., the 30 days from January 1st to January 30th, 2025). Furthermore, the preceding period information E13, like the input period information E3, includes the period average, which is the average value of three measurement periods, including the same measurement period as the forecast period (e.g., 9:00 to 9:29), and the measurement periods before and after the forecast period (e.g., 8:30 to 8:59 and 9:30 to 9:59). For example, the period average AG11 shown in Figure 6 is the average of the cumulative gas consumption CG13 for the same measurement period as the forecast period and the cumulative gas consumption for the measurement periods before and after the forecast period. Similarly, the other period averages AE11, AO11, AH11, AK11, AF11, and AB11 are also the average values ​​of the corresponding values ​​over the three measurement periods. The subordinate management server 40 calculates the operational information H1 and the period average for all measurement periods that are the same as the forecast period in the immediately preceding period. Therefore, the immediately preceding period information E13 includes the operational information H1 and period average for 30 measurement periods.

[0065] In S44, the subordinate management server 40 calculates predictive statistics E14. Predictive statistics E14 corresponds to the statistics E4 described above. Predictive statistics E14 includes the item mean and item variance of operational information H1 for the preceding period (e.g., 30 days). Item means MG11, ME11, MO11, MH11, MK11, MF11, MB11 are calculated by averaging each value over the 30 days from January 1 to January 30, 2025. Item variances VG11, VE11, VO11, VH11, VK11, VF11, VB11 are variance values ​​calculated based on the corresponding values ​​and item means.

[0066] In S50, the lower management server 40 inputs the prediction condition information E12 calculated in S40, the previous period information E13 calculated in S42, and the prediction statistical information E14 calculated in S44 to the learner 60. The learner 60 then performs calculations to calculate the predicted power consumption for each of the pieces of information E12, E13, and E14. As mentioned earlier, the learner 60 has already performed the learning process shown in Figure 4 based on the learning information E1, which includes the measurement condition information E2 corresponding to the prediction condition information E12, the input period information E3 corresponding to the previous period information E13, and the statistical information E4 corresponding to the prediction statistical information E14, as well as the actual information T1. Therefore, the learner 60 uses the learning result data D1 (see Figure 1) stored in memory 54, i.e., the weights of the connections between each neuron in the neural network constructed in the learning process shown in Figure 4, and the threshold values ​​of each neuron, to perform calculations.

[0067] In S52, the lower-level management server 40 obtains the predicted power consumption FE1 calculated by the learner 60. The predicted power consumption FE1 is the power consumption that is predicted to be consumed by the hot water heating system 10 during the prediction period, as received from the upper-level management server 70.

[0068] In S54, the lower-level management server 40 transmits the acquired predicted power consumption FE1 to the upper-level management server 70. This allows the calculated predicted power consumption FE1 to be transmitted to the aggregator 72. Once the lower-level management server 40 transmits the predicted power consumption FE1 to the upper-level management server 70, it terminates the process shown in Figure 5.

[0069] Similarly, the lower-level management server 40 inputs prediction condition information, previous period information, and prediction statistics information for other houses in the area surrounding house 2, in addition to house 2, into the learning device 60, calculates the predicted power consumption for the hot water heating system of those other houses during the prediction period, and transmits it to the higher-level management server 70. In this way, the lower-level management server 40 can perform power consumption prediction processing for multiple houses and transmit multiple predicted power consumption values ​​to the higher-level management server 70.

[0070] The higher-level management server 70 determines whether a reduction in power consumption is necessary in the area surrounding house 2 based on multiple predicted power consumption amounts FE1, etc., received from the lower-level management server 40. For example, if the total predicted power consumption of multiple houses, including house 2, exceeds a threshold percentage of the power supplied by the power company during a specific prediction period, the higher-level management server 70 determines that a reduction in power consumption is necessary during that specific prediction period and sends a reduction instruction to the lower-level management server 40. The reduction instruction includes the house whose power consumption should be reduced (e.g., house 2) and a reduction period indicating the period during which the power consumption should be reduced. When the lower-level management server 40 receives a reduction instruction from the aggregator 72 via the higher-level management server 70, it sends the reduction instruction to the target house.

[0071] Referring to Figure 7, the power consumption reduction process performed by the kitchen remote control 12 will be explained. The power consumption reduction process is a process to reduce the power consumption of the hot water heating system 10 of the house 2 in accordance with the reduction instruction described above. The kitchen remote control 12 performs the process shown in Figure 7 in response to receiving a reduction instruction from the lower management server 40.

[0072] In S60, the kitchen remote control 12 determines whether or not it has received the timer instruction mentioned above. If the time for filling the bathtub, specified by the timer instruction, is not stored in the kitchen remote control 12's memory (not shown), the kitchen remote control 12 determines that it has not received the timer instruction (NO in S60), skips S62 and S64, and proceeds to S70. If the time for filling the bathtub, it determines that it has received the timer instruction (YES in S60), and proceeds to S62. In this embodiment, the time for filling the bathtub, it indicates the time at which the bathtub filling operation is completed.

[0073] In S62, the kitchen remote control 12 determines whether the bath filling period, during which the bath filling operation is performed according to the timer instruction, is included in the reduction period included in the reduction instruction. The bath filling period is the period from the bath filling start time to the bath filling completion time. Based on the amount of high-temperature water supplied to the bathtub 4, the heating capacity of the hot water supply burner 32S, and the bath filling set temperature specified by the user, the kitchen remote control 12 calculates the bath filling start time in order to complete the bath filling operation by the bath filling completion time. If the bath filling period is not included in the reduction period (NO in S62), the kitchen remote control 12 skips S64 and proceeds to S70. If the bath filling period is included in the reduction period (YES in S62), the kitchen remote control 12 proceeds to S64.

[0074] In S64, the kitchen remote control 12 changes the start time for filling the bathtub. The kitchen remote control 12 advances the start time so that the bathtub filling operation is completed before the reduction period begins. When the advanced start time for filling the bathtub arrives, the kitchen remote control 12 controls, for example, the HP heat source unit 20 and the gas heat source unit 30 to start filling the bathtub 4. This allows the bathtub filling operation to be completed before the reduction period begins. In this way, even if a timer instruction is received (YES in S60), it is possible to prevent the bathtub filling operation using the HP heat source unit 20 and the gas heat source unit 30 from being performed within the reduction period. For this reason, for example, it is possible to prevent an increase in the power consumption of the HP heat source unit 20 and the gas heat source unit 30 during the reduction period, and contribute to the realization of demand response. Furthermore, since the bathtub filling operation is completed by the bathtub filling completion time specified by the timer instruction, it is possible to prevent a decrease in user convenience. In this embodiment, the kitchen remote control 12 also changes the start time for filling the bathtub so that the bathtub filling operation is completed at the start time of the reduction period. As a result, the actual completion time of the bath filling operation approaches the completion time specified by the timer, which helps to suppress the decrease in the temperature of the water in the bathtub 4 due to heat dissipation before the user uses the bathtub 4.

[0075] In S70, the kitchen remote control 12 compares the current time with the reduction period. If the current time has not yet reached the reduction period (NO in S70), it returns to S60. If the kitchen remote control 12 has reached the reduction period (YES in S70), it proceeds to S80.

[0076] In S80, the kitchen remote control 12 monitors whether it has received the above-mentioned bath filling instruction. If the kitchen remote control 12 has received the bath filling instruction (YES in S80), it proceeds to S82. In addition, if the kitchen remote control 12 receives the bath filling instruction from the user via the bathroom remote control 14, it also determines YES in S80 and proceeds to S82. If the kitchen remote control 12 has not received the bath filling instruction (NO in S80), it proceeds to S90.

[0077] In S90, the kitchen remote control 12 monitors whether it has received the above-mentioned reheat command. If the kitchen remote control 12 does not receive the reheat command (NO in S90), it skips S92 and proceeds to S100. If the kitchen remote control 12 does receive the reheat command (YES in S90), it proceeds to S92. In addition, if the kitchen remote control 12 receives a reheat command from the user via the bathroom remote control 14, it also determines YES in S90 and proceeds to S92.

[0078] In S92, the kitchen remote control 12 ignites the heating burner 32H of the gas heat source unit 30 and activates the bathtub water circulation pump (not shown) described above. As a result, the water in the bathtub 4 passes through the bathtub water circulation circuit and the reheating heat exchanger of the heating circuit, and is heated by the heat of the heating water heated by the heating burner 32H. At this time, the kitchen remote control 12 stops the operation of the HP unit 22 of the HP heat source unit 20. This allows the reheating operation of the bathtub 4 water to be performed using the gas heat source unit 30 while the HP heat source unit 20 is stopped, when a reheating instruction is received during the reduction period (YES in S70, YES in S90). Therefore, it is possible to prevent the HP heat source unit 20 from operating and increasing power consumption during the reduction period, and contribute to the realization of demand response.

[0079] If a bath filling command is received (YES in S80), the kitchen remote control 12 executes the bath filling operation in S82. At this time, the kitchen remote control 12 stops the operation of the HP heat source unit 20 and stops heating the tank 25 by the HP unit 22. Furthermore, the kitchen remote control 12 heats the water from the tank 25 with the hot water burner 32S of the gas heat source unit 30 and supplies it to the bathtub 4. This makes it possible to perform the bath filling operation using the gas heat source unit 30 while keeping the HP heat source unit 20 stopped when a bath filling command is received during the reduction period (YES in S70, YES in S80). Therefore, it is possible to prevent the HP heat source unit 20 from operating and increasing power consumption during the reduction period, and contribute to the realization of demand response.

[0080] In S100, the kitchen remote control 12 determines whether the reduction period has ended. The kitchen remote control 12 repeats the processes from S80 to S92 until the reduction period ends, and when the reduction period ends (YES in S100), it terminates the process shown in Figure 7.

[0081] (Effects of this embodiment) As described above, in the power consumption prediction system 100 of this embodiment, the learning device 60, which has already performed the learning process shown in Figure 4, is input with prediction condition information E12, which includes the predicted maximum temperature TH2 and predicted minimum temperature TL2 on the prediction day; the preceding period information E13, which includes the cumulative gas consumption C13 and total power consumption SE13 during the measurement period of the preceding period corresponding to the prediction period; and prediction statistical information E14, which includes the item mean and item variance (S50 in Figure 5), to calculate the predicted power consumption FE1 for the prediction period. Therefore, the predicted power consumption FE1 can be predicted more accurately than in the conventional technology. As a result, for example, the accurate predicted power consumption can be transmitted to the higher-level management server 70 (S54 in Figure 5), so that the higher-level management server 70 can make a more appropriate decision on whether or not to send a reduction instruction. This can contribute to the realization of demand response.

[0082] Furthermore, by inputting predictive statistical information E14, which includes item mean and item variance, into the learning device 60, it is possible to reduce the influence of extremely high or low values ​​in the operating information H1 and calculate the predicted power consumption FE1 more accurately.

[0083] Furthermore, the prediction condition information E12 includes weekly information, and the preceding period information E13 includes on-time, bath filling time, heat retention time, combustion time, and boiling time. For example, if the prediction period is winter, the operating time of the hot water heating system 10 tends to increase, while if the prediction period is summer, the operating time of the hot water heating system 10 tends to decrease. In other words, the energy consumption of the hot water heating system 10 may change depending on the week of the prediction period, that is, which week of the year the prediction period falls into. Also, if the on-time of the kitchen remote control 12 increases during the same measurement period as the prediction period in the preceding period, it is presumed that the operating time of the hot water heating system 10 increased during that prediction period. Furthermore, if the reheating time increases during the same measurement period as the prediction period in the preceding period, it is presumed that the operating time of the hot water heating system 10 increased during that prediction period. In this embodiment, the learner 60 performs calculation processing to calculate the predicted power consumption based on these values ​​as well, so that the predicted power consumption can be predicted more accurately.

[0084] (Correspondence) The hot water heating system 10 is an example of "equipment." The HP heat source unit 20 is an example of an "electric heat source unit," and the gas heat source unit 30 is an example of a "combustion heat source unit." The forecast condition information E12, the previous period information E13, and the forecast statistics information E14 are examples of "energy consumption information." The forecast period is an example of the "first period," each measurement period which is the same as the forecast period of the previous period is an example of the "second period," and the forecast day is an example of the "third period." The predicted maximum temperature TH2 included in the forecast condition information E12 is an example of "climate information." The higher-level management server 70 is an example of an "external device." Weekly information is an example of "period information."

[0085] The processes S40 and S42 in Figure 5 are examples of processes executed by the "Acquisition Unit". The process S50 is an example of a process executed by the "Calculation Unit". The process S54 is an example of a process executed by the "Transmission Unit". The process S64 in Figure 7 is an example of a process executed by the "Device Control Unit".

[0086] (Modification 1) Instead of the upper management server 70, the lower management server 40 may determine whether or not a reduction in power consumption is necessary in the area surrounding the house 2. In this modification, for example, the lower management server 40 may send a reduction instruction to the house 2. In this modification, the lower management server 40 does not need to perform the process in S54 of Figure 5.

[0087] (Modification 2) The power consumption prediction system 100 does not need to include a subordinate management server 40. In that case, for example, the kitchen remote control 12 of the house 2 may perform the power consumption prediction process shown in Figure 5.

[0088] (Modification 3) The total power consumption does not have to include the third power consumption of the kitchen remote control 12, the fourth power consumption of the bathroom remote control 14, and the fifth power consumption of the dressing room remote control 16. In other words, the total power consumption only needs to be the sum of the first power consumption of the HP heat source unit 20 and the second power consumption of the gas heat source unit 30 of the hot water heating system 10.

[0089] (Modification 4) In the hot water heating system 10, instead of the kitchen remote control 12, for example, the tank unit 24 and / or gas heat source unit 30 of the HP heat source unit 20 may communicate with the lower management server 40 via the router 7 of the house 2. In that case, for example, the tank unit 24 and / or gas heat source unit 30 may perform the power consumption prediction processing shown in Figure 5.

[0090] (Modification 5) The higher-level management server 70 may, in addition to or instead of a reduction instruction, send an increase instruction to the kitchen remote control 12 via the lower-level management server 40 to actively increase the power consumption. In this case, the kitchen remote control 12 may, for example, actively operate the HP unit 22 during the increase period included in the increase instruction to, for example, heat the water in the tank 25.

[0091] (Modification 6) The kitchen remote control 12 does not have to perform the processes S60 to S64 in Figure 7. In further modifications, the kitchen remote control 12 does not have to perform the processes S80 and S82 in Figure 7, nor does it have to perform the processes S90 and S92.

[0092] (Modification 7) The forecast condition information E12 does not have to include weekly information. Also, the preceding period information E13 does not have to include boiling time and on time. In further modifications, the forecast condition information E12 may include monthly information for the forecast date instead of weekly information. In this modification, monthly information is an example of "period information". In further modifications, the forecast condition information E12 may include quarterly information indicating which quarter the forecast period is included in instead of weekly information. In this modification, quarterly information is an example of "period information".

[0093] (Modification 8) The learner 60 may be configured to retrain at predetermined intervals. For example, after the end of the prediction period, the lower management server 40 may store the actual amount of power consumed by the hot water heating system 10 during the prediction period in the memory 54. Furthermore, the actual power consumption in the memory 54 and the predicted power consumption FE1 calculated by the process in Figure 5 may be input to the learner 60, and the weights of the connections between each neuron in the neural network constructed by the learning process in Figure 4, as well as the thresholds of each neuron, may be readjusted so as to reduce the error between the two.

[0094] The technical elements described herein or in the drawings demonstrate technical usefulness individually or in various combinations, and are not limited to the combinations described in the claims at the time of filing. Furthermore, the technologies illustrated herein or in the drawings can achieve multiple objectives simultaneously, and achieving even one of these objectives constitutes technical usefulness in itself. [Explanation of Symbols]

[0095] 2: House, 4: Bathtub, 6: Faucet, 7: Router, 8: Internet, 10: Hot water heating system, 12: Kitchen remote control, 14: Bathroom remote control, 16: Dressing room remote control, 20: HP heat source unit, 22: HP unit, 24: Tank unit, 25: Tank, 30: Gas heat source unit, 32H: Heating burner, 32S: Hot water burner, 34: Panel heater, 36: Bathroom heater, 38: Floor heating, 40: Lower management server, 42: Communication I / F, 46: Display unit, 48: Operation unit, 50: Control unit, 52: CPU, 54: Memory 60: Learning device, 70: Higher-level management server, 72: Aggregator, 100: Power consumption prediction system, D1: Learning result data, E1: Learning information, E2: Measurement condition information, E3: Input period information, E4: Statistical information, E12: Prediction condition information, E13: Previous period information, E14: Prediction statistical information, FE1: Predicted power consumption, H1: Operation information, P1: Installed equipment information, S1: First power sensor, S2: Second power sensor, S3: Gas sensor, S4: Third power sensor, S5: Fourth power sensor, S6: Fifth power sensor, T1: Actual information

Claims

1. A power consumption prediction system that predicts the power consumption of equipment, The aforementioned equipment comprises an electric heat source and a combustion heat source, The aforementioned power consumption prediction system is An acquisition unit that acquires energy consumption information relating to the energy consumption of the aforementioned equipment, A calculation unit calculates the predicted amount of power consumption for the first period from the present time onward, based on the energy consumption information acquired by the acquisition unit. Equipped with, The aforementioned energy consumption information is, The amount of power consumed by the equipment during the second period, which corresponds to the first period and is prior to the present time, The amount of gas consumed by the combustion-type heat source during the second period, Climate information relating to the climate near the equipment during the third period including the first period, It includes at least, The calculation unit calculates the predicted power consumption by executing calculation processing of a trained learner that has performed machine learning to calculate the predicted power consumption during the first period based on the energy consumption information. Power consumption prediction system.

2. The power consumption prediction system according to claim 1, further comprising a transmission unit that transmits the predicted power consumption amount calculated by the calculation unit to an external device.

3. The power consumption prediction system further comprises a server capable of communicating with the equipment, The server comprises the acquisition unit and the calculation unit, The power consumption prediction system according to claim 1.

4. The power consumption prediction system according to claim 3, wherein the server further comprises a transmission unit that transmits the predicted power consumption amount calculated by the calculation unit to an external device.

5. The power consumption prediction system according to claim 2 or 4, further comprising an equipment control unit that controls the equipment based on an external instruction received from the external device that has received the predicted power consumption.

6. The power consumption prediction system according to claim 5, wherein the equipment control unit restricts the operation of the equipment during a predetermined reduction period when the external instruction includes a reduction instruction for reducing the power consumption of the equipment during a predetermined reduction period.

7. The aforementioned equipment and machinery are The aforementioned equipment can perform a hot water filling operation by supplying heated water to the bathtub. The equipment is configured to allow specifying the time when the hot water filling operation will be completed before the hot water filling operation starts. The equipment control unit modifies the time to start the hot water filling operation so that the hot water filling operation is completed before the reduction period if the period during which the equipment performs the hot water filling operation in order to complete the hot water filling operation at the specified hot water filling completion time is included in the reduction period. The power consumption prediction system according to claim 6.

8. The aforementioned equipment is capable of performing a hot water filling operation, which supplies water heated by the equipment to the bathtub. When the equipment control unit receives a hot water filling instruction to perform the hot water filling operation during the reduction period, it stops the operation of the electric heat source and supplies the water heated by the combustion heat source to the bathtub during the hot water filling operation. The power consumption prediction system according to claim 6.

9. The aforementioned equipment is capable of performing a reheating operation to heat the water in the bathtub. When the equipment control unit receives a reheating instruction to perform the reheating operation during the reduction period, it stops the operation of the electric heat source and heats the water in the bathtub with the combustion heat source during the reheating operation. The power consumption prediction system according to claim 6.

10. The aforementioned equipment further comprises a remote control that is communicatively connected to the electric heat source and the combustion heat source, The aforementioned electric heat source is configured to be able to heat the water in the tank, The aforementioned energy consumption information is further, When one year is divided into multiple unit periods, period information indicating which unit period the first period is included in, In the second period, the boiling time information relating to the time it took for the electric heat source to boil the water in the tank, On-time information relating to the time the remote control was powered on during the second period, Further including, The power consumption prediction system according to claim 1.

11. A power consumption prediction method for predicting the power consumption of equipment, The aforementioned equipment comprises an electric heat source and a combustion heat source, The aforementioned power consumption prediction method is A process for acquiring energy consumption information related to the energy consumption of the aforementioned equipment, A calculation step that calculates the predicted amount of power consumption in the first period from the present time onward, based on the energy consumption information obtained by the acquisition step, Equipped with, The aforementioned energy consumption information is, The amount of power consumed by the equipment during the second period, which corresponds to the first period and is prior to the present time, The amount of gas consumed by the combustion-type heat source during the second period, Climate information relating to the climate near the equipment during the third period including the first period, It includes at least, In the calculation step, the predicted power consumption is calculated by executing the calculation process of a trained learner that has performed machine learning to calculate the predicted power consumption during the first period based on the energy consumption information. Method for predicting power consumption.

12. A learning device capable of calculating the predicted power consumption of equipment, The aforementioned equipment comprises an electric heat source and a combustion heat source, The learning device is An acquisition unit that acquires energy consumption information relating to the energy consumption of the aforementioned equipment, A learning unit performs machine learning on a learning device to calculate the predicted power consumption of the equipment during the first period from the present time onward, based on the energy consumption information acquired by the acquisition unit. Equipped with, The aforementioned energy consumption information is, The amount of power consumed by the equipment during the second period, which corresponds to the first period and is prior to the present time, The amount of gas consumed by the combustion-type heat source during the second period, Climate information relating to the climate near the equipment during the third period including the first period, including at least, Learning device.

13. A learning method for calculating the predicted power consumption of equipment, The aforementioned equipment comprises an electric heat source and a combustion heat source, The aforementioned learning method is A process for acquiring energy consumption information related to the energy consumption of the aforementioned equipment, A learning step in which the learning device performs machine learning to calculate the predicted power consumption of the equipment in the first period from the present time based on the energy consumption information obtained by the acquisition step, Equipped with, The aforementioned energy consumption information is, The amount of power consumed by the equipment during the second period, which corresponds to the first period and is prior to the present time, The amount of gas consumed by the combustion-type heat source during the second period, Climate information relating to the climate near the equipment during the third period including the first period, including at least, Learning methods.

Images