JP3497240B2 - Air quality detection device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sensor for detecting dirt in indoor air of a building air conditioner, a house, and the like, and a detector equipped with the sensor.

[0002]

2. Description of the Related Art In recent years, energy saving has progressed in houses and buildings.
However, airtightness is improved to improve the efficiency of air conditioning.
On the other hand, such airtightness is caused by smoking, for example.
Dust, H ₂Formed from gases such as CO, CO, and building materials
Aldehyde, volatile organic gas (VOC gas), stove
CO, NO generated from combustion appliances such as_x, SO_x, CO₂etc
Gas, CO generated by human breathing₂, Also cigarettes
Smells such as odors, body odors, rotten odors, and bad odors may accumulate in the room.
May pollute the indoor air and cause adverse effects and discomfort to people in the room.
You In buildings, CO
₂Despite being ventilated based on concentration
Sick Building Sindh caused by the above VOC gas etc.
ROHM is a problem. Therefore, such a room
While keeping energy conserving and comfort in the contaminated air inside, ventilation and emptying
Required to be discharged and removed by air purifier
It

There is also a demand for measurement using human odor sensation as an index of indoor air quality pollution.

For this reason, it is known that indoor air pollution is detected by using CO ₂ concentration and gas concentration, and ventilation and air cleaning are performed.

Conventionally, this type of air quality detecting device has generally been constructed as shown in Japanese Patent Application Laid-Open No. 1-181026. The configuration will be described below with reference to FIG.

As shown in FIG. 29, the carbon monoxide sensor 1
01A, hydrogen sensor 101B, propane sensor 101
C, the oxygen sensor 101D, and the smoke sensor 101E output analog signals to the comparators 102A to 102E, respectively. The reference levels of the comparators 102A to 102E are set in advance. When the indoor air pollution increases and the sensor output exceeds the reference level, the output becomes, for example, "H".

Further, the outputs of the comparators 102A to 102E are
It is input to the arithmetic means 103. The operation means 103 performs an OR operation and outputs one of the outputs of the sensors 101A to 101E.
If the reference level is exceeded at any time, a signal is output to the output means 104 to operate the ventilation fan.

[0008]

In such a conventional air quality detecting device, when setting the reference level and obtaining the sensor output, the reference value is obtained from the sensor signal and the difference between the reference value and the sensor output is used. It is common to use the sensor signal in the clean outside air as the reference value.

However, in the room to be measured,
Compared to the air level (clean outside air), there is gas generated from building materials, furniture, etc. as a background. For example, hydrocarbons such as methane, alcohols, VOC gas, aldehydes, etc. High concentrations of these gases impair health, but they exist at low concentration levels that pose no health problems, and their concentrations and components differ from room to room. Further, the person occupying the room perceives the state of indoor air quality and the amount of dirt as, for example, a smell based on the state where the background gas exists. Therefore, when trying to measure the odor sensation of a person in the room with a sensor, it is necessary to measure the background and to measure the amount of contamination with reference to the background. Therefore, it is necessary to accurately obtain the background.

The present invention solves the above-mentioned problems, and detects the indoor pollution state with high accuracy by accurately obtaining the background level from the slope which is the stability of the sensor output and its duration using a fuzzy method. The first purpose is to be able to.

A second object is to accurately detect a background level peculiar to each room, building, or area by learning the background level with a neural network.

A third object is to accurately detect a polluted state of indoor air by accurately obtaining an air level (clean outside air) by using a fuzzy method from a sensor signal, a duration of inclination, and a background level. It is in.

A fourth object is to learn the air level in the room by a neural network,
It is to accurately detect the air level peculiar to each building or area.

A fifth object is to accurately determine the state of contamination of the indoor air or the cause of the contamination by using a fuzzy method from the amount of contamination of the indoor air obtained from the background level and the amount of contamination of the indoor air obtained from the air level. It is to judge well.

A sixth object is to determine the state of indoor air pollution by using a fuzzy method from the indoor air pollution amount obtained from the background level, the indoor air pollution amount obtained from the air level, and the inclination of the sensor signal. The purpose is to accurately and quickly determine the cause of contamination.

A seventh purpose is to learn the determined state discriminant value by a neural network, and
The purpose is to accurately determine the state of indoor air pollution that is unique to each building or area.

The eighth object is to accurately determine the pollution level (contamination intensity) of each pollution state by using the fuzzy method from the indoor air pollution amount obtained from the background level and the indoor air pollution amount obtained from the air level. It is to calculate well.

The ninth object is peculiar to each room, each building, or each area by learning the calculated pollution level from the pollution intensity (or odor intensity) entered and declared by the occupants by a neural network. The purpose is to calculate the pollution intensity of indoor air as the pollution intensity corresponding to the sensation of the person in the room.

[0019]

A first means for achieving the first object of the present invention is to provide one or more sensors 1 for detecting dirt in the room air and outputting a dirt signal, and A sensor signal input unit for inputting a signal from one or a plurality of sensors 1, a signal sampling unit for sampling the signal of the sensor input by the sensor signal input unit at regular intervals, and a signal sampling unit for sampling A sampling signal output unit that outputs a signal, a sampling signal input unit A that inputs the sampling signal output by the sampling signal output unit, and a sampling signal of the sensor input by the sampling signal input unit A that is constant from the past to the present Outputs the sensor signal storage unit that stores the time and the sensor signal for the fixed time stored in the sensor signal storage unit A stored signal output unit, a stored signal input unit for inputting a stored signal output from the stored signal output unit, and a sensor signal output variation per unit time from the stored signal input by the stored signal input unit (hereinafter, Called tilt)
A tilt calculation unit to be obtained, a tilt output unit that outputs the tilt calculated by the tilt calculation unit, a tilt input unit A that inputs the tilt signal output from the tilt output unit, and the tilt input unit A
A count unit that counts the slope duration input by each slope level, and a slope duration output unit that outputs the slope duration counted by the count unit,
A background level calculation fuzzy table storage section in which a membership function for calculating the background level is prepared in advance as a background level calculation fuzzy table, and membership stored in the background level calculation fuzzy table storage section A membership function output unit that outputs a function, a sampling signal input unit B that inputs the sampling signal of the sensor that is output from the sampling signal output unit, and a slope duration time that inputs the signal output from the slope duration time output unit An input unit A and a membership function input unit for inputting a membership function output from the membership function output unit are provided, and the slope duration input by the slope duration input unit A and the membership function input unit are Membership function entered And a background level calculation unit that obtains an output value (hereinafter, background level) of the sensor in a normal state of indoor air (hereinafter, referred to as background), and a background level signal calculated by the background level calculation unit. A background level output section for outputting the sampling signal, a sampling signal input section C for inputting a sampling signal from the sampling signal output section, and a background level input section for inputting a signal of the background level output section. The background level value input by the ground level input section is stored as a reference value 1 and the sampling signal input section C is stored.
It is provided with a dirt amount 1 calculation unit for calculating the dirt amount 1 of the indoor air by comparing the sampling signal of the sensor input in step 1, and a dirt amount 1 output unit for outputting the calculated dirt amount to the outside.

Second aspect for achieving the second object of the present invention
Means comprises a background level input section for inputting the background level signal output from the background level output section as the reference value 1 and a sampling signal input section D for inputting the sampling signal of the sensor from the sampling signal output section. A reference for taking in the reference value 1 input at the background level input section and the sampling signal of each sensor input at the sampling signal input section D, and learning the correlation between the reference value 1 and the sampling signal of the sensor by a neural network. The reference value 1 learning unit and the reference value 1 output unit that outputs the reference value obtained by the learning of the reference value 1 learning unit to the reference value 1 input unit A of the stain amount 1 calculation unit are provided.

A third aspect for achieving the third object of the present invention
The means is an air level calculation fuzzy table storage unit that prepares a membership function for calculating the air level in advance as an air level calculation fuzzy table, and membership stored in the air level calculation fuzzy table storage unit. A membership function output unit that outputs a function, a slope duration input unit B that inputs the signal output from the slope duration time output unit, and a reference value 1 input that inputs the reference value 1 output from the reference value 1 output unit A section B and a membership function input section for inputting the membership function output from the membership function output section, and the slope duration input in the slope duration input section B and the reference value 1 input section B The room air is usually calculated from the reference value entered in step 1 and the membership function entered in the membership function input section. An air level calculation unit that calculates a signal of the sensor (hereinafter referred to as an air level) when it becomes equivalent to the air of the outside air, an air level output unit that outputs the air level calculated by the air level calculation unit, and sampling The air level input by the air level input section is provided with a sampling signal input section E for inputting the sampling signal of the sensor from the signal output section and an air level input section for inputting the air level output by the air level output section. The signal is stored as a reference value 2, and the stored reference value 2 is compared with the sampling signal of the sensor input by the sampling signal input unit E to calculate a contamination amount 2 of indoor air. The dirt amount 2 output section for outputting the quantity 2 and the dirt quantity 1 from the dirt quantity 1 output section and the dirt quantity 2 from the dirt quantity 2 output section are excluded. It is obtained and an output unit for outputting to.

A fourth object for achieving the fourth object of the present invention.
Means includes an air level input section for inputting an air level signal output from the air level output section and a sampling signal input section F for inputting a sampling signal of the sensor from the sampling signal output section. A reference value 2 learning unit that takes in the air level input in step S and the sampling signal of the sensor input in the sampling signal input unit F and learns the correlation between the reference value 2 and the sampling signal of the sensor by a neural network; The reference value 2 output section that outputs the reference value 2 obtained by the learning of the 2 learning section to the dirt amount 2 calculation section is provided.

A fifth aspect for achieving the fifth object of the present invention.
Means of the background gas, the state of pollution of indoor air,
When determining whether a person is in a smoking state or a person is in the room, an alcohol sensor is a sensor that detects background gas and outputs a detection signal, and a sensor that detects a smoking state and outputs a detection signal is H.
₂ sensors or H _2S sensor or both sensors, a sensor that detects the presence of a person and outputs a detection signal, a H _2S sensor or a CO ₂ sensor or both sensors, and a membership function for determining the state of indoor air From a state determination fuzzy table storage section prepared in advance as a state determination fuzzy table, a membership function output section for outputting the membership function stored in the state determination fuzzy table storage section, and a contamination amount 1 output section. Dirt amount 1 to be input Dirt amount 1 Input part A and Dirt amount 2 Dirt amount 2 from output part
A dirt amount 2 input section A for inputting and a membership function input section for inputting a membership function from the membership function output section, and the dirt amount 1 and dirt amount 2 inputs inputted at the dirt amount 1 input section A A state determination unit that determines the cause (state) of contamination of the indoor air from the dirt amount 2 input in the section A and the membership function input in the membership function input unit, and the state determination of the indoor air determined by the state determination unit A state determination signal output section for outputting a signal, an output signal for the contamination level 1 from the contamination level 1 output section, an output signal for the contamination level 2 from the contamination level 2 output section, and an output signal for the state determination signal output section. An output section for outputting the state determination signal to the outside is provided.

A sixth aspect for achieving the sixth object of the present invention.
Means for detecting a background gas and outputting a detection signal, an H ₂ sensor and an H _2S sensor for detecting a smoking state and outputting a detection signal, and a human presence in the room and outputting a detection signal. H _{2 S} sensor or CO ₂ sensor and tilt input section B for inputting a tilt signal from the tilt output section
And dirt amount 1 Input dirt amount 1 from the output portion dirt amount 1 input portion A and dirt amount 2 Input dirt amount 2 from the stain portion 2 input portion A and the member from the membership function output portion A membership function input unit for inputting a ship function is provided, and the tilt signal input at the tilt input unit B and the contamination amount 1
A state determination unit that determines the cause (state) of indoor air contamination from the contamination amount 1 and the contamination amount 2 input in the input unit A and the contamination amount 2 input in the input unit A and the membership function input in the membership function input unit It is equipped with and.

A seventh aspect for achieving the seventh object of the present invention.
Means for inputting a state determination signal from the state determination signal output section and a contamination amount 1 input section B for inputting a contamination amount 1 from the contamination amount 1 output section and a contamination amount 2 output section. A dirt amount 2 input section B for inputting a dirt amount 2 is provided, and the state determination signal input by the status determination signal input section and the dirt amount 1 input section B and the dirt amount 1 and dirt amount input at the dirt amount 2 input section B A state discrimination learning unit that learns the correlation with 2 by a neural network, a state discrimination learning output unit that outputs the discrimination state obtained by the learning of the state discrimination learning unit, and a discrimination state signal from the state discrimination learning output unit. And an output unit for outputting the stain amount 1 from the stain amount 1 output unit and the stain amount 2 from the stain amount 2 output unit to the outside.

Eighth Object to Achieve the Eighth Object of the Present Invention
The means is stored in the pollution level calculation fuzzy table storage unit and a pollution level calculation fuzzy table storage unit that prepares a membership function for calculating the pollution level of indoor air as a pollution level calculation fuzzy table in advance. A membership function output section that outputs a membership function and a dirt amount 1 input dirt quantity 1 from the dirt quantity 1 output section and a dirt quantity 2 dirt quantity 2 that inputs a dirt quantity 2 from the output quantity The input section C and the membership function input section for inputting the membership function from the membership function output section are provided, and the stain amount 1 and the stain amount 2 input in the stain amount 1 input unit C A pollution level calculation unit for calculating a pollution level of indoor air from the quantity 2 and the membership function input in the membership function input unit; A contamination level output unit that outputs the contamination level calculated by the bell calculation unit, and a contamination amount 1 signal and a contamination amount 2 signal from the contamination amount 1 output unit and the contamination amount 2 output unit, and discrimination from the state discrimination learning output unit An output unit for outputting the status signal and the pollution level signal from the pollution level output unit to the outside is provided.

A ninth aspect for achieving the ninth object of the present invention.
The measures are the evaluation standard table, which is a standard for the indoor occupants to evaluate the pollution intensity of the indoor air (or a standard indicating the degree of contamination such as odor intensity), and the indoor air pollution felt by the indoor occupants. Input a declared value input unit for inputting the strength based on the evaluation standard table, a declared value output unit for outputting the declared value input in the declared value input unit, and a declared value signal from the declared value output unit. A declared value input section and a contamination level input section for inputting a contamination level from the contamination level output section; and a declared value signal inputted from the declared value input section and a contamination level signal inputted by the contamination level input section. A pollution level learning unit that learns the correlation of the above by a neural network, a pollution level learning output unit that outputs the pollution level obtained by the learning of the pollution level learning unit, a contamination amount 1 output unit, and a contamination amount 2 output. And an output unit for outputting the contamination amount 1 signal and the contamination amount 2 signal from the unit, the discrimination state signal from the state discrimination learning output unit, and the contamination level signal from the contamination level learning output unit to the outside. is there.

[0028]

According to the present invention, by the configuration of the above-mentioned first means, the reference value 1 is calculated by using the fuzzy method, so that the background level, which is a reference value for obtaining the amount of indoor air pollution, is calculated. Since it can be obtained with high accuracy, it is possible to detect the contamination state in the room with high accuracy.

Further, with the configuration of the second means, it is possible to accurately detect the background level peculiar to each room, each building, or each area by learning the background level by the neural network.

Further, with the configuration of the third means, by calculating the reference value 2 using the fuzzy method, the air level (clean outside air), which is a reference value for obtaining the amount of pollution of the indoor air, is accurately measured. Since it can be obtained well, it is possible to detect the contamination state of the indoor air with high accuracy.

Further, with the configuration of the fourth means, it is possible to accurately detect the air level peculiar to each room, each building, or each area by learning the air level in the room by the neural network.

Further, according to the configuration of the fifth means, it is possible to accurately determine the state of contamination of the indoor air or the cause of the contamination by determining the state of contamination of the indoor air or the cause of contamination by using the fuzzy method. Is.

In addition, with the configuration of the sixth means, it is possible to accurately and quickly determine the state of contamination of the indoor air or the cause of contamination by performing the state determination using the fuzzy method and the state detection from the inclination. It is possible.

Further, with the configuration of the seventh means, it is possible to accurately determine the state of contamination of indoor air peculiar to each room, each building, or each area by learning the determined state determination value by a neural network. Is possible.

Further, with the configuration of the eighth means, the pollution level (contamination intensity) of each pollution state is calculated using the fuzzy method, whereby the pollution level of each pollution state can be calculated accurately.

Further, by the construction of the ninth means, the calculated pollution level is learned by a neural network from the pollution intensity (or (odor intensity)) which the occupant declares and inputs, and the room, the building, or the area. It is possible to calculate the pollution intensity of the indoor air, which is unique to each, as the pollution intensity corresponding to the sensation of the person in the room.

[0037]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to FIGS. The same parts as those of the conventional example are designated by the same reference numerals and the description thereof will be omitted.

As shown in FIG. 1, the sensor 1 has a plurality of sensors a to n each of which has a different characteristic for detecting dirt in the room air, and each of the sensors a reacts in accordance with the gas concentration of the room air. From a, the sensor signal output unit n outputs the gas concentration reacted by each sensor.

The signal sampling section 2 outputs the signals from the sensor a to the sensor n output from the sensor signal output section a of the sensor 1 to the sensor signal input section 2 of the sensor 1.
a, the input signal is sampled at fixed time intervals (for example, every 10 seconds, every 1 minute), and the sampling signal output unit 2b samples the sampling signal input unit 3a, the sampling signal input unit 6a, and the sampling signal input unit 8a.
Send to each.

The sensor signal storage section 3 receives the signals of the respective sensors input by the sampling signal input section 3a,
You can store each input sensor signal for the past N signals,
When the latest sensor signal is sent, the oldest sensor signal among the stored N sensor signals is discarded, and the latest sensor signal is stored. The memory signal output unit 3b outputs N signals to the memory signal input unit 4a.

The inclination calculation section 4 inputs the signal sent from the storage signal output section 3b by the storage signal input section 4a,
The inclination of each sensor is calculated from the input signal by the formula shown in (Equation 1), and the inclination of each sensor calculated by the inclination output unit 4b is output to the inclination input unit 5a.

[0042]

[Equation 1] Here, DV is the slope, _VN is the latest sensor output, and _VN-1 is the old sensor signal.

The counting section 5 receives the tilt signal sent from the tilt output section 4b by the tilt input section 5a, ranks it according to the magnitude of the tilt, counts the duration of the tilt for each rank, and counts it. The slope duration time output unit 5b outputs the slope duration time for each rank to the slope duration time input unit 6b.

The background level calculating section 6 includes the signals of the respective sensors from the sampling signal output section 2b input by the sampling signal input section 6a and the respective sensors from the inclination duration input section 5b input by the inclination duration input section 6b. Duration of each slope of rank and the membership function output unit 7 of the background level calculation fuzzy table storage unit 7 input by the membership function input unit 6c
From a, a membership function for calculating the background level of each sensor prepared in advance in the background level calculation fuzzy table storage unit 7 and the background level of each sensor are calculated, and the background level output unit The background level calculated by 6d is output to the background level input unit 8b. The membership function shown in FIG. 2 is prepared in advance in the background level calculation fuzzy table storage unit 7, and the membership functions for calculating the background level are as shown in the table.

The dirt amount 1 calculation unit 8 takes in the background level signal of each sensor input by the background level input unit 8d, and updates and stores it as a reference value. Further, the sensor signal from the sampling signal output unit 2b is taken in by the sampling signal input unit 8a, the stain amount 1 is calculated from the stored reference value and the sensor signal, and the calculated stain amount 1 is externally output by the stain amount 1 output unit 8c. Output.

Next, the operation will be described. In FIG. 3, after turning on the power of the air quality detection device and warming up the sensor, if the air quality detection device is placed in a normal measurement atmosphere, the operations from step 9 to step 16 are performed until the power is turned off. Repeated.

The operation from step 9 to step 16 will be described. First, in step 9, the signal sampling section 2 determines whether or not the elapsed time of the sampling timing has elapsed. If not, the process proceeds to step 9. If the return has elapsed, the process proceeds to step 10, and after sampling the sensor signal of each sensor, step 11 is performed.
Proceed to.

In step 11, the signal sampling unit 2
The signal of each sensor sampled in step S12 is taken into the sensor signal storage unit 3, replaced with the oldest sensor signal among the stored N sensor signals, and stored, and step 12
Proceed to.

In step 12, the inclination (DV) is calculated by the inclination calculating section 4 based on the N sensor signals of each sensor stored in the sensor signal storage section 3, and the step 13 is performed.
Proceed to.

In step 13, the inclination (DV) of each sensor calculated by the inclination calculator 4 by the counting unit 5 is fetched and ranked according to the magnitude of the inclination, and the duration of the inclination continuously fetched for each rank is determined. The duration of the slope that has been counted and counted for each rank is sent to the background level calculation unit 6 and the process proceeds to step 14.

In step 14, the background level calculation unit 6 fetches the duration time for each rank of each sensor counted by the counting unit 5, and the background level calculation fuzzy table 7 prepares a background function for each sensor by a membership function. The ground level is calculated, and the calculated background level is used as the dirt amount 1 calculation unit 8
To step 15.

In step 15, the background level signal of each sensor received from the background level calculating section 6 by the contamination amount 1 calculating section 8 is updated and stored as a reference value, and the stored reference value and the signal sampling section 2 are fetched. The dirt amount 1 is calculated from the sensor signal.

In step 16, the dirt amount 1 signal calculated by the dirt amount 1 calculator 8 is output to the outside by the dirt amount 1 output means 8c.

After that, the process returns to step 9. In addition,
As the sensor 1, various types of sensors can be used as long as they are sensitive to pollutants that represent the polluted state of the indoor air to be detected. For example, when detecting smoking, a hydrogen sensor or a CO sensor as a cigarette sensor, an alcohol sensor when detecting a VOC gas, a carbon dioxide gas sensor when detecting a person or combustion, and a malodor sensor when detecting a bad smell. , Hydrogen sulfide sensor, ammonia sensor, etc. The number of the sensors 1 may be one or more depending on the detection target.

Further, in this embodiment, the sensor signal storage unit 3
The number of N sensor signals that can be stored is set to 2, and the difference is calculated from the two sensor signals to obtain the slope of the sensor signal trend. However, if there are two or more sensor signals that can be stored, for example, N
= 10, and the slope of the sensor signal trend can be understood by calculating the slope from the difference between the average of the first 5 (or the minimum or maximum value) and the average of the latter 5 (or the minimum or minimum). Any method may be used to calculate the number of Ns and the slope.

With the above configuration, it is possible to accurately obtain the sensor signal when there is no load fluctuation and the sensor signal is stable, using the fuzzy method from the inclination of the sensor signal trend and its duration as a reference value. is there.

As described above, according to the air quality detecting apparatus of the first embodiment of the present invention, the background level in the room can be detected accurately, so that the amount of dirt in the room air can be detected with high accuracy. Energy-saving ventilation can be performed by automatically controlling ventilation according to the amount of dirt.

Next, the second embodiment of the present invention will be described with reference to FIG.
From now on, description will be made with reference to FIG. The same parts as those of the conventional example and the first embodiment of the present invention are designated by the same reference numerals and the description thereof will be omitted.

As shown in FIG. 4, the reference value 1 learning unit 17
Inputs the sensor signals from the sampling signal output section 2b through the sampling signal input section 17a. Further, the background level value of each sensor output by the background level output unit 6d is input as the reference value 1 by the background level input unit 17b, and the reference value 1 is learned by correlating with the input sensor signals. ,
The learned reference value 1 is output to the reference value 1 input unit by the reference value 1 output unit 17c. FIG. 5 shows the neural network unit of the reference value 1 learning unit 17.

In FIG. 5, S1 to Sn are a plurality of sensors, I1 to In are input portions of sensor signals, and M
Reference numerals 1 to Mn are intermediate portions for weighting the input signals of the sensors, and O1 to On are output portions for outputting the reference values of the sensors. Then, the correlation between each sensor signal and each reference value is learned from each reference value output from O1 to On and each weighted background level value.

The dirt amount 1 calculation unit 8 includes a reference value 1 input unit 8d.
The reference value 1 signal of each sensor input in 1 is taken in, updated and stored, and the contamination amount 1 is calculated from the sampling signal of each sensor input by the sampling signal input unit 8b and the updated and stored reference value of each sensor.

Next, the operation will be described. As shown in FIG. 6, in step 18, the background level value of each sensor calculated by the background level calculation unit 6 and each sensor signal sampled by the signal sampling unit 2 are taken in and the reference value of each sensor is acquired by the neural network. 1 is learned, the reference value 1 of each sensor obtained by learning is output to the stain amount 1 calculator 8, and the process proceeds to step 15.

As described above, according to the air quality detecting apparatus of the second embodiment of the present invention, the reference value can be accurately obtained by learning the reference value by the neural network. It can detect with high accuracy. Further, since the reference value is learned, the background level peculiar to each room, each building, and each region can be accurately obtained, and more optimal energy-saving automatic ventilation control can be performed.

Next, the third embodiment of the present invention will be described with reference to FIG.
9 to 9 will be described. The same parts as those of the conventional example and the embodiment of the present invention are designated by the same reference numerals and the description thereof will be omitted.

As shown in FIG. 7, the air level calculator 1
0 is the inclination duration time of each sensor rank input by the inclination duration input unit 19a and the reference value 1 of each sensor input by the reference value 1 input unit 19b and the membership function of the air level calculation fuzzy table storage unit 20. Output unit 20
From a, the air level calculation fuzzy table storage unit 2
The membership level prepared in advance is used to calculate the air level value of each sensor, and the calculated air level value is output to the air level input section 21a by the air level output section 19d. The membership function shown in FIG. 8 is prepared in advance in the air level calculation fuzzy table 20, and the membership function for calculating the air level is as shown in the table.

The dirt amount 2 calculation unit 21 takes in the air level value of each sensor input by the air level input unit 21a and updates and stores it as the reference value 2. Further, from the stored reference value 2 and the sampling signal of each sensor input by the sampling signal input unit 21b, the contamination amount 2 of the indoor air
Is calculated, and the calculated stain amount 2 signal is used as the stain amount 2 output unit 21.
Output by c.

The output unit 104 outputs the dirty amount 1 of the indoor air from the dirty amount 1 output unit 8c and the dirty amount 2 of the indoor air from the dirty amount 2 output unit 21c to the outside.

Next, the operation will be described. As shown in FIG. 9, in step 22, the reference value 1 of each sensor calculated by the reference value 1 learning unit 9 and the signal of each sensor sampled by the signal sampling unit 2 are fetched and the magnitude comparison is performed, and “sensor signal <reference If "value 1", the process for the sensor signal ends and the process proceeds to step 15. If "sensor signal> reference value 1", the process proceeds to the following process. The reference value 1
The learning unit 9 further calculates the reference value 2 of each sensor by taking in the duration of the inclination for each rank of each sensor, each sensor signal, and the membership function from the air level calculation fuzzy table 11 from the counting unit 5. 15
Move on to. Here, the reference value 2 is an output value of each sensor output by each sensor when the indoor air is equal to the outside air.

In step 15, each sensor signal sampled by the signal sampling unit 2 by the dirt amount 1 calculator 8 and the reference value 1 of each sensor calculated by the reference value 1 learning unit 17
And are taken in to calculate the stain amount 1 in the room. On the other hand, the dirt amount 2 calculation unit 21 takes in each sensor signal sampled by the signal sampling unit 2 and the reference value 2 of each sensor calculated by the air level calculation unit 19, calculates the dirt amount 2 in the room, and proceeds to step 16. .

In step 16, the output unit 104 uses the ambient air level calculated by the stain amount 1 and stain amount 2 calculation units 21 as the reference, based on the indoor background level calculated by the stain amount 1 calculation unit 8. The indoor dirt amount 2 and the dirt amount are taken in and output to the outside.

As described above, according to the air quality detecting apparatus of the third embodiment of the present invention, the sensor signal when the indoor air is equal to the outside air can be accurately obtained by using the fuzzy method. It is possible to accurately detect the amount of dirt from when the air is cleaner. Furthermore, since the background level and the outside air level in the room can be detected at the same time, it is possible to know the amount of change in the load felt by the person in the room, and the amount of dirt in the room air felt by the person in the room and the actual (outside air level) By performing automatic ventilation control or the like according to the amount of dirt in the room air (from (1)), it is possible to perform comfortable automatic ventilation control for the occupants in energy saving.

Next, the fourth embodiment of the present invention will be described with reference to FIG.
This will be described with reference to FIGS. In addition, the conventional example,
The same parts as those of the embodiment of the present invention are designated by the same reference numerals and the description thereof will be omitted.

As shown in FIG. 10, the reference value 2 learning unit 23
Captures the sampling signal of each sensor input by the sampling signal input unit 23a. Further, the air level value of each sensor input to the air level input unit 23b is taken as a reference value 2 and associated with each sensor signal to learn the reference value 2, and the learned reference value 2 is output by the reference value 2 output unit 23c. Output.

FIG. 11 shows the neural network unit of the reference value 2 learning unit 23.

In FIG. 11, S1 to Sn denote a plurality of sensors, I1 to In denote input portions for sensor signals,
M1 to Mn are intermediate portions for weighting the input signals of the sensors, and O1 to On are output portions for outputting the reference value 2 of the sensors. Then, the correlation between each sensor signal and each reference value 2 is learned from each reference value 2 output from O1 to On and each weighted air level value.

The dirt amount 2 calculation unit 21 includes a reference value 2 input unit 2
The reference value 2 of each sensor input in 1d is taken in, updated and stored, and the contamination amount 2 is calculated from the sampling signal of each sensor sampled by the sampling signal input portion 21b and the updated and stored reference value 2 of each sensor.

Next, the operation will be described. As shown in FIG. 12, in step 24, the air level calculation unit 19
The reference value 2 of each sensor obtained by learning is obtained by taking in the background level value of each sensor calculated in step 1 and each sensor signal sampled by the signal sampling unit 2 and learning the reference value 2 of each sensor by a neural network.
Is output to the dirt amount 2 calculator 21, and the process proceeds to step 15.

As described above, according to the air quality detecting apparatus of the fourth embodiment of the present invention, the reference value 2 which is the sensor signal when the indoor air is equal to the outside air is obtained by learning the reference value 2 by the neural network. Since 2 can be obtained with high accuracy, the amount of dirt in the room air can be detected with high accuracy. Further, since the reference value is learned, the background level peculiar to each room, each building, and each region can be accurately obtained, and more optimal energy-saving automatic ventilation control can be performed.

Next, the fifth embodiment of the present invention will be described with reference to FIG.
It will be described with reference to FIGS. In addition, the conventional example,
The same parts as those of the embodiment of the present invention are designated by the same reference numerals and the description thereof will be omitted.

As shown in FIG. 13, the sensor 1 has a V
Alcohol sensor as a sensor to detect the OC odor outputs a detection signal, H _2S and as a sensor for outputting a detected detection signal smoking odor H ₂ sensor and the H _2S sensor, also as a sensor for outputting a detected body odors detection signal A CO ₂ sensor is prepared, each sensor detects each gas concentration of the air in the room, and each sensor signal output section outputs each gas concentration signal detected by each sensor.

The state discriminating section 25 has a dirt amount 1 input section 25a.
The state determination from the membership function output unit 26a of the fuzzy table storage unit 26 of the soil amount 1 signal and the soil amount 2 input unit 25b and the soil amount 2 signal input from the membership function input unit 25c The state determination membership function prepared in advance in the fuzzy table storage unit 26 is read to determine the cause (or state) of indoor air pollution, and the determined cause (or state) of indoor air pollution is output to the state determination signal output unit 25d. Is output by the output unit 104. The membership function shown in FIG. 14 is prepared in advance in the state determination fuzzy table storage unit 26, and the membership function for determining the state of indoor air is as shown in the table. Smoking odor is the amount of H ₂ sensor contamination 1
The signal is used to determine the smoking state when a dirt amount of 1 is generated, and the VOC odor is _{detected when} the dirt amount 2 of the H ₂ sensor, OH sensor or H _2S sensor is used, and when the dirt amount of 2 is generated V
The OC odor is discriminated, and the body odor is discriminated by using the dirt amount 2 signal of the H _{2 S} sensor or the CO ₂ concentration signal of the CO ₂ sensor when the dirt amount 2 or the CO ₂ concentration occurs.

Next, the operation will be described. In FIG. 15, in step 27, the state determination unit 25 determines that the amount of stain 1
Contamination amount 1 signal and contamination amount 2 calculation unit 21 calculated by the calculation unit 8
The contamination amount 2 signal calculated in step 2 and the membership function for status determination prepared in the status determination fuzzy table storage unit 26 are taken in to determine the cause of contamination (or contamination state) of the indoor air, and the determined state signal Output section 10
4 and go to step 16.

In step 16, the stain amount 1 and stain amount 2 calculator 21 calculated by the stain amount 1 calculator 8 by the output unit 104.
The dirt amount 2 calculated in step 2 and the state determination signal calculated in the state determination part 25 are fetched and output to the outside.

As described above, according to the air quality detecting apparatus of the fifth embodiment of the present invention, the state of the room air is changed to smoking odor, VOC odor,
The cause of the indoor air pollution (or the state of pollution) because it is possible to accurately determine the state using the body odor as the cause.
It is possible to perform optimal ventilation control suitable for.

Next, a sixth embodiment of the present invention will be described with reference to FIG.
This will be described with reference to FIGS. In addition, the conventional example,
The same parts as those of the embodiment of the present invention are designated by the same reference numerals and the description thereof will be omitted.

As shown in FIG. 16, the state discriminating unit 25 inputs the inclination signal of each sensor input by the inclination input unit 25e, the contamination amount 1 signal input by the contamination amount 1 input unit 25a, and the contamination amount 2 input unit 25b. The dirtiness 2 signal and the membership function output unit 26a of the status determination fuzzy table storage unit 26 input by the membership function input unit 25c. The status determination membership function prepared in advance in the status determination fuzzy table storage unit 26. Is taken in to determine the cause (or state) of contamination of the indoor air, and the determined state determination signal is output to the output unit 104 by the state determination signal output unit 25d.

The membership function shown in FIG. 17 is prepared in advance in the state determination fuzzy table storage unit 26, and the membership function for determining the state of indoor air is as shown in the table. For the smoking odor, the smoke amount 1 signal of the H ₂ sensor is used to determine the smoking state when the stain amount 1 occurs, and the VO
For the C odor, the VOC odor is discriminated when the contamination amount 2 signal of the H ₂ sensor, the OH sensor or the H _2S sensor is used, and the body odor is determined by the contamination amount 2 signal of the H _2S sensor or CO.
₂ sensor CO ₂ concentration signal is used
₂ Determine the body odor condition when a concentration occurs. Furthermore, the inclination value of each sensor is monitored, and when the sensor signal changes rapidly, that is, when the indoor air changes rapidly due to smoking, the inclination of the sensor signal becomes a large value. Determine.

Next, the operation will be described. 18, in step 27, the state determination unit 25 causes the sensor inclination signal from the inclination calculation unit 4 and the contamination amount 1 signal and contamination amount 2 calculation unit 2 of each sensor calculated by the contamination amount 1 calculation unit 8 to be calculated.
The contamination amount 2 signal of each sensor calculated in 1 and the state determination membership function prepared in advance in the state determination fuzzy table storage unit 26 are taken in to determine the contamination cause (or contamination state) of the indoor air, and the determined state The signal is sent to the output unit 104 and the process proceeds to step 16.

As described above, according to the air quality detecting apparatus of the sixth embodiment of the present invention, by determining the start of the state from the trend of the sensor signal, that is, the inclination value of the sensor, the cause of the contamination of the indoor air (or the contamination state). ) Can be accurately discriminated and the state can be quickly discriminated.

Next, the seventh embodiment of the present invention will be described with reference to FIG.
It will be described with reference to FIGS. In addition, the conventional example,
The same parts as those of the embodiment of the present invention are designated by the same reference numerals and the description thereof will be omitted.

As shown in FIG. 19, the state discrimination learning unit 28
Captures the status signal input by the status determination signal input unit 28c. Further, the dirt amount 1 signal of each sensor input by the dirt amount 1 input unit 28a and the dirt amount 2 signal of each sensor input by the dirt amount 2 input unit 28b are taken in and are respectively associated with the status determination signals to determine the state. And the learned state determination signal is output to the output unit 104 by the state determination learning output unit 28d.

FIG. 20 shows the neural network section of the state discrimination learning section 28.

In FIG. 20, S11, S21, ...
Is a signal of the amount of contamination of each sensor, S12, S22, ...
Is a dirt amount 2 signal of each sensor, I11 to Imn are input portions of dirt amount, M11 to Mmn are intermediate portions for weighting dirt amount, O1 to O3 are “O1 = smoking odor”,
The output unit outputs the states of “O2 = VOC odor” and “O3 = body odor”. Then, the correlation between each state output from O1 to O3 and each weighted stain amount is learned.

Next, the operation will be described. 21, in step 29, the state determination learning unit 28 causes the state signal of the indoor air from the state determination unit 25 and the contamination amount 1 signal and the contamination amount 2 calculation unit 21 of each sensor calculated by the contamination amount 1 calculation unit 8. The calculated contamination amount 2 signal of each sensor is taken in to learn the state determination of the cause of contamination (or contamination state) of the indoor air by the neural network, and the learned state determination signal is sent to the output unit 104, and the process proceeds to step 16.

As described above, according to the air quality detecting apparatus of the seventh embodiment of the present invention, by learning the cause of pollution of indoor air (or the pollution state) by the neural network, each room, each building, and each area. It is possible to accurately determine the cause of indoor air pollution, which is peculiar to, and to perform more optimal energy-saving automatic ventilation control.

Next, the eighth embodiment of the present invention will be described with reference to FIG.
2 to 24, description will be given. In addition, the conventional example,
The same parts as those of the embodiment of the present invention are designated by the same reference numerals and the description thereof will be omitted.

As shown in FIG. 22, the contamination level calculation unit 3
0 is the dirt amount 1 signal of each sensor input by the dirt amount 1 input unit 30a and the dirt amount 2 input by the dirt amount 2 input unit 30b.
The contamination level calculation membership function prepared in advance in the contamination level calculation fuzzy table storage unit 31 from the membership function output unit 31a of the contamination level calculation fuzzy table storage unit 31 is fetched by the signal and the membership function input unit 30c. The pollution level of each cause (or pollution state) of the indoor air is calculated, and the calculated pollution level value is output to the output unit 104 by the pollution level output unit 30d.

Contamination level calculation fuzzy table storage unit 3
23, the membership function shown in FIG. 23 is prepared in advance, and the membership function for calculating the indoor air pollution level is as shown in the table.

Next, the operation will be described. 24, in step 32, the contamination level calculation unit 30 calculates the contamination amount 1 signal of each sensor calculated by the contamination amount 1 calculation unit 8 and the contamination amount 2 signal of each sensor calculated by the contamination amount 2 calculation unit 21 and the contamination level. A membership function for calculating the pollution level prepared in advance in the calculation fuzzy table storage unit 31 is taken in to calculate each pollution level of the pollution cause (or the pollution state) of the indoor air, and the calculated pollution level signal is output. Send to section 104 and proceed to step 16.

As described above, according to the air quality detecting apparatus of the eighth embodiment of the present invention, it is possible to accurately obtain the pollution level for each of the causes of contamination (or the state of contamination) of indoor air. Ventilation control etc. can be performed.

Next, the ninth embodiment of the present invention will be described with reference to FIG.
5 to 27, description will be made. In addition, the conventional example,
The same parts as those of the embodiment of the present invention are designated by the same reference numerals and the description thereof will be omitted.

As shown in FIG. 25, the pollution level learning unit 3
Reference numeral 3 captures a signal of a reported value of odor (or odor intensity, pollution intensity, etc.) of the indoor air felt by a person in the room through the report value input unit 33b. The signal input to the declared value input unit 33b is based on the evaluation standard table 34b shown in FIG. 28, which evaluates the odor that indoor occupants feel in the room air, and the evaluation result is input to the declared value input unit, which is an external input unit. The evaluation value can be input by the unit 34, and the input evaluation value is output to the declared value input unit 33b by the declared value output unit 34a. The input level of the declared value input unit 34 has five levels: “strong odor”, “odor”, “weak odor”, “slight odor”, and “no odor”. Further, the pollution level learning unit 33 takes in the pollution level signal calculated by the pollution level calculation unit 30, and associates the pollution level signal with the odor declaration value signal. FIG. 26 shows the neural network unit of the contamination level learning unit 33. In FIG. 26, X,
Y and Z are the level values of the state of the indoor air, and are "X = smoking odor level", "Y = VOC odor level", and "Z = body odor level". IX to IZ are input portions for state level values, MX to MZ are intermediate portions for weighting state level values, O1 to O5 are “O1 = strong smell”, “O2”
The output unit outputs the states of “= smell”, “O3 = slightly smell”, “O4 = slightly smell”, and “O5 = not smell”. Then, the correlation between the level signal of each indoor air state input from the pollution level calculation unit 30 and each odor intensity output from O1 to O5 is learned.

The input level of the declared value input section 34 of the present embodiment is set to 5 levels of "strongly smelling", "smelling", "weakly smelling", "slightly smelling", and "not smelling". Any combination of expressions and levels may be used as long as the levels are expressed.

Next, the operation will be described. In FIG. 27, step 35 is a state level value of each indoor air pollution calculated by the pollution level calculation section 30 and a declared value input section 34.
The odor intensity according to the level of each pollution state of the indoor air is learned by taking in the odor report value of the indoor air sensed by the indoor occupant input by the above, and the learned odor intensity signal is output to the pollution level learning unit output means 33c. Is output to the output unit 104, and the process proceeds to step 16.

As described above, according to the air quality detecting apparatus of the ninth embodiment of the present invention, the relationship between the pollution level of each pollution state of the indoor air and the odor declaration value for the odor felt by the person in the room is detected by the neural network. By learning, it is possible to perform more optimal energy-saving automatic ventilation control, etc. that matches the sense of odor of the occupants.

[0106]

As is apparent from the above embodiments, according to the present invention, the background level in the room can be detected with high accuracy, so that the amount of dirt in the indoor air can be detected with high accuracy. It is possible to provide an air quality detection device having an energy-saving effect such as performing optimum automatic ventilation control according to the amount of dirt.

Further, since the reference value can be obtained with high accuracy by learning the reference value with the neural network, the amount of dirt in the room air can be detected with high accuracy. Furthermore, because the standard value is being learned, each room,
It is possible to accurately obtain a background level peculiar to each building or region, and it is possible to provide an air quality detection device having an energy-saving effect such as more optimal automatic ventilation control.

Further, by using the fuzzy method in calculating the air level reference value, the sensor signal when the indoor air is equal to the outside air can be accurately obtained, so that the amount of dirt from when the indoor air is cleaner can be accurately obtained. Can be detected. Furthermore, since the background level and the outside air level in the room can be detected at the same time, it is possible to know the amount of change in the load felt by the person in the room, and the amount of dirt in the room air felt by the person in the room and the actual (outside air level) By performing the automatic ventilation control or the like according to the amount of dirt in the room air (from (1)), it is possible to provide an air quality detection device having an effect of performing comfortable automatic ventilation control for occupants in energy saving.

Further, by learning the reference value 2 by the neural network, the reference value 2 which is the sensor signal when the indoor air is equal to the outside air can be accurately obtained. It can detect well. Furthermore, because the standard value is being learned, each room,
A background level peculiar to each building and each region can be accurately obtained, and an air quality detection device having an energy-saving effect such as more optimal automatic ventilation control can be provided.

In addition, smoking odor, VOC
Since it is possible to accurately determine the state by using odors and body odors as the causes, it is possible to provide an air quality detection device having an energy-saving effect such as optimal automatic ventilation control suitable for the cause (or the state of contamination) of indoor air.

Further, by determining the start of the state from the trend of the sensor signal, that is, the inclination value of the sensor, not only the cause of contamination of the indoor air (or the state of contamination) can be accurately determined, but also quickly (the initial stage of the state). Since the state can be discriminated, it is possible to provide an air quality detection device having an energy saving effect such as automatic ventilation control by predicting the state of contamination of indoor air.

Further, by learning the cause of contamination (or the state of contamination) of indoor air by using the neural network, it is possible to accurately determine the cause of contamination of indoor air that is unique to each room, building, or area. Therefore, it is possible to provide an air quality detection device having an energy-saving effect such as more optimal automatic ventilation control.

Further, since the pollution level for each cause (or pollution state) of the indoor air can be accurately obtained, it is possible to provide an air quality detection device having an energy saving effect such as more optimal automatic ventilation control.

Further, by learning the relation between the pollution level of each pollution state of the indoor air and the odor declaration value for the odor felt by the indoor occupants by a neural network,
It is possible to provide an air quality detection device that has an energy-saving effect such as automatic ventilation control that is more optimal for the sense of odor of the occupants.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an air quality detection device according to a first embodiment of the present invention.

2A is a diagram showing a membership function of a fuzzy rule according to the first embodiment of the present invention; FIG. 2B is a diagram showing a membership function of a fuzzy rule according to the first embodiment of the present invention; The figure which shows the membership function of the fuzzy rule of 1st Example (d) The figure which shows the membership function of the fuzzy rule of 1st Example of this invention.

FIG. 3 is a flowchart showing the operation of the air quality detection device of the first embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration of an air quality detection device according to a second embodiment of the present invention.

FIG. 5 is a diagram showing a neural network according to a second embodiment of the present invention.

FIG. 6 is a flowchart showing the operation of the air quality detection device of the second embodiment of the present invention.

FIG. 7 is a block diagram showing a configuration of an air quality detection device according to a third embodiment of the present invention.

8A is a diagram showing a membership function of a fuzzy rule according to a third embodiment of the present invention; FIG. 8B is a diagram showing a membership function of a fuzzy rule according to the third embodiment of the present invention; The figure which shows the membership function of the fuzzy rule of 3rd Example (d) The figure which shows the membership function of the fuzzy rule of 3rd Example of this invention.

FIG. 9 is a flowchart showing the operation of the air quality detection device of the third embodiment of the present invention.

FIG. 10 is a block diagram showing a configuration of an air quality detecting device according to a fourth embodiment of the present invention.

FIG. 11 is a diagram showing a neural network according to a fourth embodiment of the present invention.

FIG. 12 is a flowchart showing the operation of the air quality detection device of the fourth embodiment of the present invention.

FIG. 13 is a block diagram showing a configuration of an air quality detecting device according to a fifth embodiment of the present invention.

14A is a diagram showing a membership function of a fuzzy rule according to a fifth embodiment of the present invention; FIG. 14B is a diagram showing a membership function of a fuzzy rule according to the fifth embodiment of the present invention; The figure which shows the membership function of the fuzzy rule of 5th Example (d) The figure which shows the membership function of the fuzzy rule of 5th Example of this invention (e) The membership of the fuzzy rule of 5th Example of this invention Diagram showing functions

FIG. 15 is a flowchart showing the operation of the air quality detection device of the fifth embodiment of the present invention.

FIG. 16 is a block diagram showing the configuration of an air quality detecting device according to a sixth embodiment of the present invention.

17A is a diagram showing a membership function of a fuzzy rule according to a sixth embodiment of the present invention; FIG. 17B is a diagram showing a membership function of a fuzzy rule according to the sixth embodiment of the present invention; The figure which shows the membership function of the fuzzy rule of 6th Example (d) The figure which shows the membership function of the fuzzy rule of 6th Example of this invention (e) The membership of the fuzzy rule of 6th Example of this invention Diagram showing functions

FIG. 18 is a flowchart showing the operation of the air quality detecting device of the sixth embodiment of the present invention.

FIG. 19 is a block diagram showing the configuration of an air quality detecting device according to a seventh embodiment of the present invention.

FIG. 20 is a diagram showing a neural network according to a seventh embodiment of the present invention.

FIG. 21 is a flowchart showing the operation of the air quality detection device of the seventh embodiment of the present invention.

FIG. 22 is a block diagram showing the configuration of an air quality detecting device according to an eighth embodiment of the present invention.

23A is a diagram showing a membership function of a fuzzy rule of an eighth embodiment of the present invention; FIG. 23B is a diagram showing a membership function of a fuzzy rule of an eighth embodiment of the present invention; The figure which shows the membership function of the fuzzy rule of 8th Example (d) The figure which shows the membership function of the fuzzy rule of 8th Example of this invention.

FIG. 24 is a flowchart showing the operation of the air quality detecting device of the eighth embodiment of the present invention.

FIG. 25 is a block diagram showing the configuration of an air quality detecting device according to a ninth embodiment of the present invention.

FIG. 26 is a diagram showing a neural network according to a ninth embodiment of the present invention.

FIG. 27 is a flowchart showing the operation of the air quality detection device of the ninth embodiment of the present invention.

FIG. 28 is a diagram showing an evaluation standard table.

FIG. 29 is a block diagram showing the configuration of a conventional air quality detection device.

[Explanation of symbols]

1 sensor 2 signal sampling unit 2a sensor signal input unit 2b sampling signal output unit 3 sensor signal storage unit 3a sampling signal input unit 3b memory signal output unit 4 inclination calculation unit 4a memory signal input unit 4b inclination output unit 5 counting unit 5a inclination input Section 5b slope duration output section 6 background level calculation section 6a sampling signal input section 6b slope duration input section 6c membership function input section 6d backbrand level output section 7 background level calculation fuzzy table storage section 7a membership function output Part 8 Contamination amount 1 calculation unit 8a Sampling signal input unit 8b Back brand level input unit 8c Contamination amount 1 output unit 8d Reference value 1 input unit 17 Reference value 1 learning unit 17a Sampling signal input unit 17b Background level input unit 17c Reference value 1 output section 1 Air level calculation unit 19a Inclination duration input unit 19b Reference value 1 input unit 19c Membership function input unit 19d Air level output unit 20 Air level calculation fuzzy table storage unit 20a Membership function output unit 21 Contamination amount 2 calculation unit 21a Air level Input unit 21b Sampling signal input unit 21c Stain amount 2 output unit 21d Reference value 2 input unit 23 Reference value 2 learning unit 23a Sampling signal input unit 23b Air level input unit 23c Reference value 2 output unit 25 State determination unit 25a Stain amount 1 input Part 25b Dirt amount 2 input part 25c Membership function input part 25d State discrimination signal output part 25e Tilt joint input part 26 State discrimination fuzzy table storage part 26a Membership function output part 28 State discrimination learning part 28a Dirt amount 1 input part 28b Dirt Quantity 2 input section 28c state Separate signal input unit 28d State discrimination learning output unit 30 Contamination level calculation unit 30a Contamination amount 1 input unit 30b Contamination amount 2 input unit 30c Membership function input unit 30d Contamination level output unit 31 Contamination level calculation fuzzy table storage unit 31a Membership function Output unit 33 Contamination level learning unit 33a Contamination level input unit 33b Declaration value input unit 33c Contamination level learning output means 34 Declaration value input unit (external input unit) 34a Declaration value output unit 104 Output unit

Continuation of front page (56) Reference JP-A-1-199142 (JP, A) JP-A-4-270955 (JP, A) JP-A-4-369338 (JP, A) JP-A-5-44973 (JP , A) JP 5-130410 (JP, A) JP 7-309546 (JP, A) JP 62-172252 (JP, A) JP 5-96942 (JP, A) JP 3-134704 (JP, A) JP-A-2-242038 (JP, A) JP-A-63-49653 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) F24F 11/02

Claims (9)

(57) [Claims] 1. One or a plurality of sensors 1 for detecting dirt in the air in a room and outputting a dirt signal, a sensor signal input section for inputting a signal from the one or a plurality of sensors 1, and the sensor signal. A signal sampling unit that samples the signal of the sensor input by the input unit at regular time intervals, a sampling signal output unit that outputs the signal sampled by the signal sampling unit, and a sampling signal output by the sampling signal output unit A sampling signal input unit A, and a sensor signal storage unit that stores the sampling signal of the sensor input by the sampling signal input unit A for a certain time from the past to the present,
A storage signal output unit that outputs a sensor signal for a fixed time stored in the sensor signal storage unit, a storage signal input unit that inputs the storage signal output from the storage signal output unit, and an input using the storage signal input unit The inclination calculation unit for obtaining the sensor signal output variation per unit time (hereinafter referred to as inclination) from the stored signal, the inclination output unit for outputting the inclination calculated by the inclination calculation unit, and the inclination output unit for outputting the inclination. A tilt input unit A for inputting a tilt signal, a counting unit for counting the tilt duration input by the tilt input unit A for each tilt level, and a tilt for outputting the tilt duration counted by the counting unit. A background output fuzzy table and a membership function for calculating the background level are prepared in advance as a background level fuzzy table. A land level calculation fuzzy table storage unit, a membership function output unit that outputs a membership function stored in the background level calculation fuzzy table storage unit, and a sampling signal of the sensor output from the sampling signal output unit. A sampling signal input unit B for input, a slope duration input unit A for inputting the signal output from the slope duration output unit, and a membership function input unit for inputting the membership function output from the membership function output unit. The sensor output in a normal state of indoor air (hereinafter, referred to as background) from the inclination duration input by the inclination duration input section A and the membership function input by the membership function input section. The background for which the value (hereinafter, background level) is calculated Land level calculation unit, background level output unit that outputs the background level signal calculated by the background level calculation unit, sampling signal input unit C that inputs the sampling signal from the sampling signal output unit, and the background A background level input section for inputting a signal of the level output section, and the background level value input by the background level input section is stored as a reference value 1 and the sampling signal of the sensor input by the sampling signal input section C. Comparing with the indoor air pollution amount 1
An air quality detection device comprising: a dirt amount 1 calculation unit for calculating the above; and a dirt amount 1 output unit for outputting the calculated dirt amount to the outside. 2. A background level input section for inputting the background level signal output from the background level output section as a reference value 1 and a sampling signal input section D for inputting the sampling signal of the sensor from the sampling signal output section. The reference value 1 input by the background level input section and the sampling signal of each sensor input by the sampling signal input section D are taken in, and the correlation between the reference value 1 and the sampling signal of the sensor is learned by a neural network. Reference value 1 learning section,
The air quality detection device according to claim 1, further comprising a reference value 1 output unit that outputs the reference value obtained by the learning of the reference value 1 learning unit to the reference value 1 input unit A of the dirt amount 1 calculation unit. 3. An air level calculation fuzzy table storage unit that previously prepares a membership function for calculating the air level as an air level calculation fuzzy table, and members stored in the air level calculation fuzzy table storage unit. A membership function output unit that outputs a ship function, a slope duration input unit B that inputs the signal output from the slope duration time output unit, and a reference value 1 that inputs the reference value 1 output from the reference value 1 output unit Input section B
And a membership function input unit for inputting the membership function output from the membership function output unit, and the slope duration input in the slope duration input unit B and the reference value 1 input unit B are input. Air level calculation to calculate the signal of the sensor (hereinafter referred to as air level) when the indoor air becomes equal to the normal outside air from the reference value and the membership function input in the membership function input section. Section, an air level output section for outputting the air level calculated by the air level calculation section,
The sampling signal input section E for inputting the sampling signal of the sensor from the sampling signal output section and the air level input section for inputting the air level output by the air level output section are provided, and the air input by the air level input section is provided. A fouling amount 2 calculation unit that stores a level signal as a reference value 2 and compares the stored reference value 2 with a sampling signal of the sensor input in the sampling signal input unit E to calculate a fouling amount 2 of indoor air; 2. A stain amount 2 output unit for outputting a stain amount 2; and an output unit for outputting the stain amount 1 from the stain amount 1 output unit and the stain amount 2 from the stain amount 2 output unit to the outside. Or the air quality detection device according to 2. 4. An air level input section for inputting an air level signal output from an air level output section, and a sampling signal input section F for inputting a sampling signal of a sensor from a sampling signal output section. A reference value 2 learning unit that takes in the air level input in the section and the sampling signal of the sensor input in the sampling signal input unit F, and learns the correlation between the reference value 2 and the sampling signal of the sensor by a neural network; The air chamber detection device according to claim 1, 2 or 3, further comprising a reference value 2 output unit that outputs the reference value 2 obtained by the learning of the value 2 learning unit to the dirt amount 2 calculation unit. 5. When an indoor air pollution state is determined to be a background gas, a smoking state, or a person's room, an alcohol sensor is a sensor that detects the background gas and outputs a detection signal, and a smoking state is detected and detected. determination and H ₂ sensor or H _2S sensor or both sensors the sensor, occupancy sensor for outputting a detected detection signal and H _2S sensor or CO ₂ sensor or both sensors human, a state of the room air that outputs a signal A state determination fuzzy table storage unit that prepares a membership function for doing as a state determination fuzzy table in advance, and a membership function output unit that outputs the membership function stored in the state determination fuzzy table storage unit, Dirt amount 1 Dirt amount 1 input from the output unit Dirt amount 1 input unit A and Dirt amount 2 Dirt amount 2 input from the output unit Dirt input 2 input parts A and a membership function input part for inputting the membership function from the membership function output part are provided, and the dirt amount 1 and dirt amount 2 input parts A inputted in the dirt amount 1 input part A are inputted. A state determination section that determines the cause (state) of indoor air pollution from the dirt amount 2 and the membership function that is input by the membership function input section, and a state that outputs a state determination signal of the indoor air that is determined by the state determination section A determination signal output unit, an output signal of the contamination amount 1 from the contamination amount 1 output unit, an output signal of the contamination amount 2 from the contamination amount 2 output unit, and a state determination signal output from the state determination signal output unit. The air quality detection device according to claim 1, 2, 3, or 4, further comprising an output unit that outputs the signal to the outside. 6. An alcohol sensor for detecting a background gas and outputting a detection signal, an H ₂ sensor and an H _2S sensor for detecting a smoking state and outputting a detection signal, and detecting a person's presence in the room and outputting a detection signal. H _{2 S} sensor or CO ₂ sensor, a tilt input section B for inputting a tilt signal from the tilt output section, and a dirt quantity 1 input from a dirt quantity 1 output section A dirt quantity 1 input section A and a dirt quantity 2 output section The dirt amount 2 input section A for inputting the dirt quantity 2 from the above and a membership function input section for inputting the membership function from the membership function output section are provided, and the tilt signal and dirt quantity inputted at the tilt input section B 1 A state judgment that determines the cause (state) of indoor air pollution from the contamination amount 1 and the contamination amount 2 input in the input unit A and the contamination amount 2 input in the input unit A and the membership function input in the membership function input unit Claim 1 and 2 and a part,
The air quality detection device according to 3, 4, or 5. 7. A state determination signal input section for inputting a state determination signal from the state determination signal output section, and a contamination amount 1 input section B and a contamination amount 2 output section for inputting a contamination amount 1 from a contamination amount 1 output section. And a dirt amount 2 input section B for inputting the dirt amount 2 of
The state determination signal input from the state determination signal input section and the contamination amount 1 input section B and the contamination amount 2 input section B contamination amount 1
From the state discrimination learning output unit that outputs the discrimination state obtained by the learning of the state discrimination learning unit, and the state discrimination learning output unit. 6. An output unit for outputting the discrimination state signal, the stain amount 1 from the stain amount 1 output unit, and the stain amount 2 from the stain amount 2 output unit to the outside. Alternatively, the air chamber detection device according to item 6. 8. A pollution level calculation fuzzy table storage unit that prepares a membership function for calculating the pollution level of indoor air as a pollution level calculation fuzzy table in advance, and is stored in the pollution level calculation fuzzy table storage unit. The membership function output section that outputs the membership function and the dirt amount 1 input section C that inputs the dirt quantity 1 from the dirt quantity 1 output section and the dirt quantity 2 that inputs the dirt quantity 2 from the dirt quantity 2 output section Two input parts C and a membership function input part for inputting the membership function from the membership function output part are provided, and the dirt amount 1 and dirt amount 2 input parts C input by the dirt amount 1 input part C are inputted. A pollution level calculation unit for calculating the pollution level of the indoor air from the dirt amount 2 and the membership function input in the membership function input unit; A contamination level output unit that outputs the contamination level calculated by the dye level calculation unit, a contamination amount 1 signal and a contamination amount 2 signal from the contamination amount 1 output unit and the contamination amount 2 output unit, and a state determination learning output unit. An output section for outputting the discrimination state signal and the contamination level signal from the contamination level output section to the outside.
The air quality detection device according to 5, 6, or 7. 9. An evaluation standard table which is a standard for an indoor occupant to evaluate the pollution intensity of the indoor air (or a standard indicating the degree of contamination such as odor intensity) and the indoor air that the indoor occupant feels. Input the declared value input section for inputting the pollution intensity based on the evaluation standard table, the declared value output section for outputting the declared value input in the declared value input section, and the declared value signal from the declared value output section And a pollution level input section for inputting the pollution level from the pollution level output section, wherein the declared value signal input from the declared value input section and the pollution level signal input by the pollution level input section A pollution level learning unit that learns the correlation between the pollution level learning unit and a pollution level learning unit that outputs the pollution level obtained by the learning of the pollution level learning unit, and a contamination amount 1
Output for outputting the contamination amount 1 signal and the contamination amount 2 signal from the output unit and the contamination amount 2 output unit, the discrimination state signal from the state discrimination learning output unit, and the contamination level signal from the contamination level learning output unit to the outside And a section,
The air chamber detection device according to 3, 4, 5, 6, 7 or 8.