JP2002022278A - Hot water feeding device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a hot water feeding device in which the most suitable temperature is ratiocinated in response to a hot water feeding state data by a ratiocinating means 32 comprised of a neural network and hot water is supplied at the most suitable temperature. SOLUTION: A ratiocinating means 32 for use in ratiocinating the most suitable temperature is comprised of a neural network including an input layer 41, a middle layer 42 and an output layer 43. State data concerning hot water feeding operation is inputted to the input layer 41, thereby the most suitable temperature is ratiocinated. In addition, when the hot water feeding is performed not in compliance with a ratiocinated temperature but its temperature is specified by a hot water temperature specifying means 9, the state data when the specified temperature is specified is stored as indicated data 51, 52 and a studying of weighting of the ratiocinating means 32 is performed under a timing of turned-off state of the power supply, for example.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a water heater that detects a situation surrounding a water heater, infers an optimum water temperature from the situation, and supplies hot water at that temperature.

[0002]

2. Description of the Related Art Heretofore, as this kind of invention, there has been known an apparatus in which the temperature of hot water provided by a water heater is determined in advance, and the hot water is supplied at the temperature when hot water is supplied.

[0003] As for the shower, Japanese Patent Laid-Open Publication No.
There is an invention described in Japanese Patent No. 48118. The present invention relates to a shower water heater, comprising: an environmental temperature for providing a shower; a timer for generating a calendar signal from a current date and a time zone; and a control rule storage for storing a relationship between these. Therefore, it is intended to set the hot water temperature by performing control inference based on this rule. ADVANTAGE OF THE INVENTION According to this invention, the hot water of the temperature which a user feels comfortable regardless of a season can be provided, and it is very convenient.

[0004]

However, although such a convenient water heater for a shower has conventionally existed,
In the related art, the relationship between the environmental temperature and the calendar signal of the current month, day, and time zone had to be created in advance by a predetermined rule, and it was very troublesome to create the rule.

Further, in the case of optimizing the hot water temperature by creating such rules, when the factors to be further considered increase, the burden increases exponentially, and the operation time for determining the hot water temperature is increased. Is also greatly increased. In fact, the conditions for determining the temperature that a user of a water heater would want are of even more concern.

Further, if a rule is created in advance as in the prior art, it is not possible to automatically set the temperature according to the user's individual preference, and the method must be performed uniformly regardless of the user's preference. could not. Further, even when the user's preference changes, the optimum temperature cannot be automatically changed accordingly.

Accordingly, the present invention has been made in view of such circumstances. First, the inference means is configured using a neural network, and it is not necessary to limit the number of elements to be considered unnecessarily. An object of the present invention is to provide a water heater capable of inferring an optimum water temperature based on conditions suitable for a situation surrounding a water heater and reliably providing hot water at an optimum temperature.

Further, in the present invention, when the user of the water heater specifies the temperature provided by the water heater, learning of the neural network is performed using the specified temperature and the state of the situation element at that time as teaching data. It is possible to infer at the optimum temperature corresponding to the individual preference of the user, and even when the preference of the user gradually changes, the inference of the optimal temperature automatically changes by learning according to the preference. Therefore, an object of the present invention is to provide a water heater that can supply hot water at an optimum temperature according to a user's individual preference or a change in the preference.

According to the present invention, a plurality of locations such as a bath, a shower, a kitchen, and a washroom are provided.
Each of the optimum temperatures can be estimated, and the weighting coefficient of the inference means corresponding to a location having a high possibility of use is set in advance. Therefore, when hot water is supplied, the optimal temperature corresponding to the hot water supply location is immediately inferred. It is another object of the present invention to provide a water heater capable of supplying hot water at the optimum temperature.

[0010]

In order to achieve the above object, according to the present invention, a situation detecting means for detecting a situation to be supplied with hot water, and a situation detected by the situation detecting means as input data. The system comprises inference means which is a neural network for inferring and outputting information about a temperature that a user of the water heater would want, and hot water supply control means for supplying hot water based on the content inferred by the inference means. It is characterized by the following.

Further, the present invention provides a hot water supply system in which hot water is supplied at a temperature specified by a hot water temperature specifying means for specifying a hot water temperature, or the temperature is automatically set without using the hot water temperature specifying means. A water heater, a situation detecting means for detecting a situation when hot water is supplied, and an inference means for inferring information about a temperature that a user of the water heater would want as a situation detected by the situation detecting means as input data, Hot water supply control means for supplying hot water based on the content inferred by the inference means, a temperature specified by the hot water user by the hot water temperature specifying means, and a situation detected by the situation detecting means at that time And learning means for learning the inference means using the above as a teaching signal. That is, the invention according to claim 2 provides a hot water temperature designation means for adjusting the temperature of hot water provided from a water heater by a user, an automatic designation means for inferring an optimum temperature by inference and automatically setting the temperature, and a hot water supply. When the temperature of the hot water is designated by the hot water temperature designation means, the hot water supply control means performs hot water supply control to supply hot water at the designated temperature, and is automatically controlled by the automatic designation means. At the time of designation, the hot water supply control means infers information relating to a temperature that a user of the water heater would want as a situation detection means for detecting a situation to be supplied with hot water, and the situation detected by the situation detection means as input data. The hot water supply is controlled based on the information inferred based on the inference means which is a neural network that outputs the hot water. According to the third aspect of the present invention, a hot water temperature specifying means for a user to adjust the temperature of hot water provided from a water heater, an automatic specifying means for inferring an optimum temperature by inference and automatically setting the same, and controlling hot water supply Hot water supply control means, and when the temperature of the hot water is designated by the hot water temperature designation means, the hot water supply control means performs hot water supply control to supply hot water at the designated temperature, and is automatically designated by the automatic designation means. In the case of, the hot water supply control means infers information relating to a temperature which a user of the water heater would want as a situation detection means for detecting a situation in which hot water should be supplied, and a situation detected by the situation detection means as input data. The supply of hot water is controlled based on the information inferred based on the inference means consisting of a neural network to be output, and further, the temperature when the hot water is specified by the hot water specifying means and the situation at that time As teaching signal and a situation where the detection means detects, characterized by comprising a learning means for performing learning of the inference means.

According to a fourth aspect of the present invention, the water heater is any one of a water heater that supplies hot water to a kitchen, a water heater that supplies hot water to a bathroom, and a water heater that supplies hot water to a washroom. There is a feature.

According to a fifth aspect of the present invention, a first hot water supply system is provided.
A hot water supply point, and at least two or more second hot water supply points for supplying hot water, a situation detecting means for detecting a situation of hot water supply, a neuron unit of an input layer and an output layer A plurality of neuron units, and a synapse having a weighting coefficient connects the neuron units to each other, and inputs the detection data of the situation detecting means to an input layer, thereby providing a user of the water heater. Inference means comprising a neural network that outputs information on a hot water supply temperature that would be desired from an output layer; data storage means for storing a weight coefficient of a synapse of the inference means corresponding to each of the hot water supply locations; Hot water supply possibility determining means for determining the possibility of supplying hot water to the hot water supply point, and the hot water supply possibility determining means First weighting factor setting means for setting the weighting coefficient stored in the data storage means to the inference means; and hot water supply points and actual hot water supply points determined by the hot water supply possibility determining means during hot water supply. And a second weighting factor setting unit for resetting the weighting factor of the inference unit to a weighting factor corresponding to an actual hot water supply location when the two do not match, and inferring by the inference unit. And hot water supply control means for supplying hot water at a temperature according to the obtained information.

According to a sixth aspect of the present invention, a first hot water supply system is provided.
A water supply unit having at least two hot water supply points, a second hot water supply point for supplying hot water, a state detection means for detecting a state of hot water supply, a neuron unit of an input layer and an output layer A plurality of neuron units between the neuron units, a synapse having a weighting factor connects the neuron units, and inputs the detection data of the situation detecting means to an input layer, and Inference means consisting of a neural network that outputs information on a hot water supply temperature that a user would desire from an output layer, and data storage means for storing a weight coefficient of a synapse of the inference means in correspondence with each of the hot water supply locations, Weighting factor setting means for setting the weighting factor stored in the data storage means to the inference means; and By sequentially setting the weighting coefficient of each hot water supply point stored in the stage to the inference means, information about the hot water supply temperature that the user of the water heater corresponding to each hot water supply point would like is inferred in advance, When hot water is supplied, hot water supply control means for supplying hot water based on the inferred contents corresponding to the hot water supply location is provided.

[0015]

The inference means 32 for inferring the optimum temperature uses a neural network. The input layer 41 inputs user information of a water heater and situation data such as detection data of a sensor for detecting a situation of a hot water supply point. And infer the optimal temperature.

When the temperature is designated by the hot water temperature designation means 9 without depending on the inferred automatic setting of the optimum temperature, the designated temperature at that time, status data and the like are stored.
For example, the weight of the inference means 32 is learned based on the power-off timing.

In the case of a water heater that supplies hot water to a plurality of locations, where the inferred temperature range is different for each hot water supply location, the weight of the inference means 32 corresponding to the hot water supply location is stored. A weight corresponding to the hot water supply location having the highest availability in the availability determining means 27 is set in the inference means 32 in advance. Alternatively, the optimum temperature corresponding to each hot water supply point is inferred and set in advance at a predetermined cycle or at a predetermined timing.

[0018]

Embodiments of the present invention will be described below with reference to the drawings. First, FIG. 1 shows an example of use of a water heater to which the present invention is applied.

The water heater 1 can supply hot water to a kitchen, a washroom, a shower room, a bathroom, a living room, a bedroom, or the like. Instructions regarding hot water supply, such as hot water temperature, can be given from the main remote controller 2 in the kitchen, the flow remote controller 3 in the bathroom, and the sub remote controllers 4 and 5 in other places.

Next, the configuration of the present invention will be described with reference to the block diagram of FIG. Control device 21 for controlling water heater 1 and the like
The control device 21 has a main remote controller 2, a flow remote controller 3, and sub remote controllers 4, 5 through a transmission line.
Is connected, and the water heater 1 can be remotely controlled. Then, in accordance with the operation instruction, control device 21 controls hot water supply device 1 to supply hot water to hot water supply point 18 such as a bathroom.

Each hot water supply location is provided with a detection sensor 19 for detecting the situation at that location. Further, an inference means 32 is connected to the control device 21.
The optimum temperature inferred in 2 is instructed to the control device 21. Further, the inference means 32 is provided with a learning means 34, a teaching means 35, and a data storage means 33 for reliably inferring the optimum temperature.

Next, each component will be described in detail. First, the remote controller will be described.
The remote controller 3 and the sub remote controllers 4 and 5
1. It is connected to the inference means 32 and the data storage means 33, instructs the control device 21 to control and transmits and stores the instruction contents to the data storage means 33 as teaching data for learning the inference means 32. Further, the inference means 32
Indicates necessary information for inference as necessary.

FIG. 3 shows an external view of the main remote controller 2 provided in the kitchen. Above the main remote controller 2, there is a display means 6 for displaying control contents such as the temperature of hot water supplied by the water heater 1. Below that is the operation unit 7,
The operation of the water heater 1 is performed by operating the operation unit 7. The operation unit 7 has a power switch 8, hot water temperature designation means 9, automatic designation means 10, mode designation means 11 and 12, and other means for setting reservations and the like.

The hot water temperature designation means 9 is for the user to adjust the temperature of the hot water provided from the water heater 1. Automatic designation means 1
0 indicates that the optimum temperature inferred by the inference means 32 is automatically set.

The mode designating means 11 and 12 are the inference means 32
When the inference is performed by using the learning mode or when the learning is performed by the learning unit 34, the process is performed in an individual mode. There are two modes, A to D and SP, where A to D are specified by the user of the water heater 1 as personal information. As for the temperature of the bath, the temperature of the shower, and the like, since the user has individual preferences, the user specifies who uses the temperature.

For example, in the case of a family of four, a father, mother, eldest son, and eldest daughter, the father is determined in advance as A, the mother as B, the eldest son as C, and the eldest daughter as D. Since the pressing of the D button is input to the inference means 32 as the situation data, for example, when the father uses the A button, the A button is pressed to infer the optimum temperature corresponding to the individual preference corresponding to the person of the A button. Is available.

On the other hand, the SP mode specifies a different temperature range. Usually, the temperature range inferred by the inference means 32 is in the range of about 30 ° C to 50 ° C. At other high or low temperatures, the mode of the hot water temperature designation means 9 or SP is used.

In the SP mode, a temperature range is designated in advance by a designation unit (not shown), and the optimum temperature is inferred from the contents learned by the learning unit 34 using the use state in the range as teaching data. For example, when it is desired to use at a lower temperature of cold water frequently used in a shower, setting the temperature range of the SP mode to 15 ° C. to 35 ° C. enables
If the P mode is specified, the optimum temperature is inferred in the range of 15 ° C. to 35 ° C. using the usage state in that range as teaching data.

Next, FIG. 4 shows an external view of the remote controller 3 provided in the bath. The flow remote controller 3 is capable of giving an instruction from the bathroom to the water heater 1 and displaying the contents thereof.

In addition, although not shown, the sub remote controllers 4 and 5 provided at other hot water supply points such as a shower room can perform operation instructions and display corresponding to the hot water supply points.

Next, the control device 21 will be described. The control device 21 includes a storage unit 22 and a control unit 23. The control unit 23 includes an operation control unit 24, a hot water supply control unit 25, a clock unit 26, a usability determination unit 27, a weight coefficient setting unit 28, a delay Means 29 and other control means 30.

The storage means 22 stores control procedures and operation procedures, and contents necessary for controlling the water heater 1, such as temporary storage of transmission data. The operation control unit 24 controls signals with the operation units of the remote controllers 2 to 5. The hot water supply control means 25 controls a burner, a valve and the like of the water heater 1 in accordance with the content of the operation instruction and the like.

The clock means 26 is a function of a calendar of month, day and time and a clock. The availability determining means 27 determines which hot water supply location, such as a kitchen, a washroom, or a bathroom, is most likely to be used, based on the usage status up to now.

The weighting factor setting means 28 includes an inference means 32 corresponding to the hot water supply location stored in the data storage means 33.
Is set in the inference means 32. The weight coefficient setting means 28 constitutes first and second weight coefficient setting means.

The delay means 29 is for delaying a predetermined time at a predetermined timing, for example, when the power switch is turned off, and is for ensuring that the learning means 34 learns when the power is turned off. is there. Other control means 3
0 controls other operations, for example, controls data exchange between the storage unit 22 and the data storage unit 33.

Next, the inference means 32 will be described. The inference means 32 infers the optimum temperature of hot water from detection data of the sensor 19 for detecting the state of the hot water supply location, the contents of the clock means 26, user information of the water heater 1, and the like.
FIG. 5 shows specific contents.

As shown in FIG. 5, the inference means 32 comprises an input layer 41, an intermediate layer 42, and an output layer 43. The input layer 41 has a plurality of neuron units I1 to Ii.
2 has neuron units M1 to Mj, and the output layer 43
Has a plurality of neuron units O1 to Ok.

Then, the neuron unit I1 of the input layer
ＩIi and the neuron units M1 to Mj in the intermediate layer have a synapse A1 having a weighting factor.
1 to Aij, and the neuron units M1 to Mj in the intermediate layer and the neuron units O1 to O in the output layer.
k are connected by synapses B11 to Bjk having weight coefficients. Since the neuron unit and the synapse itself are already known, the detailed description will be omitted.

Each of the neuron units I1 to I1 of the input layer 41
Information for inference is input to Ii, user information (unspecified or A to D) for unit I1, month and day information for unit I2, time information for unit I3,
The temperature information of the hot water supply point is input to the unit I4, the outside air temperature is input to the unit I5, the humidity of the hot water supply point is input to the unit I6, and other environmental information is input to the units I7 to Ii.
These constitute the situation data for inference.

When these pieces of information are input to the units of the input layer 41, synapses between the input layer 41 and the intermediate layer 42,
Signals are transmitted to the units of the output layer 43 through the units M1 to Mj of the intermediate layer 42 and the synapse between the intermediate layer 42 and the output layer 43, and the units O1 to O of the output layer 43 are transmitted.
From k, a signal is output to the inferred temperature determination means 44. The units O1 to Ok of the output layer 43 correspond to each temperature.
The temperature range inferred by the inference means 32 is, for example, 36 ° C. to 45 ° C. in the case of a bath.
Each unit is shifted by 1 ° C. so that the unit O10 is 36 ° C., such as 5 ° C., the unit O2 is 44 ° C., and the unit O3 is 43 ° C.

In addition, in the case of a shower room, the inferred temperature range is 30 ° C. to 45 ° C., and in the kitchen, it is 20 ° C. to 6 ° C.
0 ° C., 25 ° C. to 50 ° C. in the washroom, and the range specified by the input means (not shown) in the SP mode. In these cases, since there are ten units of the output layer 43, the temperature of each hot water supply point The range from high temperature to low temperature corresponds to units O1 to O10 by dividing the range into ten equal parts. The switching of the temperature range is performed based on a signal from the control unit 23. As a result, it is possible to reliably infer in a necessary temperature range corresponding to the hot water supply location.

The inferred temperature judging means 44 infers from the output states of the units O1 to O10 of the output layer 43 that the temperature corresponding to the unit having the highest output level is usually the optimum temperature.

Next, the data storage means 33 will be described with reference to FIG. The contents stored in the data storage means 33 include synapse weighting coefficient data 50, kitchen,
The teaching data 51 corresponds to a hot water supply location such as a bathroom, and the teaching data 52 when the SP switch is designated. The synapse weighting coefficient data 50 stores weighting coefficients (A11 to Aij, B11 to Bjk) for coupling between neuron units, corresponding to each hot water supply location and SP mode.

Therefore, when inferring by the inference means 32,
Inference can be made using a weight coefficient corresponding to a hot water supply location or the SP mode.

The teaching data 51 corresponding to the hot water supply location is teaching data for learning by the learning means 34, and is used when the temperature of hot water to be supplied is specified by the hot water temperature specifying means 9 of the remote controllers 2 to 5. It stores status data such as the content of the clock means 26 corresponding to the designated temperature and the detection status of the sensor 19. The SP teaching data 52 stores situation data corresponding to the designated temperature designated by the hot water temperature designation means 9 when this mode is designated by the SP switch.

Next, the learning means 34 performs learning based on the teaching signal from the teaching means 35, and determines the weight coefficient of the inference means 32. The teaching means 35 supplies signals relating to the situation data and the designated temperature as teaching data corresponding to each hot water supply point.

Then, the learning means 34 performs learning by the back propagation method. That is, first, data is input by the teaching means 35, the result is output, and the degree of coupling of the weight coefficients between neuron units is changed so as to reduce the error of the result. Then, input data is input again, and the same operation is repeated until convergence. Since the learning itself by the back propagation method is already known, the details will be omitted.

Next, the operation of inferring the optimum temperature corresponding to the hot water supply point according to the present invention will be described with reference to FIG. This is an operation to be performed when the automatic designating means 10 is designated. When the automatic designation means 10 is not designated, inference is not performed and only the temperature designation by the hot water temperature designation means 9 is accepted.

First, in order to prepare for hot water supply, the usability determining means 27 determines which part of the kitchen, bathroom, shower room, living room, etc. is highly usable (S1). This is performed by comprehensive evaluation of the usage probability based on the time period up to now, and the usage probability by detecting each operation such as turning on / off the lighting switch in a bathroom or a kitchen.

Then, the weight coefficient stored in the weight data 50 of the data storage unit 33 corresponding to the hot water supply point with the highest availability is set in the inference unit 32 in advance (S2). Then, it is determined whether or not the hot water temperature is specified by the hot water temperature specifying means using remote controllers 2 to 5 (S3). If not, it is determined whether or not there is an instruction to start hot water supply (S4). . If there is none, steps S1 to S4 are repeated.

Here, when there is an instruction to start hot water supply in step S4, it is determined whether the SP switch 12 has been pressed and the SP mode has been designated (S5). If not specified, it is determined whether the type of weight set in the inference means 32 matches the hot water supply location (S6, S7), and if they do match, the situation data detected from the sensor 19 or the like at that time. The inference means 32 infers the optimal temperature from (S9).

If they do not match in step S7, the weight coefficient corresponding to the hot water supply point is changed and set (S8), and the inference means 32 infers the optimum temperature from the situation data detected from the sensor 19 and the like at that time. (S9).
If the SP switch 12 has been pressed in step S5, the weight coefficient stored in the weight coefficient data of the data storage unit 33 is changed and set (S8), and the optimum temperature is inferred (S9).

Then, the optimum temperature inferred by the inference means 32 is set as the hot water supply temperature (S10), and hot water is supplied to the corresponding hot water supply location (S15).

On the other hand, in step S3,
If the hot water temperature is specified in 5, the sensor 1 corresponding to the specified temperature and the data necessary for inferring at that time
The status data such as 9 is acquired (S11). Then, a hot water supply point corresponding to the designated temperature is determined (S12), and the data is stored in the data storage means 33 as teaching data (S13), and stored as data for learning. Then, the designated temperature is set as the hot water supply temperature (S1).
4) Then, hot water is supplied to the corresponding location at that temperature (S15).
When the hot water supply ends (S16), the process returns to step S1, and the same operation is repeated.

Next, the operation of learning the weight coefficient will be described with reference to FIG. Since the learning is performed when the hot water supply ends and the power is turned off, first, it is determined whether or not the power switch 8 is turned off (S31). When the switch is turned off, the delay unit 29 operates (S32), the time required for learning, the control unit 21, the inference unit 32, the data storage unit 33, and the learning unit 3
4. Power is supplied to the teaching means 35. Then, it is determined whether there is new teaching data in the data storage means 33 (S33). If there is no new teaching data, there is no need for learning, and all power is turned off and the process ends (S4).
9).

If there is new teaching data, learning is performed in the order of the kitchen weight factor, the bathroom weight factor, the shower room weight factor, the wash surface, the living room weight factor, and the SP mode weight factor.

First, it is determined whether or not there is a kitchen in the teaching data 51 (S34). If there is, the optimum temperature is inferred by the inference means 32 from the situation data at that time, and if there is a match. It is determined whether or not to perform (S35). If there is at least one mismatch, learning of the kitchen weight is performed, and the weighting factor data 50 is calculated based on the weighting factor of the learning result.
(S36). Learning is performed by the back propagation method described above. Step S34
If there is no teaching data for kitchen or if the designated temperature and the inferred temperature are all the same in step S35, there is no need for learning, and the process proceeds to the learning of the next bathroom weighting coefficient.

Hereinafter, learning of the weight coefficient for the bathroom (S37-
S39), learning of weight coefficient for shower room (S40)
-S42), learning of weighting factors for the washroom, living room, etc. (S43-
S45), learning of SP mode weighting coefficients (S46 to S46)
48) is performed in the same manner as in the case of the learning of the kitchen weight coefficient described above. If all the operations are completed, all the power is turned off and the operation is completed (S49).

Next, another embodiment will be described with reference to FIG. As shown in FIG. 9, the inference means 32 includes an input layer 61,
It is composed of an intermediate layer 62 and an output layer 63, and is connected by a synapse having a weight coefficient between the neuron unit of the input layer and the neuron unit of the intermediate layer, and between the neuron unit of the intermediate layer and the neuron unit of the output layer. .

Then, hot water supply location information is added to the input information itself input to the input layer 61. Therefore, the synapse weighting coefficient data 50 for storing the weighting coefficient corresponding to each hot water supply point, which is required in the embodiment, becomes unnecessary, and the usability determining means 27 and the weighting coefficient setting means 28
Without providing a heater, the optimum temperature corresponding to the hot water supply location can be inferred, and the configuration can be simplified. However, in this case, it is desirable to make the temperature range of the output layer 63 of the inference means 32 common.

In the above embodiment, the optimum temperature is inferred by the inference means 32 at the start of hot water supply. This is to make inferences with solid situation data,
The optimum temperature of each hot water supply point may be inferred by the inference means 32 at a constant cycle, and the hot water may be provided at the previously inferred temperature during hot water supply or during hot water supply.

The inference means 32 may make an inference at some other timing, for example, at a timing when the lighting switch at that location is turned on, and supply hot water at the previously inferred temperature when hot water is supplied. In this case,
Because there is a time lag between hot water supply and inference time, if the situation data changes during that time, it may not be possible to reliably infer the optimal temperature, but it is excellent in that hot water can be supplied more immediately. I have.

In the above embodiment, the intermediate layers 42 and 62
However, it is needless to say that two or more intermediate layers may be provided.

In the above embodiment, all of the situation data is input to the input layers 41 and 61. However, a part of the situation data is directly input to the intermediate layers 42 and 62 or the output layers 43 and 63. You may do it.

In the above-described embodiment, hot water may be provided to a plurality of locations at the same time. In this case, priorities may be determined in advance, or priority instructing means may be provided in each remote controller, and the remote controller may be provided with the instruction. The priority may be given to the hot water supply temperature at the location where the water supply was performed.

Further, in the above embodiment, the case of the hot water heater for supplying hot water to a plurality of locations has been described. However, the present invention is also applied to the case of a hot water heater used in an individual location such as a kitchen, a bathroom, a shower room and the like. Of course, it is of course possible to infer the optimal temperature. In this case, the weight coefficient data 50 of the data storage unit 33, the usability determination unit 27, and the like are unnecessary.

In the above embodiment, some data are shown as the situation data detected by the situation detecting means.
The data may be composed of a part of the data shown, or may be further added.

[0068]

As described above, in the present invention, the optimum temperature can be inferred from the situation data when hot water is provided by the inference means comprising a neural network for the temperature of hot water provided by the water heater. Therefore, there is no need to remember how many times your favorite temperature is for the situation at that time, there is no need to set the temperature each time, and you can provide hot water at the optimal temperature, It has the effect of being convenient.

According to the present invention, as the water heater is used, when the temperature is designated, the temperature and the situation data at that time are stored as teaching data, and learning of the inference means is performed based on the teaching data. So you can do
There is an effect that the learning setting can be automatically performed when the user's preference of the water heater changes or the user's individual preference.

In a water heater that supplies hot water to a plurality of locations such as a kitchen and a bath, the weight of the inference means corresponding to each hot water location is stored in the data storage means. In addition to the above, it is possible to appropriately infer, and it is possible to infer in a different temperature range for each hot water supply location.

Further, the availability determining means determines a hot water supply location which is highly likely to be supplied prior to hot water supply, and the weight of the location is set in the inference means in advance. Can be immediately inferred. Further, if the inferring means is inferred in advance at a predetermined cycle and timing, the inferred optimum temperature is set in advance, and hot water is supplied at that temperature when hot water is supplied, so that there is an effect that hot water can be supplied more immediately.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining an example of a usage example of the present invention.

FIG. 2 is a block diagram for explaining the present invention.

FIG. 3 is an external view of a main remote controller.

FIG. 4 is an external view of a remote controller.

FIG. 5 is an explanatory diagram of inference means.

FIG. 6 is an explanatory diagram of a data storage unit.

FIG. 7 is a flowchart showing an operation when inferring an optimum temperature.

FIG. 8 is a flowchart showing a learning operation based on teaching data.

FIG. 9 is an explanatory diagram of another embodiment.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 hot water heater 2 main remote controller 3 remote controller 9 hot water temperature designation means 10 automatic designation means 19 sensor 21 controller 25 hot water supply control means 26 clock means 27 availability determination means 28 weight coefficient setting means 32 inference means 33 data storage means 34 learning Means 35 Teaching means 50 Weighting coefficient data 51 Teaching data of kitchen etc. 52 Teaching data of SP

[Procedure amendment]

[Submission date] April 26, 2001 (2001.4.2
6)

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Full text

[Correction method] Change

[Correction contents]

[Document Name] Statement

[Title of the Invention] Water heater

[Claims]

1. A hot water temperature designation means for a user to adjust the temperature of hot water provided from a water heater, an automatic designation means for inferring an optimum temperature by inference and automatically setting the same, and a hot water supply control for controlling hot water supply. Means, when the temperature of hot water is designated by the hot water temperature designation means, the hot water supply control means performs hot water supply control to provide hot water at the designated temperature, and at the time of automatic designation by the automatic designation means, The hot water supply control means, a situation detection means for detecting a situation in which hot water is to be supplied,
And information inferred on the basis of the situation detected by the situation detecting means as input data and inferring means comprising a neural network for inferring and outputting information relating to a temperature that a user of the water heater would desire. Learning means for controlling the supply of hot water and for learning the inference means using the temperature designated by the hot water temperature designation means and the situation detected by the situation detection means at that time as a teaching signal; And a water heater comprising:

2. The water heater is one of a water heater for supplying hot water to a kitchen, a water heater for supplying hot water to a bathroom, and a water heater for supplying hot water to a washroom. The water heater according to claim 1.

3. A water heater that has at least two hot water supply points, a first hot water supply point for supplying hot water and a second hot water supply point for supplying hot water, and detects a situation in which hot water is to be supplied. Means, further comprising one or more neuron units between the neuron unit of the input layer and the neuron unit of the output layer, wherein a synapse having a weight coefficient connects the neuron units, and the situation detection means is provided in the input layer. Inference means comprising a neural network for inputting the detection data of the above and outputting information on a hot water supply temperature that a user of the water heater would want from an output layer; and a weighting factor of a synapse of the inference means to each of the hot water supply points. Data storage means for storing the hot water supply possibility, and hot water supply possibility determination means for determining the possibility of hot water supply to which of the hot water supply locations, and the hot water supply possibility determination means The first weighting factor setting means for setting the weighting factor stored in the data storage means at the most likely location in the inference means, and the hot water supply possibility determining means at the time of hot water supply. A second weight for resetting the weighting factor of the inference means to a weighting factor corresponding to the actual hot water supply location if the hot water supply location determined in step 1 is coincident with or not matching the actual hot water supply location. A water heater comprising: coefficient setting means; and hot water supply control means for supplying hot water at a temperature according to the information inferred by the inference means.

4. A water heater having at least two hot water supply points, a first hot water supply point for supplying hot water and a second hot water supply point for supplying hot water, wherein a situation detection for detecting a situation in which hot water is to be supplied is provided. Means, and one or more neuron units between the neuron unit of the input layer and the neuron unit of the output layer, wherein a synapse having a weight coefficient connects the neuron units to each other, and The detection data of the detection means is input, the inference means which is a neural network which outputs information on a hot water supply temperature that a user of the water heater would want from an output layer, and the weight coefficient of the synapse of the inference means is set to Data storage means for storing corresponding to the hot water supply location; weight coefficient setting means for setting the weight coefficient stored in the data storage means to the inference means; By setting the weighting factor of each hot water supply point stored in the data storage means in the inference means in turn by the coefficient setting means, the hot water supply temperature that a user of the water heater corresponding to each hot water supply point would want And hot water supply control means for supplying hot water based on the inferred contents corresponding to the hot water supply location when hot water is supplied.

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a water heater that detects a situation surrounding a water heater, infers an optimum water temperature from the situation, and supplies hot water at that temperature.

[0002]

2. Description of the Related Art Heretofore, as this kind of invention, there has been known an apparatus in which the temperature of hot water provided by a water heater is determined in advance, and the hot water is supplied at the temperature when hot water is supplied.

[0003] As for the shower, Japanese Patent Laid-Open Publication No.
There is an invention described in Japanese Patent No. 48118. The present invention relates to a shower water heater, comprising: an environmental temperature for providing a shower; a timer for generating a calendar signal from a current date and a time zone; and a control rule storage for storing a relationship between these. Therefore, it is intended to set the hot water temperature by performing control inference based on this rule. ADVANTAGE OF THE INVENTION According to this invention, the hot water of the temperature which a user feels comfortable regardless of a season can be provided, and it is very convenient.

[0004]

However, although such a convenient water heater for a shower has conventionally existed,
In the related art, the relationship between the environmental temperature and the calendar signal of the current month, day, and time zone had to be created in advance by a predetermined rule, and it was very troublesome to create the rule.

Further, in the case of optimizing the hot water temperature by creating such rules, when the factors to be further considered increase, the burden increases exponentially, and the operation time for determining the hot water temperature is increased. Is also greatly increased. In fact, the conditions for determining the temperature that a user of a water heater would want are of even more concern.

Further, if a rule is created in advance as in the prior art, it is not possible to automatically set the temperature according to the user's individual preference, and the method must be performed uniformly regardless of the user's preference. could not. Further, even when the user's preference changes, the optimum temperature cannot be automatically changed accordingly.

Accordingly, the present invention has been made in view of such circumstances. First, the inference means is configured using a neural network, and it is not necessary to limit the number of elements to be considered unnecessarily. An object of the present invention is to provide a water heater capable of inferring an optimum water temperature based on conditions suitable for a situation surrounding a water heater and reliably providing hot water at an optimum temperature.

Further, in the present invention, when the user of the water heater specifies the temperature provided by the water heater, learning of the neural network is performed using the specified temperature and the state of the situation element at that time as teaching data. It is possible to infer at the optimum temperature corresponding to the individual preference of the user, and even when the preference of the user gradually changes, the inference of the optimal temperature automatically changes by learning according to the preference. Therefore, an object of the present invention is to provide a water heater that can supply hot water at an optimum temperature according to a user's individual preference or a change in the preference.

According to the present invention, a plurality of locations such as a bath, a shower, a kitchen, and a washroom are provided.
Each of the optimum temperatures can be estimated, and the weighting coefficient of the inference means corresponding to a location having a high possibility of use is set in advance. Therefore, when hot water is supplied, the optimal temperature corresponding to the hot water supply location is immediately inferred. It is another object of the present invention to provide a water heater capable of supplying hot water at the optimum temperature.

[0010]

Means for Solving the Problems In order to achieve the above object, the invention according to claim 1 is a hot water temperature designation means for adjusting the temperature of hot water provided from a water heater by a user, and an optimum temperature by inference. And automatic hot water supply control means for controlling hot water supply. When the hot water temperature is specified by the hot water temperature specifying means, the hot water supply control means Hot water supply control to provide hot water at
At the time of automatic designation by the automatic designation means, the hot water supply control means, a situation detection means for detecting a situation to be hot water supply,
And information inferred on the basis of the situation detected by the situation detecting means as input data and inferring means comprising a neural network for inferring and outputting information relating to a temperature that a user of the water heater would desire. Learning means for controlling the supply of hot water and for learning the inference means by using the temperature designated by the hot water designation means and the situation detected by the situation detection means at that time as a teaching signal; When,
It is characterized by having.

According to a second aspect of the present invention, the water heater is any one of a water heater that supplies hot water to a kitchen, a water heater that supplies hot water to a bathroom, and a water heater that supplies hot water to a washroom. There is a feature.

[0012] According to a third aspect of the present invention, a first hot water supply system is provided.
A hot water supply point, and at least two or more second hot water supply points for supplying hot water, a situation detecting means for detecting a situation of hot water supply, a neuron unit of an input layer and an output layer A plurality of neuron units, and a synapse having a weighting coefficient connects the neuron units to each other, and inputs the detection data of the situation detecting means to an input layer, thereby providing a user of the water heater. Inference means comprising a neural network that outputs information on a hot water supply temperature that would be desired from an output layer; data storage means for storing a weight coefficient of a synapse of the inference means corresponding to each of the hot water supply locations; Hot water supply possibility determining means for determining the possibility of supplying hot water to the hot water supply point, and the hot water supply possibility determining means First weighting factor setting means for setting the weighting coefficient stored in the data storage means to the inference means; and hot water supply points and actual hot water supply points determined by the hot water supply possibility determining means during hot water supply. And a second weighting factor setting unit for resetting the weighting factor of the inference unit to a weighting factor corresponding to an actual hot water supply location when the two do not match, and inferring by the inference unit. And hot water supply control means for supplying hot water at a temperature according to the obtained information.

According to a fourth aspect of the present invention, a first hot water supply system is provided.
A water supply unit having at least two hot water supply points, a second hot water supply point for supplying hot water, a state detection means for detecting a state of hot water supply, a neuron unit of an input layer and an output layer A plurality of neuron units between the neuron units, a synapse having a weighting factor connects the neuron units, and inputs the detection data of the situation detecting means to an input layer, and Inference means consisting of a neural network that outputs information on a hot water supply temperature that a user would desire from an output layer, and data storage means for storing a weight coefficient of a synapse of the inference means in correspondence with each of the hot water supply locations, Weighting factor setting means for setting the weighting factor stored in the data storage means to the inference means; and By sequentially setting the weighting coefficient of each hot water supply point stored in the stage to the inference means, information about the hot water supply temperature that the user of the water heater corresponding to each hot water supply point would like is inferred in advance, When hot water is supplied, hot water supply control means for supplying hot water based on the inferred contents corresponding to the hot water supply location is provided.

In the embodiment of the present invention, hot water is supplied.
Situation detecting means for detecting a situation to be performed, and the situation detecting means
The user of the water heater uses the situation detected in
A function to infer and output information about the temperature
Inference means comprising a neural network, and
Hot water supply control means for supplying hot water based on the inferred contents
It is desirable to have the following. Embodiments of the present invention
, The hot water temperature is specified by the hot water temperature specifying means.
Hot water at a hot temperature, or without depending on the means for specifying hot water temperature
A water heater that automatically sets the temperature and supplies hot water,
Status detection means for detecting the status of hot water supply;
The condition detected by the condition detection means is used as input data for the water heater.
Inference that infers information about the temperature that a user may want
Means and hot water based on the content inferred by the inference means.
Hot water supply control means for supplying water and a user of the water heater
The temperature at the time specified by the
The situation detected by the output means as a teaching signal,
It is desirable to have learning means for learning the steps.
Also, in the embodiment of the present invention, provided from the water heater
Hot water temperature designation means for the user to adjust the temperature of the hot water
Automatic designation means for inferring the optimal temperature and automatically setting the temperature;
Hot water supply control means for controlling hot water supply;
When the temperature of the hot water is designated by the
The control means is a hot water supply control for providing hot water at the designated temperature.
In the case of automatic designation by the automatic designation means,
The hot water supply control means detects a situation in which hot water should be supplied.
Means, and the status detected by the status detecting means are input data.
Information about the temperature that a water heater user would want as a
A neural network that infers and outputs information
Providing hot water based on information inferred based on logical means
It is desirable to control hot water supply.

[0015]

The inference means 32 for inferring the optimum temperature uses a neural network. The input layer 41 inputs user information of a water heater and situation data such as detection data of a sensor for detecting a situation of a hot water supply point. And infer the optimal temperature.

When the temperature is designated by the hot water temperature designation means 9 without depending on the inferred automatic setting of the optimum temperature, the designated temperature at that time, status data and the like are stored.
For example, the weight of the inference means 32 is learned based on the power-off timing.

In the case of a water heater that supplies hot water to a plurality of locations, where the inferred temperature range is different for each hot water supply location, the weight of the inference means 32 corresponding to the hot water supply location is stored. A weight corresponding to the hot water supply location having the highest availability in the availability determining means 27 is set in the inference means 32 in advance. Alternatively, the optimum temperature corresponding to each hot water supply point is inferred and set in advance at a predetermined cycle or at a predetermined timing.

[0018]

Embodiments of the present invention will be described below with reference to the drawings. First, FIG. 1 shows an example of use of a water heater to which the present invention is applied.

The water heater 1 can supply hot water to a kitchen, a washroom, a shower room, a bathroom, a living room, a bedroom, or the like. Instructions regarding hot water supply, such as hot water temperature, can be given from the main remote controller 2 in the kitchen, the flow remote controller 3 in the bathroom, and the sub remote controllers 4 and 5 in other places.

Next, the configuration of the present invention will be described with reference to the block diagram of FIG. Control device 21 for controlling water heater 1 and the like
The control device 21 has a main remote controller 2, a flow remote controller 3, and sub remote controllers 4, 5 through a transmission line.
Is connected, and the water heater 1 can be remotely controlled. Then, in accordance with the operation instruction, control device 21 controls hot water supply device 1 to supply hot water to hot water supply point 18 such as a bathroom.

Each hot water supply location is provided with a detection sensor 19 for detecting the situation at that location. Further, an inference means 32 is connected to the control device 21.
The optimum temperature inferred in 2 is instructed to the control device 21. Further, the inference means 32 is provided with a learning means 34, a teaching means 35, and a data storage means 33 for reliably inferring the optimum temperature.

Next, each component will be described in detail. First, the remote controller will be described.
The remote controller 3 and the sub remote controllers 4 and 5
1. It is connected to the inference means 32 and the data storage means 33, instructs the control device 21 to control and transmits and stores the instruction contents to the data storage means 33 as teaching data for learning the inference means 32. Further, the inference means 32
Indicates necessary information for inference as necessary.

FIG. 3 shows an external view of the main remote controller 2 provided in the kitchen. Above the main remote controller 2, there is a display means 6 for displaying control contents such as the temperature of hot water supplied by the water heater 1. Below that is the operation unit 7,
The operation of the water heater 1 is performed by operating the operation unit 7. The operation unit 7 has a power switch 8, hot water temperature designation means 9, automatic designation means 10, mode designation means 11 and 12, and other means for setting reservations and the like.

The hot water temperature designation means 9 is for the user to adjust the temperature of the hot water provided from the water heater 1. Automatic designation means 1
0 indicates that the optimum temperature inferred by the inference means 32 is automatically set.

The mode designating means 11 and 12 are the inference means 32
When the inference is performed by using the learning mode or when the learning is performed by the learning unit 34, the process is performed in an individual mode. There are two modes, A to D and SP, where A to D are specified by the user of the water heater 1 as personal information. As for the temperature of the bath, the temperature of the shower, and the like, since the user has individual preferences, the user specifies who uses the temperature.

For example, in the case of a family of four, a father, mother, eldest son, and eldest daughter, the father is determined in advance as A, the mother as B, the eldest son as C, and the eldest daughter as D. Since the pressing of the D button is input to the inference means 32 as the situation data, for example, when the father uses the A button, the A button is pressed to infer the optimum temperature corresponding to the individual preference corresponding to the person of the A button. Is available.

On the other hand, the SP mode specifies a different temperature range. Usually, the temperature range inferred by the inference means 32 is in the range of about 30 ° C to 50 ° C. At other high or low temperatures, the mode of the hot water temperature designation means 9 or SP is used.

In the SP mode, a temperature range is designated in advance by a designation unit (not shown), and the optimum temperature is inferred from the contents learned by the learning unit 34 using the use state in the range as teaching data. For example, when it is desired to use at a lower temperature of cold water frequently used in a shower, setting the temperature range of the SP mode to 15 ° C. to 35 ° C. enables
If the P mode is specified, the optimum temperature is inferred in the range of 15 ° C. to 35 ° C. using the usage state in that range as teaching data.

Next, FIG. 4 shows an external view of the remote controller 3 provided in the bath. The flow remote controller 3 is capable of giving an instruction from the bathroom to the water heater 1 and displaying the contents thereof.

In addition, although not shown, the sub remote controllers 4 and 5 provided at other hot water supply points such as a shower room can perform operation instructions and display corresponding to the hot water supply points.

Next, the control device 21 will be described. The control device 21 includes a storage unit 22 and a control unit 23. The control unit 23 includes an operation control unit 24, a hot water supply control unit 25, a clock unit 26, a usability determination unit 27, a weight coefficient setting unit 28, a delay Means 29 and other control means 30.

The storage means 22 stores control procedures and operation procedures, and contents necessary for controlling the water heater 1, such as temporary storage of transmission data. The operation control unit 24 controls signals with the operation units of the remote controllers 2 to 5. The hot water supply control means 25 controls a burner, a valve and the like of the water heater 1 in accordance with the content of the operation instruction and the like.

The clock means 26 is a function of a calendar of month, day and time and a clock. The availability determining means 27 determines which hot water supply location, such as a kitchen, a washroom, or a bathroom, is most likely to be used, based on the usage status up to now.

The weighting factor setting means 28 includes an inference means 32 corresponding to the hot water supply location stored in the data storage means 33.
Is set in the inference means 32. The weight coefficient setting means 28 constitutes first and second weight coefficient setting means.

The delay means 29 is for delaying a predetermined time at a predetermined timing, for example, when the power switch is turned off, and is for ensuring that the learning means 34 learns when the power is turned off. is there. Other control means 3
0 controls other operations, for example, controls data exchange between the storage unit 22 and the data storage unit 33.

Next, the inference means 32 will be described. The inference means 32 infers the optimum temperature of hot water from detection data of the sensor 19 for detecting the state of the hot water supply location, the contents of the clock means 26, user information of the water heater 1, and the like.
FIG. 5 shows specific contents.

As shown in FIG. 5, the inference means 32 comprises an input layer 41, an intermediate layer 42, and an output layer 43. The input layer 41 has a plurality of neuron units I1 to Ii.
2 has neuron units M1 to Mj, and the output layer 43
Has a plurality of neuron units O1 to Ok.

Then, the neuron unit I1 of the input layer
ＩIi and the neuron units M1 to Mj in the intermediate layer have a synapse A1 having a weighting factor.
1 to Aij, and the neuron units M1 to Mj in the intermediate layer and the neuron units O1 to O in the output layer.
k are connected by synapses B11 to Bjk having weight coefficients. Since the neuron unit and the synapse itself are already known, the detailed description will be omitted.

Each of the neuron units I1 to I1 of the input layer 41
Information for inference is input to Ii, user information (unspecified or A to D) for unit I1, month and day information for unit I2, time information for unit I3,
The temperature information of the hot water supply point is input to the unit I4, the outside air temperature is input to the unit I5, the humidity of the hot water supply point is input to the unit I6, and other environmental information is input to the units I7 to Ii.
These constitute the situation data for inference.

When these pieces of information are input to the units of the input layer 41, synapses between the input layer 41 and the intermediate layer 42,
Signals are transmitted to the units of the output layer 43 through the units M1 to Mj of the intermediate layer 42 and the synapse between the intermediate layer 42 and the output layer 43, and the units O1 to O of the output layer 43 are transmitted.
From k, a signal is output to the inferred temperature determination means 44. The units O1 to Ok of the output layer 43 correspond to each temperature.
The temperature range inferred by the inference means 32 is, for example, 36 ° C. to 45 ° C. in the case of a bath.
Each unit is shifted by 1 ° C. so that the unit O10 is 36 ° C., such as 5 ° C., the unit O2 is 44 ° C., and the unit O3 is 43 ° C.

In addition, in the case of a shower room, the inferred temperature range is 30 ° C. to 45 ° C., and in the kitchen, it is 20 ° C. to 6 ° C.
0 ° C., 25 ° C. to 50 ° C. in the washroom, and the range specified by the input means (not shown) in the SP mode. In these cases, since there are ten units of the output layer 43, the temperature of each hot water supply point The range from high temperature to low temperature corresponds to units O1 to O10 by dividing the range into ten equal parts. The switching of the temperature range is performed based on a signal from the control unit 23. As a result, it is possible to reliably infer in a necessary temperature range corresponding to the hot water supply location.

The inferred temperature judging means 44 infers from the output states of the units O1 to O10 of the output layer 43 that the temperature corresponding to the unit having the highest output level is usually the optimum temperature.

Next, the data storage means 33 will be described with reference to FIG. The contents stored in the data storage means 33 include synapse weighting coefficient data 50, kitchen,
The teaching data 51 corresponds to a hot water supply location such as a bathroom, and the teaching data 52 when the SP switch is designated. The synapse weighting coefficient data 50 stores weighting coefficients (A11 to Aij, B11 to Bjk) for coupling between neuron units, corresponding to each hot water supply location and SP mode.

Therefore, when inferring by the inference means 32,
Inference can be made using a weight coefficient corresponding to a hot water supply location or the SP mode.

The teaching data 51 corresponding to the hot water supply location is teaching data for learning by the learning means 34, and is used when the temperature of hot water to be supplied is specified by the hot water temperature specifying means 9 of the remote controllers 2 to 5. It stores status data such as the content of the clock means 26 corresponding to the designated temperature and the detection status of the sensor 19. The SP teaching data 52 stores situation data corresponding to the designated temperature designated by the hot water temperature designation means 9 when this mode is designated by the SP switch.

Next, the learning means 34 performs learning based on the teaching signal from the teaching means 35, and determines the weight coefficient of the inference means 32. The teaching means 35 supplies signals relating to the situation data and the designated temperature as teaching data corresponding to each hot water supply point.

Then, the learning means 34 performs learning by the back propagation method. That is, first, data is input by the teaching means 35, the result is output, and the degree of coupling of the weight coefficients between neuron units is changed so as to reduce the error of the result. Then, input data is input again, and the same operation is repeated until convergence. Since the learning itself by the back propagation method is already known, the details will be omitted.

Next, the operation of inferring the optimum temperature corresponding to the hot water supply point according to the present invention will be described with reference to FIG. This is an operation to be performed when the automatic designating means 10 is designated. When the automatic designation means 10 is not designated, inference is not performed and only the temperature designation by the hot water temperature designation means 9 is accepted.

First, in order to prepare for hot water supply, the usability determining means 27 determines which part of the kitchen, bathroom, shower room, living room, etc. is highly usable (S1). This is performed by comprehensive evaluation of the usage probability based on the time period up to now, and the usage probability by detecting each operation such as turning on / off the lighting switch in a bathroom or a kitchen.

Then, the weight coefficient stored in the weight data 50 of the data storage unit 33 corresponding to the hot water supply point with the highest availability is set in the inference unit 32 in advance (S2). Then, it is determined whether or not the hot water temperature is specified by the hot water temperature specifying means using remote controllers 2 to 5 (S3). If not, it is determined whether or not there is an instruction to start hot water supply (S4). . If there is none, steps S1 to S4 are repeated.

Here, when there is an instruction to start hot water supply in step S4, it is determined whether the SP switch 12 has been pressed and the SP mode has been designated (S5). If not specified, it is determined whether the type of weight set in the inference means 32 matches the hot water supply location (S6, S7), and if they do match, the situation data detected from the sensor 19 or the like at that time. The inference means 32 infers the optimal temperature from (S9).

If they do not match in step S7, the weight coefficient corresponding to the hot water supply point is changed and set (S8), and the inference means 32 infers the optimum temperature from the situation data detected from the sensor 19 and the like at that time. (S9).
If the SP switch 12 has been pressed in step S5, the weight coefficient stored in the weight coefficient data of the data storage unit 33 is changed and set (S8), and the optimum temperature is inferred (S9).

Then, the optimum temperature inferred by the inference means 32 is set as the hot water supply temperature (S10), and hot water is supplied to the corresponding hot water supply location (S15).

On the other hand, in step S3,
If the hot water temperature is specified in 5, the sensor 1 corresponding to the specified temperature and the data necessary for inferring at that time
The status data such as 9 is acquired (S11). Then, a hot water supply point corresponding to the designated temperature is determined (S12), and the data is stored in the data storage means 33 as teaching data (S13), and stored as data for learning. Then, the designated temperature is set as the hot water supply temperature (S1).
4) Then, hot water is supplied to the corresponding location at that temperature (S15).
When the hot water supply ends (S16), the process returns to step S1, and the same operation is repeated.

Next, the operation of learning the weight coefficient will be described with reference to FIG. Since the learning is performed when the hot water supply ends and the power is turned off, first, it is determined whether or not the power switch 8 is turned off (S31). When the switch is turned off, the delay unit 29 operates (S32), the time required for learning, the control unit 21, the inference unit 32, the data storage unit 33, and the learning unit 3
4. Power is supplied to the teaching means 35. Then, it is determined whether there is new teaching data in the data storage means 33 (S33). If there is no new teaching data, there is no need for learning, and all power is turned off and the process ends (S4).
9).

If there is new teaching data, learning is performed in the order of the kitchen weight factor, the bathroom weight factor, the shower room weight factor, the wash surface, the living room weight factor, and the SP mode weight factor.

First, it is determined whether or not there is a kitchen in the teaching data 51 (S34). If there is, the optimum temperature is inferred by the inference means 32 from the situation data at that time, and if there is a match. It is determined whether or not to perform (S35). If there is at least one mismatch, learning of the kitchen weight is performed, and the weighting factor data 50 is calculated based on the weighting factor of the learning result.
(S36). Learning is performed by the back propagation method described above. Step S34
If there is no teaching data for kitchen or if the designated temperature and the inferred temperature are all the same in step S35, there is no need for learning, and the process proceeds to the learning of the next bathroom weighting coefficient.

Hereinafter, learning of the weight coefficient for the bathroom (S37-
S39), learning of weight coefficient for shower room (S40)
-S42), learning of weighting factors for the washroom, living room, etc. (S43-
S45), learning of SP mode weighting coefficients (S46 to S46)
48) is performed in the same manner as in the case of the learning of the kitchen weight coefficient described above. If all the operations are completed, all the power is turned off and the operation is completed (S49).

Next, another embodiment will be described with reference to FIG. As shown in FIG. 9, the inference means 32 includes an input layer 61,
It is composed of an intermediate layer 62 and an output layer 63, and is connected by a synapse having a weight coefficient between the neuron unit of the input layer and the neuron unit of the intermediate layer, and between the neuron unit of the intermediate layer and the neuron unit of the output layer. .

Then, hot water supply location information is added to the input information itself input to the input layer 61. Therefore, the synapse weighting coefficient data 50 for storing the weighting coefficient corresponding to each hot water supply point, which is required in the embodiment, becomes unnecessary, and the usability determining means 27 and the weighting coefficient setting means 28
Without providing a heater, the optimum temperature corresponding to the hot water supply location can be inferred, and the configuration can be simplified. However, in this case, it is desirable to make the temperature range of the output layer 63 of the inference means 32 common.

In the above embodiment, the optimum temperature is inferred by the inference means 32 at the start of hot water supply. This is to make inferences with solid situation data,
The optimum temperature of each hot water supply point may be inferred by the inference means 32 at a constant cycle, and the hot water may be provided at the previously inferred temperature during hot water supply or during hot water supply.

The inference means 32 may make an inference at some other timing, for example, at a timing when the lighting switch at that location is turned on, and supply hot water at the previously inferred temperature when hot water is supplied. In this case,
Because there is a time lag between hot water supply and inference time, if the situation data changes during that time, it may not be possible to reliably infer the optimal temperature, but it is excellent in that hot water can be supplied more immediately. I have.

In the above embodiment, the intermediate layers 42 and 62
However, it is needless to say that two or more intermediate layers may be provided.

In the above embodiment, all of the situation data is input to the input layers 41 and 61. However, a part of the situation data is directly input to the intermediate layers 42 and 62 or the output layers 43 and 63. You may do it.

In the above-described embodiment, hot water may be provided to a plurality of locations at the same time. In this case, priorities may be determined in advance, or priority instructing means may be provided in each remote controller, and the remote controller may be provided with the instruction. The priority may be given to the hot water supply temperature at the location where the water supply was performed.

Further, in the above embodiment, the case of the hot water heater for supplying hot water to a plurality of locations has been described. However, the present invention is also applied to the case of a hot water heater used in an individual location such as a kitchen, a bathroom, a shower room and the like. Of course, it is of course possible to infer the optimal temperature. In this case, the weight coefficient data 50 of the data storage unit 33, the usability determination unit 27, and the like are unnecessary.

In the above embodiment, some data are shown as the situation data detected by the situation detecting means.
The data may be composed of a part of the data shown, or may be further added.

[0068]

As described above, in the present invention, the optimum temperature can be inferred from the situation data when hot water is provided by the inference means comprising a neural network for the temperature of hot water provided by the water heater. Therefore, there is no need to remember how many times your favorite temperature is for the situation at that time, there is no need to set the temperature each time, and you can provide hot water at the optimal temperature, It has the effect of being convenient.

According to the present invention, as the water heater is used, when the temperature is designated, the temperature and the situation data at that time are stored as teaching data, and learning of the inference means is performed based on the teaching data. So you can do
There is an effect that the learning setting can be automatically performed when the user's preference of the water heater changes or the user's individual preference.

In a water heater that supplies hot water to a plurality of locations such as a kitchen and a bath, the weight of the inference means corresponding to each hot water location is stored in the data storage means. In addition to the above, it is possible to appropriately infer, and it is possible to infer in a different temperature range for each hot water supply location.

Further, the availability determining means determines a hot water supply location which is highly likely to be supplied prior to hot water supply, and the weight of the location is set in the inference means in advance. Can be immediately inferred. Further, if the inferring means is inferred in advance at a predetermined cycle and timing, the inferred optimum temperature is set in advance, and hot water is supplied at that temperature when hot water is supplied, so that there is an effect that hot water can be supplied more immediately.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining an example of a usage example of the present invention.

FIG. 2 is a block diagram for explaining the present invention.

FIG. 3 is an external view of a main remote controller.

FIG. 4 is an external view of a remote controller.

FIG. 5 is an explanatory diagram of inference means.

FIG. 6 is an explanatory diagram of a data storage unit.

FIG. 7 is a flowchart showing an operation when inferring an optimum temperature.

FIG. 8 is a flowchart showing a learning operation based on teaching data.

FIG. 9 is an explanatory diagram of another embodiment.

[Description of Signs] 1 Hot water supply device 2 Main remote control 3 Flow remote control 9 Hot water temperature designation means 10 Automatic designation means 19 Sensor 21 Control device 25 Hot water supply control means 26 Clock means 27 Usability determination means 28 Weight coefficient setting means 32 Inference means 33 Data storage means 34 Learning means 35 Teaching means 50 Weight coefficient data 51 Teaching data of kitchen etc. 52 Teaching data of SP

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 3L034 DA02 5H004 GA26 GB07 HA01 HB01 JA03 JA22 JA29 KD44 LA16 LA19 MA02 MA51 MA60 5H323 AA31 BB15 CA05 CB10 DA04 EE04 FF01 FF10 KK05 LL12 LL13

Claims (6)

[Claims] 1. A situation detecting means for detecting a situation in which hot water is to be supplied; and a situation detected by the situation detecting means as input data.
Inference means comprising a neural network for inferring and outputting information about a temperature that a user of the water heater would want, and hot water supply control means for supplying hot water based on the content inferred by the inference means. Water heater characterized by doing. 2. A hot water temperature designating means for a user to adjust the temperature of hot water provided from a water heater, an automatic designating means for automatically inferring an optimum temperature by inference, and a hot water supply control for controlling hot water supply. Means, when the temperature of hot water is designated by the hot water temperature designation means, the hot water supply control means performs hot water supply control to provide hot water at the designated temperature, and at the time of automatic designation by the automatic designation means, The hot water supply control means is a situation detection means for detecting a situation in which hot water is to be supplied,
And information inferred based on inference means consisting of a neural network for inferring and outputting information relating to the temperature that the user of the water heater would want as the input data using the situation detected by the situation detection means as input data. A water heater characterized by controlling hot water supply. 3. A hot water temperature designating means for a user to adjust the temperature of hot water provided from a water heater, an automatic designating means for automatically inferring an optimum temperature by inference, and a hot water supply control for controlling hot water supply. Means, when the temperature of hot water is designated by the hot water temperature designation means, the hot water supply control means performs hot water supply control to provide hot water at the designated temperature, and at the time of automatic designation by the automatic designation means, The hot water supply control means is a situation detection means for detecting a situation in which hot water is to be supplied,
And information inferred on the basis of the situation detected by the situation detecting means as input data and inferring means comprising a neural network for inferring and outputting information relating to a temperature that a user of the water heater would desire. Learning means for controlling the supply of hot water and for learning the inference means using the temperature designated by the hot water temperature designation means and the situation detected by the situation detection means at that time as a teaching signal; And a water heater comprising: 4. The water heater is one of a water heater that supplies hot water to a kitchen, a water heater that supplies hot water to a bathroom, and a water heater that supplies hot water to a washroom. The water heater according to any one of claims 1 to 3. 5. A water heater that has at least two hot water supply points, a first hot water supply point for supplying hot water and a second hot water supply point for supplying hot water, and detects a situation in which hot water is to be supplied. Means, further comprising one or more neuron units between the neuron unit of the input layer and the neuron unit of the output layer, wherein a synapse having a weight coefficient connects the neuron units, and the situation detection means is provided in the input layer. Inference means comprising a neural network for inputting the detection data of the above and outputting information on a hot water supply temperature that a user of the water heater would want from an output layer; Data storage means for storing the hot water supply possibility; hot water supply possibility determining means for determining the possibility of hot water supply; The first weighting factor setting unit that sets the weighting factor stored in the data storage unit at the most likely location among the discrimination units in the inference unit, and in the hot water supply, the hot water supply possibility determining unit A second weighting factor for determining whether or not the hot water supply location and the actual hot water supply location determined in accordance with each other match or disagree, and in the case of a mismatch, resetting the weighting factor of the inference means to a weighting factor corresponding to the actual hot water supply location. A water heater comprising: setting means; and hot water supply control means for supplying hot water at a temperature according to the information inferred by the inference means. 6. A water heater having at least two hot water supply points, a first hot water supply point for supplying hot water and a second hot water supply point for supplying hot water, wherein a situation detection for detecting a situation in which hot water is to be supplied is provided. Means, one or more neuron units between the input layer neuron unit and the output layer neuron unit, and a synapse having a weighting coefficient connects the neuron units to each other, and the input layer has Inference means comprising a neural network that inputs detection data of the detection means and outputs information about a hot water supply temperature that a user of a water heater would want from an output layer, and a weighting coefficient of a synapse of the inference means, Data storage means for storing corresponding to the hot water supply location; weight coefficient setting means for setting the weight coefficient stored in the data storage means to the inference means; The coefficient setting means sequentially sets the weighting coefficient of each hot water supply point stored in the data storage means in the inference means, so that the water heater temperature corresponding to each hot water supply point that the user of the water heater corresponding to each hot water supply point may desire. And a hot water supply control means for inferring information in advance and supplying hot water based on the inferred contents corresponding to the hot water supply location when hot water is supplied.