JP7726686B2 - Gas detection device, humidity correction method and control device - Google Patents

Description

The present invention relates to a gas detection device, a humidity correction method, and a control device.

In order to distinguish between odors, which are collections of multiple odorous components, there has been active development in recent years of odor sensors that pattern odors by arranging adsorption films with multiple different chemical properties into an array to form a multi-array sensor. Using such odor sensors makes it possible to distinguish between pleasant and unpleasant odors, and applications of odor sensors are beginning to be explored in areas where human sensory evaluation has traditionally been relied upon, such as indoor and vehicle environmental management, factory process management, and environmental monitoring to detect early-stage fires and foul odors that affect the human body.

For example, odor sensors use a detection element with an adsorption film on a quartz crystal oscillator. The resonant frequency of the quartz crystal oscillator decreases in proportion to the weight of odor components adsorbed to the adsorption film, so the amount of odor components can be detected based on the change in resonant frequency. For this reason, a decrease in sensitivity level due to deterioration over time of the adsorption film that adsorbs odor components affects the reliability of the odor sensor. The main cause of deterioration of the adsorption film over time is that adsorbed odor components are unable to desorb from the adsorption film during sensor use and remain trapped in the adsorption film, causing the odor components to accumulate.

For example, Patent Document 1 discloses a method for improving the accuracy of correcting sensor values due to deterioration over time by using a correction means with two stages of parameters depending on the degree of deterioration. Furthermore, Patent Document 2 discloses a method for performing correction by adding a new sensor that serves as a reference for correcting the sensor's response amount.

Patent No. 6730110 Patent No. 6428779

However, the method in Patent Document 1 has parameters with only two stages, making it difficult to flexibly respond to the degree of deterioration, and differences are likely to occur in long-term environments or when parameters change. Furthermore, if an individual sensor exhibits changes over time that are unique to that sensor, the correction will no longer function properly.

On the other hand, the method described in Patent Document 2 requires the preparation of a reference for correcting the gas sensor, which incurs additional costs. Furthermore, the reference used must have the same characteristics as the sensor used for measurement, including its time-dependent change characteristics.

In light of the above circumstances, the object of the present invention is to provide a gas detection device, humidity correction method, and control device that can flexibly respond to changes in the sensor over time while improving the accuracy of correction of measurement values.

A gas detection device according to one aspect of the present invention includes a detection element, a humidity sensor, and a control device.
The detection element generates a change in output upon adsorption of odor components contained in the detection target gas.
The humidity sensor detects the humidity of the detection target gas.
The control device includes an acquisition unit, a storage unit, a measurement unit, and an update unit.
The acquisition unit acquires a measurement value of the detection element and a measurement value of the humidity sensor.
The storage unit stores a correction coefficient.
The measurement unit measures the odor component based on the output of the detection element that has been humidity-corrected using the correction coefficient.
The update unit updates the correction coefficient to an update correction coefficient calculated based on first history data, which is a history of measurement values of the detection element and the humidity sensor.

The measurement value of the detection element may be the minimum or maximum resonant frequency obtained within the measurement period, or the resonant frequency at the end of the measurement period.

The measurement values of the detection element and the humidity sensor may be linked to each other in terms of time.

The control device may further include a determination unit that determines whether the correction coefficient should be updated to the update correction coefficient based on a determination criterion set based on second historical data, which is a history of the correction coefficient.

The determination unit may be configured to set the determination criterion based on the average value of the correction coefficients in the second history data, and to update the correction coefficient to the update correction coefficient only if the update correction coefficient satisfies the determination criterion.

The detection element may include multiple detection elements.

The detection element may be a vibration device having an adsorption film that adsorbs the odor components, and in which adsorption of the odor components causes a change in resonant frequency.

A humidity correction method according to one aspect of the present invention is a humidity correction method for a gas detection device, comprising:
In a first period, a measurement value of a detection element that produces an output change due to adsorption of odor components contained in the detection target gas and a measurement value of a humidity sensor that detects the humidity of the detection target gas are obtained;
measuring the odor component based on the output of the detection element that has been humidity-corrected using a correction coefficient;
An update correction coefficient is calculated based on history data that is a history of measurement values of the detection element and the humidity sensor.

A control device according to one aspect of the present invention is a control device for a gas detection device,
an acquisition unit that acquires, in a first period, a measurement value of a detection element that produces an output change due to adsorption of odor components contained in the detection target gas and a measurement value of a humidity sensor that detects the humidity of the detection target gas;
a storage unit that stores the correction coefficient;
a measurement unit that measures the odor components based on the output of the detection element that has been humidity-corrected using the correction coefficient; and
and an updating unit that updates the correction coefficient to a correction coefficient calculated based on first history data that is a history of measurement values of the detection element and the humidity sensor.

This invention makes it possible to flexibly respond to changes in the sensor over time while improving the accuracy of measurement value correction.

1 is a schematic diagram of a gas detection device according to an embodiment of the present invention. FIG. 2 is a schematic diagram showing an example of the configuration of a detection element. FIG. 10 is an explanatory diagram of a method for calculating a correction coefficient for humidity correction. FIG. 10 is an explanatory diagram of a method for calculating a correction coefficient for humidity correction. FIG. 4 is a schematic diagram showing frequency changes of a detection element. 4 is a flowchart showing an example of a processing procedure of a control device in the gas detection device. FIG. 10 is a diagram illustrating an example of history data of correction coefficients. 10A and 10B are diagrams illustrating a method for determining whether or not an update correction coefficient is adopted. 10A and 10B are diagrams illustrating a method for determining whether or not an update correction coefficient is adopted. 10A and 10B are diagrams illustrating a method for determining whether or not an update correction coefficient is adopted.

Embodiments of the present invention will be described below with reference to the drawings.

[Gas detection device]
FIG. 1 is a schematic diagram of a gas detection device 100 according to this embodiment.
The gas detection device 1 is installed, for example, inside a distribution panel, and is configured as an alarm device that detects specific odor components in gas that is generated when an abnormality such as a cable fire occurs, and issues an alarm, etc.

As shown in FIG. 1, the gas detection device 100 includes a gas sensor 2, a control device 4, and an alarm device 5.

The gas sensor 2 has a housing 20, a first intake port 21, a second intake port 22, and a sensor chamber 60.

The first intake port 21 communicates with the sensor chamber 60 via a first flow path 31. The second intake port 22 communicates with the sensor chamber 60 via a second flow path 32.
First air intake 21 and second air intake 22 are inlets for taking in outside air G into sensor chamber 60. Outside air G is a gas to be detected by gas detection device 100, and in this embodiment, is the air inside the switchboard. In the description of this embodiment, the gas includes moisture and odor components.

A first pump 31P is disposed in the first flow path 31 to draw outside air G from the first intake port 21 into the sensor chamber 60 .
A second pump 32P is disposed in the second flow path 32, which draws outside air G into the sensor chamber 60 from the second air intake port 22. A filter 32F is disposed in the second flow path 32, which adsorbs odor components and moisture contained in the outside air G drawn in from the second air intake port 22 to generate clean air.

In this embodiment of the present invention, clean air refers to gas from which odorous components and moisture contained in the outside air G have been removed using a filter. However, a separate device may be provided to generate clean gas that does not contain odorous components or moisture, and the clean gas may be sent from that device to the sensor chamber 60 as clean air.

The first pump 31P and the second pump 32P are typically configured as diaphragm pumps, but other gas senders that can send or suck gas at a constant flow rate, such as fans or blowers, can also be used.
The first pump 31P and the second pump 32P are driven individually by the control device 4, and by driving either one of them, either the first flow path 31 or the second flow path 32 is selected as the outside air intake flow path.

The filter 32F can be made of materials with high humidity removal properties, such as silica gel or molecular sieves, to remove humidity and hydrophilic odor components, or materials with high odor absorption properties, such as activated carbon or zeolite, to remove odor components, or a combination of these.

The sensor chamber 60 houses multiple QCM (Quartz Crystal Microbalance) sensors 10 as detection elements and a humidity sensor 70.

The sensor chamber 60 forms a chamber into which outside air G is taken in and measurements are performed. The sensor chamber 60 is connected to the first flow path 31 and the second flow path 32, and outside air G is introduced into the sensor chamber 60 via the first air intake 21 or the second air intake 22. Although not shown, the sensor chamber 60 is connected to an exhaust passage that exhausts the introduced gas.

The QCM sensor 10 is a detection element whose resonant frequency changes when it adsorbs odor components contained in the outside air G. Each QCM sensor 10 constitutes a detection channel (ch). The number of QCM sensors 10 corresponds to the number of channels, and this number is not particularly limited. In this embodiment, an example is given in which 16 QCM sensors 10 are provided, but at least one channel is sufficient.

An oscillator circuit 50 is connected to each QCM sensor 10. Each oscillator circuit 50 is driven by the control device 4. Each oscillator circuit 50 vibrates each QCM sensor 10 at a predetermined resonant frequency. The oscillator circuit 50 is connected to a counter circuit (not shown) that counts the oscillation frequency of the QCM sensor 10. The counter circuit counts changes in the resonant frequency of each QCM sensor 10 and outputs the count value to the control device 4. In this embodiment, the measurement value of each QCM sensor 10 is the count value output by the counter circuit.

The humidity sensor 70 measures the relative humidity of the outside air G. The measurement value output from the humidity sensor 70 is output to the control device 4. The control device 4 corrects the measurement value of each QCM sensor 10 based on the humidity measured by the humidity sensor 70, as will be described later.
The humidity sensor 70 can be a conventional type such as a capacitance type that detects humidity based on the output value of capacitance or a semiconductor type that detects humidity based on the output value of resistance, but is not limited to these. A hydrophilic adsorption film may be formed on the QCM sensor, and the humidity of each QCM sensor 10 may be corrected based on the change in resonant frequency when odor components are adsorbed.

The sensor chamber 60 may further include a temperature sensor that detects the temperature of the outside air G. In this case, based on the temperature detected by the temperature sensor, the temperature-induced change in resonant frequency detected by each QCM sensor 10 can be canceled out, and the resonant frequency change of each QCM sensor 10 that is not affected by temperature can be detected.

Each QCM sensor 10, each oscillator circuit 50, and humidity sensor 70 may be mounted on a common wiring board, or may each be mounted on a separate wiring board.

Each QCM sensor 10 is configured with a quartz crystal oscillator as an oscillator and an adsorption film disposed on the quartz crystal oscillator that adsorbs odor components contained in the outside air G. The basic structure is the same, with the only difference being the type of adsorption film.

As shown in FIG. 2, the QCM sensor 10 includes a quartz crystal oscillator 13, an electrode 11, an adsorption film 12, lead lands 16A and 16B, leads 14A and 14B, pin terminals 19A and 19B, and a holder 18. The quartz crystal oscillator 13 is, for example, an AT-cut quartz crystal plate. The resonant frequency of the quartz crystal oscillator 13 is not particularly limited and is, for example, 32 MHz.

The resonant frequency of the quartz crystal oscillator 13 of the QCM sensor 10 decreases in proportion to the increase in the weight of the gas adsorbed by the adsorption film 12. Therefore, the change in resonant frequency is calculated for each QCM sensor 10 to detect the presence or absence of odor components in the outside air G, or to measure the amount or concentration of one or more odor components. Furthermore, based on the calculation results, the type or intensity of the odor contained in the outside air G can be determined.

In this embodiment, a quartz oscillator is used as the oscillator of the detection element, but this is not limited to this. For example, instead of a quartz oscillator, other oscillator devices such as a ceramic oscillator, a surface acoustic wave element, a piezoelectric thin film resonator, a cantilever, or a diaphragm can also be used.

Electrodes 11 are formed on both sides of the quartz crystal oscillator 13, and an adsorption film 12 is formed on the electrode 11 formed on one side of the quartz crystal oscillator 13. Lead land 16A is formed integrally with the electrode 11 formed on one side, and lead land 16B is formed integrally with the electrode 11 formed on the other side.

The adsorption film 12 of each QCM sensor 10 is made of a different type of material. The material that makes up the adsorption film 12 is selected arbitrarily depending on the type of odor component to be detected. The adsorption film 12 is configured to be able to adsorb one or more types of odor component.

Leads 14A and 14B are made of a metal spring material. One end of lead 14A is electrically connected to electrode 11 formed on one surface via lead land 16A, and the other end is connected to pin terminal 19A. One end of lead 14B is electrically connected to electrode 11 formed on the other surface via lead land 16B, and the other end is connected to pin terminal 19B.
The pin terminals 19A and 19B are supported by a holder 18 provided on the wiring board, and the holder 18 supports the crystal oscillator 13 so that it can vibrate freely.

The control device 4 detects the presence or absence of odor components in the outside air G, which is the gas to be detected, based on the measurement values of each QCM sensor 10, or measures the amount or concentration of one or more odor components.
In this case, the adsorption film 12 of each QCM sensor 10 adsorbs not only odor components but also moisture contained in the detection target gas, so the measurement value of each QCM sensor 10 also contains information about frequency changes due to moisture adsorption.
Therefore, in order to accurately measure only the frequency change due to the adsorption of odor components, the control device 4 performs a correction (hereinafter also referred to as humidity correction) to remove the humidity component from the measurement value of each QCM sensor 10.
Furthermore, the measurement device 4 is configured to be able to periodically update the correction coefficient for humidity correction in order to suppress a decrease in measurement accuracy due to deterioration over time of the adsorption film 12 of each QCM sensor 10 .

[Control device]
The control device 4 will be described in detail below.
The control device 4 is an information processing device having a CPU and memory. The control device 4 controls the driving of the first pump 31P, the second pump 32P, the oscillation circuit 50, the humidity sensor 70, etc., and acquires information on the amount of change in the resonant frequency of each QCM sensor 10 and the humidity of the outside air G from the oscillation circuit 50 and the humidity sensor 70.

As shown in FIG. 1, the control device 4 has, as CPU functional blocks, an acquisition unit 41, a measurement unit 42, a calculation unit 43, an update unit 44, and a determination unit 45. The control device 4 further includes a storage unit 46 configured using a semiconductor memory or the like.

(Storage part)
The storage unit 46 stores a program for causing the acquisition unit 41, the measurement unit 42, the calculation unit 43, the determination unit 45, and the update unit 44 to operate as functional blocks of a CPU.

The memory unit 46 stores the resonant frequency of each QCM sensor 10 acquired by the acquisition unit 41 when the adsorption film 12 and sensor chamber 60 are cleaned with clean air. Cleaning the adsorption film 12 refers to desorbing odor components from the adsorption film 12 with clean air. This clean air can be obtained by taking in outside air G through the second air intake 22 and introducing it into the sensor chamber 60 through the filter 32F of the second flow path 32. The resonant frequency of each QCM sensor 10 when the adsorption film 12 and sensor chamber 60 are cleaned with clean air is hereinafter also referred to as the frequency zero point of each QCM sensor 10. The zero point is acquired in advance for each QCM sensor 10 and stored in the memory unit 46.

The memory unit 46 stores the measurement values of each QCM sensor 10 acquired by the acquisition unit 41. The measurement values of each QCM sensor 10 correspond to the amount of change in the resonant frequency of each QCM sensor 10 when outside air G is introduced into the sensor chamber 60 from the first intake port 21 through the first flow path 31.

The storage unit 46 stores the measurement value of the humidity sensor 70 acquired by the acquisition unit 41 .
The memory unit 46 temporally links the measurement values of each QCM sensor 10 with the measurement values of the humidity sensor 70, and stores historical data (hereinafter also referred to as first historical data) which is the history of the measurement values of each of these QCM sensors 10 and humidity sensors 70.

The storage unit 46 associates the correction coefficients calculated by the calculation unit 43 with information on the date and time of the calculation, and stores history data (hereinafter also referred to as second history data) that is a history of these.
The storage unit 46 also stores in advance reference detection patterns for each different type of odor when various odors are detected in the sensor chamber 60 .

The memory unit 46 stores an initial value of a unique correction coefficient for each QCM sensor 10. The initial value of the correction coefficient is a correction coefficient that is referenced when an unused gas detection device 100 is used for the first time. This correction coefficient is calculated based on the humidity-dependent characteristics of each QCM sensor 10 that have been measured in advance in an odor-free environment.

(Method for determining the initial value of the correction coefficient)
Here, a method for calculating the initial values of the correction coefficients will be described.
When calculating the initial values of the correction coefficients, reference is made to the time-varying output of each QCM sensor 10 and humidity sensor 70 measured using air as the gas. An example of the measurement results is shown in FIG.
In FIG. 3, the horizontal axis indicates measurement time, the left vertical axis indicates the resonance frequency of the QCM sensor 10, and the right vertical axis indicates the relative humidity of the atmosphere, which is the gas to be detected by the humidity sensor 70.
By measuring the output pattern of each QCM sensor 10 using a gas that does not contain odor components, the relationship between the output of each QCM sensor 10 and the environmental factor (humidity in this example) can be determined.

The resonant frequency of each QCM sensor 10 gradually decreases from zero depending on the amount of moisture adsorbed in the gas, and vibrates at a constant resonant frequency after a predetermined time has elapsed since the start of measurement. In the example shown, the QCM sensor 10 for detection channel number 8, "ch8," has the largest frequency change. This indicates that the adsorption film 12 of the QCM sensor 10 for "ch8" is more likely to adsorb moisture than the adsorption films 12 of the other channels, and has a higher affinity for moisture.

On the other hand, with the humidity sensor 70, as the humidity in the measurement environment increases with the start of air introduction, the detected humidity gradually increases from the start of measurement.

3, "dch" is the frequency variation from the zero point of each QCM sensor 10, and "dch min" is the minimum value of the resonant frequency of each QCM sensor 10. "ch8 min" means the minimum value of the resonant frequency of the QCM sensor 10 of "ch8".
Furthermore, "dhumidity" (hereinafter abbreviated as "dhmd") is the measurement value of the humidity sensor 70, and "dhmd min" is the measurement value of the humidity sensor 70 when the resonant frequency of the QCM sensor 10 of "ch8", which has the largest frequency fluctuation among all the QCM sensors 10, shows the minimum value.

The data shown in Figure 3 shows how, when air is exposed to the QCM sensor 10, odor components contained in the air are adsorbed by the QCM sensor 10, thereby lowering its resonant frequency. As described above, the measurement value of the humidity sensor 70 is adopted when the resonant frequency of the QCM sensor 10 for the detection channel with the largest frequency fluctuation is at its minimum. In the case shown in Figure 3, the measurement value of the humidity sensor is adopted when the resonant frequency of the QCM sensor 10 for "ch8" with the largest frequency fluctuation is at its minimum.
On the other hand, there are cases where the moisture adsorbed to the QCM sensor 10 is desorbed by the atmosphere. In this case, the resonant frequency of the QCM sensor 10 increases, and therefore the measurement value of the humidity sensor 70 is adopted as the measurement value of the humidity sensor when the resonant frequency of the QCM sensor 10 of the detection channel with the largest frequency fluctuation reaches its maximum value.
Thus, when the air is applied to the QCM sensor 10, the resonant frequency of the QCM sensor 10 may either decrease or increase, so that the absolute values of the minimum and maximum resonant frequencies of the QCM sensor 10 in the detection channel with the largest frequency fluctuation may be compared, and the measurement value showing the largest of these may be used as the reference measurement value when determining the measurement value of the humidity sensor 70. As yet another method, the measurement value of the QCM sensor 10 in the detection channel with the largest frequency fluctuation at the end of the measurement period may be used.

Next, the difference between the humidity at the start of measurement and the measurement value of the humidity sensor 70 at any measurement time is taken as the humidity difference, and the frequency fluctuation of each QCM sensor 10 against the humidity difference is plotted in Figure 4. As shown in Figure 4, there is a very clear linear relationship between the frequency fluctuation of each QCM sensor 10 and the humidity difference, as shown in the following equation.
(dch)=a1×(dhmd)+b1…(1)

In the above equation, the constants a1 and b1 are unique values for each QCM sensor 10, and typically, the constant a1 corresponds to the initial value of the correction coefficient. Note that, in addition to the constant a1, the constant b1 and coefficients of equations other than the linear equation may also be used for humidity correction.
The above correction formula is obtained by regression analysis from the frequency fluctuation (dch) of the QCM sensor 10 obtained by multiple measurements and the measurement value (dhmd) of the humidity sensor 70.

(Acquisition Department)
The acquisition unit 41 acquires the measurement values of each QCM sensor 10 and the measurement value of the humidity sensor 70 .
More specifically, the acquisition unit 41 acquires the zero point of the resonant frequency of each QCM sensor 10 when the adsorption film 12 and the sensor chamber 60 are cleaned by introducing clean air. Hereinafter, this process is also referred to as a refresh process. The refresh process is performed before or after each QCM sensor 10 measures the ambient air G. The refresh process lasts, for example, one minute.

The acquisition unit 41 acquires the measurement values of each QCM sensor 10 when outside air G is introduced into the sensor chamber 60 in a first cycle. Hereinafter, this process is also referred to as the gas measurement process. The first cycle is not particularly limited and is, for example, one hour. In this embodiment, the maximum change in the resonant frequency of each QCM sensor 10 during each gas measurement process is acquired as the measurement value of each QCM sensor 10.

In the first cycle, the acquisition unit 41 acquires the measurement value of the humidity sensor 70. In this embodiment, as described with reference to Fig. 3, during each gas measurement process, the output of the humidity sensor 70 when the resonant frequency of the QCM sensor 10 with the largest frequency change among the QCM sensors 10 is at its minimum value is acquired as the measurement value of the humidity sensor 70.
In this case, as described above, the larger of the absolute value of the minimum or maximum value of the resonant frequency of the QCM sensor 10 of the detection channel with the largest frequency change may be used, or the measurement value at the end of the measurement period may be used.

FIG. 5 is a schematic diagram showing an example of the change in output of the QCM sensor 10 during the refresh process and the gas measurement process.
As shown in the figure, the control device 4 executes a process of alternately performing a refresh process and a gas measurement process.

During the refresh process period before the gas measurement process, the adsorption film 12 of each QCM sensor 10 and the sensor chamber 60 are thoroughly cleaned with clean air, and environmental factors such as the vibration frequency of each QCM sensor 10 and the humidity of the sensor chamber 60 are set to zero. The refresh process period at this time is, for example, 20 minutes.
The gas measurement process is performed after the refresh process. The illustrated example shows changes in the resonant frequencies of three QCM sensors 10, ch1, ch2, and ch3, among the 16-channel QCM sensors 10. The minimum frequency value indicating the maximum change from the zero point of the resonant frequency of each channel is included in the measurement value of each QCM sensor 10. The gas measurement process is performed for a predetermined time (e.g., one minute) in a first cycle, for example, every hour.

To give a specific example, the first cycle would consist of 20 minutes of refresh processing before the gas measurement processing, followed by 1 minute of gas measurement processing, followed by 1 minute of refresh processing, followed by a 38-minute wait period. The 38-minute wait period is a period during which the pump is not driven and refresh processing and gas measurement processing are not performed. After that, the next cycle begins, and the refresh processing before the gas measurement processing begins. The first cycle is not limited to 1 hour. The gas measurement processing, refresh processing, and wait period can also be set arbitrarily. The first cycle may also be performed by sequentially performing the gas measurement processing and refresh processing without a wait period.

After the gas measurement process is completed, the refresh process is performed again. This causes the odor components and moisture adsorbed on the adsorption film 12 of each QCM sensor 10 to desorb, and the resonant frequency of each QCM sensor 10 returns to zero. The refresh process period is not particularly limited, and may be continued after the gas measurement process is completed until just before the next gas measurement process.

The zero point of the vibration frequency of each QCM sensor 10 is 32 MHz, which is the natural frequency of the quartz oscillator 13, for QCM sensors 10 that have not deteriorated over time. On the other hand, if desorption of odor components and moisture is insufficient due to deterioration of the adsorption film 12 over time, the zero point of the QCM sensor 10 will be a frequency lower than the natural frequency of the quartz oscillator 13.

(Measurement part)
During the gas measurement process, the measurement unit 42 uses the correction coefficient to humidity-correct the measurement values of each QCM sensor 10 and measure the type and amount or concentration of odor components in the target gas, i.e., outside air G. The odor components are those that are adsorbed by at least one adsorption film 12 of each QCM sensor 10 and can be identified by the reference detection pattern stored in the memory unit 46.

As described above, there is a linear relationship between humidity and the frequency fluctuation of the QCM sensor 10. If the frequency fluctuation due to the excess humidity included in the measurement value of the QCM sensor 10 is defined as "Δchhumidity" and the difference between the measurement value of the humidity sensor 70 at the start of measurement and the measurement value of the humidity sensor 70 at the time of acquisition of the measurement value of the QCM sensor 10 is defined as "ΔHumidity," the relationship between the two can be expressed as follows:
Δchhumidity=a2×ΔHumidity+b2…(2)
The constants a2 and b2 have unique values for each QCM sensor 10, and typically, the constant a2 is used as the correction coefficient. In addition to the constant a2, the constant b2 and coefficients of equations other than the linear equation may also be used for humidity correction.

The humidity correction of the measurement values of each QCM sensor 10 performed by the measurement unit 42 is assumed to include an extra fluctuation due to humidity in the measurement values of each QCM sensor 10. Therefore, if the measurement value of the QCM sensor 10 before humidity correction is Δchraw and the measurement value of the QCM sensor 10 after humidity correction is Δchsample, the measurement unit 42 performs humidity correction of the measurement values of the QCM sensors 10 by executing the calculation process shown in the following equation.
Δchsample=Δchraw−Δchhumidity…(3)

The measurement unit 42 compares the combination pattern of the humidity-corrected measurement values of each QCM sensor 10 with a reference detection pattern pre-stored in the memory unit 46, and evaluates the type and intensity of the odor using pattern recognition based on machine learning, etc.
When the measuring unit 42 detects an odor component that indicates an abnormality, or when it evaluates that the concentration of the odor component is equal to or higher than a predetermined concentration, it generates an abnormality signal and outputs it to the alarm device 5 .

The alarm device 5 has alarm devices such as a buzzer and lamp. The alarm device 5 receives an abnormality signal generated when an odor component is detected from the control device 4, and sounds a buzzer or lights or blinks a lamp. This allows an abnormality in the distribution board to be notified to the outside.

(Calculation unit)
The correction coefficient used for humidity correction of the QCM sensor 10 changes over time. That is, because the gas measurement process is repeated at a fixed cycle (every hour in this embodiment), the cleanliness level of the adsorption film 12 in the subsequent refresh process gradually decreases. As a result, as the number of days passes from the start of measurement, the fixed correction coefficient becomes unable to perform appropriate humidity correction, which may result in an erroneous evaluation of odor components. Specifically, there is a possibility that an abnormality may be determined when no abnormality has occurred, or that no abnormality has occurred when an abnormality has occurred.

To prevent such problems, the control device 4 of this embodiment is configured to be able to update the correction coefficient. The calculation unit 43 is configured to execute a process to periodically calculate an updated correction coefficient.

The calculation unit 43 calculates a correction coefficient based on first history data, which is a history of measurement values of each QCM sensor 10 and humidity sensor 70 stored in the memory unit 46. The calculation method for the correction coefficient is similar to the calculation method for the initial correction coefficient described with reference to Figures 3 and 4, but when calculating the updated correction coefficient, the measurement values of each QCM sensor 10 acquired during the gas measurement process are used as the measurement values of each QCM sensor 10. This eliminates the need to prepare a new atmospherically adjusted measurement environment, making it possible to calculate a new correction coefficient while continuing regular abnormality monitoring operation by the gas detection device 100.

The calculation unit 43 executes a process to calculate an update correction coefficient (hereinafter also referred to as an update correction coefficient) at a second period that is longer than the first period (one hour in this example), which is the operating period of the gas measurement process. The second period is not particularly limited, and is, for example, one day (24 hours).

When calculating the correction coefficient, the calculation unit 43 uses the most recent historical data from the historical data (first historical data) of the measurement values of each QCM sensor 10 and the measurement values of the humidity sensor 70 associated with them for the past 24 hours to calculate the correction formula shown in equation (1) above. The number of recent historical data is not particularly limited, and in this embodiment, historical data from the past 24 hours, i.e., the measurement values of each QCM sensor 10 and humidity sensor 70 obtained during the most recent 24 gas measurement processes, is used.

(Update Department)
The update unit 44 is configured to be able to execute a process of updating, at a second cycle (every 24 hours), the correction coefficients used in the humidity correction process in the measurement unit 42 to the update correction coefficients calculated by the calculation unit 43. This makes it possible to suppress a decrease in the accuracy of the measurement values of each QCM sensor due to deterioration of the adsorption film 12 over time, and ensure proper anomaly detection operation over a long period of time.

In this embodiment, as described below, the current correction coefficient is updated to the update correction coefficient only when the determination unit 46 allows the update.

(Judgment Department)
Ideally, the data referenced when calculating the updated correction coefficients would be measurements of the QCM sensors 10 in which there is no odor and frequency fluctuations are caused purely by humidity fluctuations. In this embodiment, the updated correction coefficients are calculated based on measurements of each QCM sensor 10 acquired during the gas measurement processing period. However, these measurements are not necessarily data measured in an odor-free environment. For example, there are times when odor components are detected but not enough to warrant evaluation of an abnormal state. Using correction coefficients calculated based on data measured under such conditions could have a significant impact on subsequent anomaly detection processing.

Therefore, in this embodiment, the newly calculated update correction coefficient is not adopted as is, but rather a determination unit 45 is provided that determines whether the calculated update correction coefficient is appropriate based on the history of past correction coefficients.

The determination unit 45 is configured to determine whether or not the current correction coefficient should be updated to the update correction coefficient calculated by the calculation unit 42 based on a determination criterion set based on second history data, which is the past history of the correction coefficients stored in the memory unit 46.
The judgment criteria are to calculate a predicted value of the next correction coefficient from past correction coefficients and its allowable range. The predicted value may be, for example, the average value of past correction coefficients, and the allowable range may be, for example, the standard deviation of the correction coefficient history multiplied by a certain coefficient.

[Operation of gas detection device]
Details of the control device 4 will be described below along with a typical operation of the gas detection device 100. Fig. 6 is a flowchart showing an example of a processing procedure of the control device 4.

(Gas measurement processing)
Control device 4 periodically executes gas measurement processing in a first cycle after starting operation of gas detection device 100. In this embodiment, the gas measurement processing is executed in the first cycle, i.e., every hour (steps 101 to 104).

As described above, in the gas measurement process, the first pump 31P is driven to draw in outside air G (air inside the switchboard) through the first intake port 21 and introduce it into the sensor chamber 60 through the first flow path 31. The control device 4 acquires the measurement values of each QCM sensor 10 and humidity sensor 70 and stores them in the memory unit 46 (step 102).

The control device 4 then performs humidity correction on the measurement values of each QCM sensor 10 using the relative humidity of the outside air G acquired by the humidity sensor 70 and the correction coefficient stored in the memory unit 46 (step 103). If the correction coefficients have not yet been updated, the initial values of the correction coefficients measured in advance as described with reference to Figures 3 and 4 are used.

The control device 4 compares the detection pattern, which is a combination of the humidity-corrected measurement values of each QCM sensor 10, with a reference detection pattern pre-stored in the memory unit 46, and evaluates the type or intensity of the odor component using pattern recognition based on machine learning or the like (step 104).
If there is no odor component indicating an abnormality or if the concentration is less than a predetermined value, it is determined that there is no odor and that the condition is normal ("Y" in step 105).
On the other hand, if the concentration of the odor component indicating an abnormality is evaluated to be equal to or greater than the predetermined value, it is determined to be abnormal ("N" in step 105), and an abnormality signal is generated and output to the alarm device 5 (step 106).

After completing the gas measurement process described above, the control device 4 executes a refresh process. During the refresh process, the first pump 31P is stopped, and the second pump 32P is driven to introduce outside air G taken in through the second intake port 22 into the sensor chamber 60 through the second flow path 32 and filter 32F. This cleans the adsorption film 12 of each QCM sensor 10 and the inside of the sensor chamber 60, and returns the resonant frequency of each QCM sensor 10 to zero (see Figure 5).

After the refresh process is completed, the control device 4 performs an operation (hereinafter also referred to as steady operation) in which the above-described gas measurement process and refresh process are repeated at a first cycle, and stores historical data for each gas measurement process, including the measurement values of each QCM sensor 10 and humidity sensor 70, the measurement values of each QCM sensor 10 after humidity correction, and odor evaluation results, as first historical data in the memory unit 46.

(Correction coefficient calculation process)
The control device 4 first executes the gas measurement process, and then periodically executes the calculation process of the update correction coefficients in the second cycle. In this embodiment, the calculation process of the update correction coefficients is executed in the second cycle, that is, every 24 hours (steps 107 to 111). The calculation process of the update correction coefficients is executed for each QCM sensor 10.

The calculation process for the update correction coefficients uses the correction formula shown in equation (1) above to calculate the update correction coefficients based on the measurement values of the QCM sensor 10 and humidity sensor 70 for the past 24 hours stored in the memory unit 46 (step 108). As described below, the control unit 4 determines whether the calculated update correction coefficients are within the allowable range (step 109). If they are within the allowable range, the control unit 4 updates the current correction coefficients with the calculated update correction coefficients, and from the next gas measurement process, humidity correction processing for the measurement values of each QCM sensor 10 is performed using the updated correction coefficients.

On the other hand, if the calculated update correction coefficient is outside the allowable range, the calculated update correction coefficient is discarded and the current correction coefficient continues to be used to perform humidity correction processing on the measurement values of each QCM sensor 10.

The above-described calculation process and update process of the update correction coefficient are repeatedly executed in a second cycle, and historical data including the calculation results of the update correction coefficient for each cycle and the updated values of the correction coefficients are stored in the memory unit 46 as second historical data.

(Decision on whether or not to adopt the update correction coefficient)
The determination of whether or not the update correction coefficient is adopted, which is executed in the process of step 109, will be described below.

Figure 7 shows an example of the correction coefficient history. The memory unit 46 stores historical data on the correction coefficient from the start of operation of the gas detection device 100 to the present for each QCM sensor 10. As shown in the figure, the value and trends of the correction coefficient differ for each QCM sensor, and it can be seen that the correction coefficient for "ch4" in particular fluctuates significantly each time it is updated.

Figure 8 is an explanatory diagram of the criteria for determining whether or not to accept an updated correction coefficient. In this figure, the horizontal axis represents the number of days, and the vertical axis represents the correction coefficient. The history of correction coefficients updated on a daily basis is shown as a line graph, and the circled points represent the newly calculated correction coefficients (the same applies to Figures 9 and 10).

In the illustrated example, the criteria are set to a range of ±3σ, which is three times the standard deviation of the average value of past correction coefficients, and this range is set as the allowable range for the correction coefficient to be updated. Therefore, in the illustrated example, since the currently calculated update correction coefficient is within the allowable range, the current correction coefficient is updated to that update correction coefficient (steps 109 and 110 in Figure 6). This type of determination is performed individually for each QCM sensor 10.

Since the allowable range reflects the previous correction coefficient, the average value of the correction coefficient can be regarded as a predicted value of the correction coefficient to be updated next time. In other words, since the average value of the correction coefficient varies depending on the number of days, and therefore the allowable range also constantly fluctuates, it is possible to update the correction coefficient to an appropriate value that reflects the actual situation.
The past correction coefficients referenced when calculating the update correction coefficients may be those from a predetermined period in the past (for example, within the past two weeks). In this case, the update correction coefficients are calculated using relatively recent past correction coefficients, making it easier to obtain correction coefficients that are in line with the actual situation.

On the other hand, as shown in Figure 9, if the newly calculated update correction coefficient falls outside the allowable range, the control device 4 determines that the update correction coefficient is inappropriate and discards it (steps 109 and 111 in Figure 6).

If some abnormality has occurred in the QCM sensor 10 within the past 24 hours, such as deterioration of the adsorption film 12, detection of a low concentration of a specific odor component, or detection of an unknown odor component, the updated correction coefficient calculated based on this data is likely to vary significantly from the previous correction coefficient. Therefore, since adopting such a correction coefficient outside the allowable range as a new correction coefficient could lead to an erroneous evaluation of the odor measurement, it was decided not to adopt the updated correction coefficient calculated this time.
Furthermore, even if the previously calculated update correction coefficient is outside the allowable range, the next calculated update correction coefficient may fall within the allowable range, as shown in FIG.
Therefore, if it is determined that the calculated update correction coefficient is outside the allowable range, the current correction coefficient is typically continued to be used. However, this is not limited to this, and for example, the average value of the correction coefficients for the most recent few days within the allowable range may be updated as the update correction coefficient.

If an update correction coefficient outside the allowable range is calculated multiple times in succession, the average value of past correction coefficients may be used, as described above.

As described above, in this embodiment, the measurement values of the QCM sensor 10 and humidity sensor 70 obtained during steady-state operation of the gas detection device 100 are numerically processed, the correction coefficients are updated periodically, and the predicted value and range of the next correction coefficient are calculated from the history of the obtained correction coefficients. Only if the value obtained by the next measurement falls within this range is the new correction coefficient adopted as the latest correction coefficient. This makes it possible to flexibly respond to changes in the sensors over time while improving the correction accuracy of the measurement values.

Furthermore, according to this embodiment, the correction coefficients are automatically calculated and updated, enabling stable, highly accurate anomaly detection over a long period of time. In particular, it makes it easier to check for abnormalities in places where people do not normally enter, such as distribution boards, thereby reducing the burden on workers.

The above describes an embodiment of the present invention, but it goes without saying that the present invention is not limited to the above embodiment and various modifications can be made.

For example, in the above embodiments, a QCM sensor was used as an example of an odor component detection element, but this is not limited to this. Other detection elements, such as semiconductor sensors, whose resistance value changes upon adsorption of odor components, may also be used. Furthermore, the detection element may be one that can detect physical changes, such as an increase in weight or expansion stress due to gas adsorption in the adsorption film, and convert these changes into an electrical signal.

Furthermore, the gas detection device according to the present invention is not limited to applications in detecting abnormalities such as fires. For example, it may be used as a sensor to detect unpleasant odor components in a room and initiate ventilation operations. The present invention can also be applied to analyzing odor components contained in unknown gases.

Furthermore, in the above embodiment, the measurement value of the humidity sensor 70 was determined based on the change in the resonant frequency of the QCM sensor 10 with the largest frequency fluctuation, and the measurement value of that humidity sensor 70 was commonly used to perform humidity correction on the measurement values of the QCM sensor 10 of each detection channel, but this is not limited to this. For example, the measurement value of the humidity sensor 70 may be determined individually for each detection channel based on the minimum or maximum value of the resonant frequency of each QCM sensor 10, and the humidity sensor measurement value determined for each detection channel may be used to perform humidity correction on the measurement values of the QCM sensor 10 of each detection channel individually.

4...Control device 10...QCM sensor (detection element)
DESCRIPTION OF SYMBOLS 12: Adsorption film 41: Acquisition unit 42: Measurement unit 43: Calculation unit 44: Update unit 45: Determination unit 46: Storage unit 70: Humidity sensor 100: Gas detection device

Claims (8)

a detection element that produces an output change upon adsorption of odor components contained in the detection target gas;
a humidity sensor for detecting the humidity of the detection target gas;
A control device and
The control device
an acquisition unit that acquires a measurement value of the detection element and a measurement value of the humidity sensor;
a calculation unit that calculates a correction coefficient based on first history data that is a history of measurement values of the detection element and the humidity sensor, and calculates an update correction coefficient using history data most recent from the first history data;
a measurement unit that measures the odor components based on the output of the detection element that has been humidity-corrected using the correction coefficient; and
a determination unit that calculates a predicted value of the update correction coefficient and a range of the predicted value based on second history data that is a history of the correction coefficient, and determines whether the update correction coefficient is within the range of the predicted value;
an updating unit that updates the correction coefficient to the update correction coefficient when the update correction coefficient is within the range of the predicted value. 2. The gas detection device according to claim 1,
A gas detection device, wherein the measurement value of the detection element used as a reference when determining the measurement value of the humidity sensor is the minimum or maximum value of the resonance frequency acquired within a measurement period, or the resonance frequency at the end of the measurement period. 3. The gas detection device according to claim 1 or 2,
The gas detection device, wherein the measurement values of the detection element and the humidity sensor are time-linked to each other. 2. The gas detection device according to claim 1 ,
The predicted value of the update correction coefficient is the average value of the correction coefficient in the second history data, and the range of the predicted value is obtained by multiplying the standard deviation of the history of the correction coefficient by a certain coefficient.
Gas detection equipment. The gas detection device according to any one of claims 1 to 4 ,
The gas detection device includes a plurality of detection elements. The gas detection device according to any one of claims 1 to 5 ,
The gas detection device, wherein the detection element is a vibration device having an adsorption film that adsorbs the odor component, and the adsorption of the odor component causes a change in resonance frequency. 1. A humidity correction method for a gas detection device, comprising:
A measurement value of a detection element that produces an output change due to adsorption of odor components contained in the detection target gas and a measurement value of a humidity sensor that detects the humidity of the detection target gas are obtained;
calculating a correction coefficient based on first history data which is a history of measurement values of the detection element and the humidity sensor, and calculating an updated correction coefficient using the most recent history data of the first history data;
measuring the odor component based on the output of the detection element that has been humidity-corrected using the correction coefficient;
calculating a predicted value of the update correction coefficient and a range of the predicted value based on second history data which is a history of the correction coefficient, and determining whether the update correction coefficient is within the range of the predicted value;
If the update correction coefficient is within the range of the predicted value, the correction coefficient is updated to the update correction coefficient.
Humidity correction method. 1. A control device for a gas detection device, comprising:
an acquisition unit that acquires a measurement value of a detection element that produces an output change due to adsorption of odor components contained in the detection target gas and a measurement value of a humidity sensor that detects the humidity of the detection target gas;
a calculation unit that calculates a correction coefficient based on first history data that is a history of measurement values of the detection element and the humidity sensor, and calculates an update correction coefficient using history data most recent from the first history data;
a measurement unit that measures the odor components based on the output of the detection element that has been humidity-corrected using the correction coefficient; and
a determination unit that calculates a predicted value of the update correction coefficient and a range of the predicted value based on second history data that is a history of the correction coefficient, and determines whether the update correction coefficient is within the range of the predicted value;
an updating unit that updates the correction coefficient to the update correction coefficient when the update correction coefficient is within the range of the predicted value.