Gas detection equipment

Abstract

To provide a gas detector including identification processing of which a processing load is lower than that of the case of using machine learning.SOLUTION: A gas detector 100 comprises a gas sensor section 10 having a first sensor to an n-th sensor (n is a natural number of 3 or more) and a determination section 20, and determines a gas type included in a gas to be measured. The determination section 20 obtains a one-dimensional first parameter P1, which is calculated on the basis of first coordinates including detection data V1 of a first sensor 101 and detection data V2 of a second sensor 102 having sensitivity characteristics different from that of the first sensor as elements, and a one-dimensional second parameter P2, which is calculated on the basis of second coordinates including detection data V3 of a third sensor 103 and detection data Vi of an i-th sensor 10i (i is a natural number of n or less and other than 3) having sensitivity characteristics different from that of the third sensor 103 as elements, and determines the gas type on the basis of third coordinates including the first parameter P1 and the second parameter P2 as elements.SELECTED DRAWING: Figure 8

Description

[Technical Field]

[0001] The present invention relates to a gas detection device for determining the type of gas contained in a gas to be measured. [Background technology]

[0002] There is a widespread demand for identifying the types of gases present in the ambient atmosphere for purposes such as fire prevention, evaluation of the properties of food and fragrances, gas leak detection, and measurement of indoor population density. Examples of methods for identifying gas types include pattern recognition techniques, specifically methods using neural networks and multivariate analysis. For example, Patent Document 1 describes an identification method using the back propagation method (BP method). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2011-169830 Summary of the Invention [Problem to be solved by the invention]

[0004] The BP method employed in Patent Document 1 is a supervised learning algorithm for a hierarchical neural network. Such a machine learning-based classification method requires the preparation of a large amount of high-quality training data to improve classification accuracy, and when there is a large amount of input information, the calculation load tends to be high. In other words, when the classification process includes machine learning, the load of preprocessing for the classification process is large, and the load of the classification process itself is also high. Although there are various methods with low processing loads, they have the problem of low classification accuracy.

[0005] An object of the present invention is to provide a gas detection device that has a small processing load and can improve identification accuracy. [Means for solving the problem]

[0006] A gas detection device according to one aspect of the present invention for solving the above problem comprises a gas sensor unit having a first sensor to an n-th sensor (n is a natural number greater than or equal to 3), and a determination unit, and is configured to determine the type of gas contained in a measured gas. The determination unit determines a one-dimensional first parameter P1 calculated based on a first coordinate system having, as elements, detection data V1 from the first sensor and detection data V2 from a second sensor having sensitivity characteristics different from those of the first sensor, and a one-dimensional second parameter P2 calculated based on a second coordinate system having, as elements, detection data V3 from a third sensor and detection data Vi from an i-th sensor (i is a natural number less than or equal to n and other than 3) having sensitivity characteristics different from those of the third sensor, and determines the type of gas based on a third coordinate system having, as elements, the first parameter P1 and the second parameter P2.

[0007] In the gas detection device of this embodiment, by appropriately selecting the types of multiple sensors possessed by the gas sensor unit and making the third coordinates obtained by detecting multiple gas species that may be contained in the measured gas different from each other in a manner that allows them to be distinguished from each other, it is possible to determine the gas species without increasing the load of the identification process performed by the determination unit.

[0008] In the gas detection device according to the present embodiment, the first parameter P1 may be a coordinate transformation value of the first coordinate or a value derived from the coordinate transformation value, and the second parameter P2 may be a coordinate transformation value of the second coordinate or a value derived from the coordinate transformation value. Dimensionality reduction using coordinate transformation is a preferred example, as it reduces the computational load. A typical example of coordinate transformation is transformation between a Cartesian coordinate system and a polar coordinate system, but this is not limiting. When the coordinate transformation value is an angle coordinate obtained by transformation from a Cartesian coordinate system to a polar coordinate system, a tangent value, which is a calculated value including an angle, is a specific example of a value derived from the coordinate transformation value.

[0009] In the gas detection device according to the present embodiment, specific examples of the first parameter P1 include angular coordinates obtained by polar transformation of the first coordinates in a linear Cartesian coordinate system, or values ​​derived from the angular coordinates. Specific examples of the second parameter P2 include angular coordinates obtained by polar transformation of the second coordinates in a linear Cartesian coordinate system, or values ​​derived from the angular coordinates. Because there is a one-to-one correspondence between Cartesian and polar coordinate systems, it is possible to maintain the characteristics in the Cartesian coordinate system even after the coordinate transformation. Furthermore, when a group of first coordinates based on data obtained by detecting the same gas species are transformed from the Cartesian coordinate system to polar coordinates, the variation in the angular coordinates is reduced. This tendency is also true for second coordinates, and therefore, it is preferable to appropriately extract the characteristics of each gas species by dimensional reduction, which determines third coordinates from the first and second coordinates.

[0010] In the gas detection device according to the above embodiment, the determination unit may have a determination coordinate system divided into a plurality of regions. In this case, the determination unit determines the gas type based on which of the plurality of regions the third coordinate in the determination coordinate system belongs to. If the third coordinate in the determination coordinate system is set to be appropriately different for each gas type, the gas type can be determined by a simple method such as region determination, thereby reducing the load of the identification process. [Effects of the Invention]

[0011] According to the present invention, a gas detection device is provided that has a small processing load and can improve identification accuracy. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a block diagram showing a configuration of a gas detection device according to an embodiment of the present invention. [Figure 2] FIG. 4 is a flowchart illustrating an identification process of the gas detection device according to the embodiment of the present invention. [Figure 3] 4 is a graph showing an example of data acquired from a first sensor of the gas detection device according to one embodiment of the present invention. [Figure 4] 4 is a graph obtained by correcting the baseline of the graph shown in FIG. 3. [Figure 5] 4 is a diagram in which first coordinates determined in an identification process by a determination unit included in the gas detection device according to one embodiment of the present invention are plotted on a two-dimensional linear orthogonal coordinate system. FIG. [Figure 6] 5A and 5B are diagrams illustrating a method in which a determination unit included in the gas detection device according to one embodiment of the present invention calculates a first parameter from a first coordinate. [Figure 7] 5A and 5B are diagrams illustrating a method in which a determination unit included in the gas detection device according to one embodiment of the present invention calculates a second parameter from a second coordinate. [Figure 8] 10 is a diagram illustrating a method for classifying odors from third coordinates determined by a determination unit included in a gas detection device according to one embodiment of the present invention. FIG. [Figure 9] FIG. 10 is a flowchart illustrating another example of the identification process of the gas detection device according to the embodiment of the present invention. [Figure 10] 10 is a graph showing baseline correction performed on data acquired from a third sensor of the gas detection device according to one embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the following description, the same components will be denoted by the same reference numerals, and the description of components that have already been described will be omitted as appropriate.

[0014] FIG. 1 is a block diagram showing the configuration of a gas detection device according to one embodiment of the present invention.

[0015] 1, a gas detection device 100 according to one embodiment of the present invention includes a gas sensor unit 10 having a first sensor 101, a second sensor 102, a third sensor 103, and an n-th sensor 10n (n is a natural number equal to or greater than 3), and a determination unit 20 that performs an identification process to identify the type of gas based on a signal input from the gas sensor unit 10. When n=3, the gas sensor unit 10 does not include the sensor corresponding to symbol 10n.

[0016] Each sensor included in the gas sensor unit 10 is a sensor that includes a detection element including, for example, a semiconductor, and outputs an electrical signal, for example, a voltage, that changes when the detection element detects a gas. As will be described later, the first sensor 101 and the second sensor 102 in the gas sensor unit 10 have different sensitivity characteristics, and the gas sensor unit 10 also includes at least one sensor whose sensitivity characteristics are different from that of the third sensor 103.

[0017] The determination unit 20 performs an identification process to identify the type of gas based on the signal input from the gas sensor unit 10. The determination unit 20 includes a receiving unit for receiving the signal from the gas sensor unit 10, a calculation unit for executing multiple processes that make up the identification process, a storage unit for temporarily storing the signal received by the receiving unit and for storing various data required for the execution of the calculation unit, and a transmitting unit for transmitting the result of the identification process to the outside.

[0018] FIG. 2 is a flowchart illustrating the identification process of the gas detection device according to one embodiment of the present invention.

[0019] As shown in FIG. 2, the determination unit 20 first acquires data from the gas sensor unit 10 (S101). The data from the gas sensor unit 10 may be composed of a signal output by the gas sensor unit 10, or may be data derived from the output signal. A specific example of such data is numerical data obtained by converting an analog electrical signal output by the gas sensor unit 10, where the signal is an analog electrical signal with a fluctuating voltage, into a digital signal and then converting it from a voltage such as a resistance. Such signal conversion and data conversion may be performed by the gas sensor unit 10 or by the determination unit 20. In other words, step S101 may include signal processing.

[0020] 3 is a graph showing an example of data acquired from the first sensor of a gas detection device according to one embodiment of the present invention. In this example, the data is digital data showing the signal strength of the first sensor over a measurement period of 10,000 seconds or more. This data shows that gases were detected during a period from about 3,000 seconds to about 4,000 seconds (first period), a period from about 7,000 seconds to about 8,000 seconds (second period), and a period from about 9,000 seconds to about 11,000 seconds (third period). Hereinafter, the gas detected during the first period will be referred to as first gas G1, the gas detected during the second period as second gas G2, and the gas detected during the third period as third gas G3.

[0021] In the identification process according to this embodiment, baseline correction is performed on the acquired data. (Step S102). FIG. 4 is a graph obtained by performing baseline correction on the graph shown in FIG. 3. The baseline correction is performed so that the average value of the signal intensity becomes zero for each of the periods from the start of measurement to the start of the first period, the period from the end of the first period to the start of the second period, and the period from the end of the second period to the start of the third period. Although such baseline correction is not essential, performing baseline correction can improve the ease of identifying each gas when the first coordinate is obtained. In this embodiment, the data after baseline correction is used as the detection data of each sensor. Specifically, the data shown in FIG. 4 is the detection data V1 of the first sensor 101.

[0022] The sensed data V2 is obtained in the same manner for the second sensor 102, which has sensitivity characteristics different from those of the first sensor 101, and a first coordinate is obtained that includes these sensed data V1 and V2 as elements (step S103). In this embodiment, the first coordinate is a two-dimensional coordinate (sensed data V2 of the second sensor 102, sensed data V1 of the first sensor 101), but is not limited to this. The first coordinate may be a three- or higher-dimensional coordinate (in which case, the three- or higher-dimensional elements are set to "0"), or may be a coordinate in which the elements of the two-dimensional coordinate are swapped (sensed data V1 of the first sensor 101, sensed data V2 of the second sensor 102).

[0023] FIG. 5 is a diagram in which the first coordinates determined in the discrimination process by the determination unit included in the gas detection device according to one embodiment of the present invention are plotted on a two-dimensional linear Cartesian coordinate system. Hereinafter, this plot will also be referred to as the "first plot." In the first plot of FIG. 5, the horizontal axis represents the detection data V2 of the second sensor 102, and the horizontal axis represents the detection data V1 of the first sensor 101. As described above, the first sensor 101 and the second sensor 102 have different sensitivity characteristics. Therefore, as shown in FIG. 5, in the first plot, the data for each gas (first gas G1, second gas G2, third gas G3) are radially separated and arranged around the origin. The greater the difference between the sensitivity characteristics of the first sensor 101 and the second sensor 102, the more easily the gases are distributed in the first plot.

[0024] As described above, since the first coordinates are arranged radially in the first plot, it is difficult to identify which gas the data arranged near the origin belongs to. Therefore, in this embodiment, a process is performed to eliminate data close to the origin in the first plot. Specifically, first, for each coordinate of the first coordinates, the distance D1 from the origin is calculated (step S104). Next, coordinates where the distance D1 is equal to or greater than a predetermined threshold D1s are extracted (step S105). In other words, coordinates where the distance D1 is less than the predetermined threshold D1s are excluded from subsequent processing. In FIG. 5, the range where the distance D1 is equal to the threshold D1s is indicated by a dashed line. Coordinates located in the area surrounded by this dashed line are excluded from subsequent processing.

[0025] 6 is a diagram illustrating a method in which a determination unit included in a gas detection device according to one embodiment of the present invention calculates a first parameter from a first coordinate. In the first graph shown in FIG. 6, coordinates where the distance D1 is less than a predetermined threshold D1s are deleted. This filtering process separates the first coordinates belonging to each gas from the other gases. Note that in FIG. 6, a portion of the third gas G3 is separated into the negative region on the vertical axis, and such data may be treated as anomalous data and excluded from subsequent processing.

[0026] Since the gases have been separated in the first plot in this way, the coordinate system is then converted (coordinate conversion) (step S106). Specifically, the linear Cartesian coordinate system of the first plot is converted to a polar coordinate system. As a result, the first coordinate is converted from coordinates (distance along the vertical axis, distance along the horizontal axis) to (distance (starting line) from the origin (pole), angle (deflection angle) from the reference axis BA1 (horizontal axis in FIG. 6)). Then, the angular coordinate consisting of this deflection angle θ1 is obtained as the one-dimensional first parameter P1.

[0027] In Figure 6, the approximate deflection angle θ11 for the first gas G1, the approximate deflection angle θ12 for the second gas G2, and the approximate deflection angle θ13 for the third gas G3 are shown by dashed lines. Specifically, the deflection angle θ11 was 38°, the deflection angle θ12 was 81°, and the deflection angle θ13 was 35°.

[0028] After the one-dimensional first parameter P1 is calculated as described above, second coordinates are calculated (step S107) including, as elements, the sensed data V3 of the third sensor 103 and the sensed data V4 of the fourth sensor 104, which has sensitivity characteristics different from those of the third sensor 103. Next, as in steps S105 and S106 for the first coordinates, the second coordinates are plotted on a linear Cartesian coordinate system (hereinafter, this plot is also referred to as the "second plot"), and a distance D2 from the origin is calculated for each second coordinate in the second plot (step S108), and coordinates where the distance D2 is equal to or greater than a threshold value D2s are extracted (step S109).

[0029] FIG. 7 shows a plot (second plot) of the filtered second coordinates. As shown in FIG. 7, the second coordinates of the second gas G2 and the third gas G3 are plotted in a similar region (almost on the horizontal axis) in the second plot. In FIG. 7, different types of markers are used to distinguish between the two. This confirms that the second coordinates of the third gas G3 include values ​​of the detection data V4 of the fourth sensor 104 that are larger than the second coordinates of the second gas G2.

[0030] Next, a coordinate system conversion (coordinate conversion) is performed on the second coordinates (step S110). Specifically, the linear Cartesian coordinate system of the second plot is converted to a polar coordinate system. As a result, the second coordinates are converted from coordinates (distance along the vertical axis, distance along the horizontal axis) to (distance (starting line) from the origin (pole), angle (deflection angle) from the reference axis BA2 (horizontal axis in FIG. 7)). Then, the angular coordinate consisting of this deflection angle θ2 is obtained as the one-dimensional second parameter P2. In FIG. 7, the deflection angle θ21 of the first gas G1 is approximately 60°, and the deflection angle θ22 of the second gas G2 and the deflection angle θ23 of the third gas G3 are all approximately 0°.

[0031] Since the first parameter P1 and the second parameter P2 are obtained by executing steps S106 and S110, a third coordinate is set using these parameters as elements (step S111). FIG. 8 is a diagram illustrating a method for classifying odors from the third coordinate determined by a determination unit included in a gas detection device according to one embodiment of the present invention. As shown in FIG. 8, the determination unit 20 has a linear Cartesian coordinate system as a determination coordinate system, with the horizontal axis representing the second parameter P2 (deflection angle θ2) and the vertical axis representing the first parameter P1 (deflection angle θ1), and the third coordinate is plotted on this coordinate system. Hereinafter, this plot will also be referred to as the "third plot."

[0032] By plotting in this manner, the first gas G1 and the third gas G3, which are located in the same region in the first plot, are properly separated, and the second gas G2 and the third gas G3, which were difficult to distinguish in the second plot, are properly separated, making it easy to visually distinguish the three gases. Specifically, when the third plot is divided into the following first region R1 to fourth region R4, the third coordinate of the first gas G1 belongs to the third region R3, the third coordinate of the second gas G2 belongs to the second region R2, and the third coordinate of the third gas G3 belongs to the first region R1. First region R1: The first parameter P1 (deflection angle θ1) is less than 50° and the second parameter P2 (deflection angle θ2) is less than 35° Second region R2: The first parameter P1 (deflection angle θ1) is 50° or more and the second parameter P2 (deflection angle θ2) is less than 35° Third region R3: The first parameter P1 (deflection angle θ1) is less than 50° and the second parameter P2 (deflection angle θ2) is 35° or more. Fourth region R4: The first parameter P1 (deflection angle θ1) is 50° or more and the second parameter P2 (deflection angle θ2) is 35° or more.

[0033] In Figure 8, the line L1 where P1 = 50° and the line L2 where P2 = 35° are both shown as dashed lines. Furthermore, for confirmation, dashed lines indicating the deflection angle θ11 (P1 = 38°), the deflection angle θ12 (P1 = 81°), the deflection angle θ13 (P1 = 35°), and the deflection angle θ21 (P2 = 60°) are shown up to positions where they can be compared with the third coordinate of the corresponding gas.

[0034] Once the third plot is obtained in this manner, the determination unit 20 performs odor classification, which determines the gas type based on which of the multiple regions (first region R1 to fourth region R4) the third coordinate in the third plot belongs to (step S112). Specifically, the determination unit 20 determines that the third coordinate belonging to the first region R1 originates from the first gas G1, that the third coordinate belonging to the second region R2 originates from the second gas G2, and that the third coordinate belonging to the third region R3 originates from the first gas G1. Note that the odor classification in step S112 may be performed using machine learning (e.g., a determination including image recognition of the third plot).

[0035] The determination unit 20 performs the above-described identification process, thereby realizing identification of the gas type.

[0036] The above-described embodiments have been described to facilitate understanding of the present invention, and are not intended to limit the present invention. Therefore, each element disclosed in the above embodiments is intended to include all design modifications and equivalents that fall within the technical scope of the present invention.

[0037] For example, in the above description, the second coordinates are set using the detection data V3 of the third sensor 103 and the detection data V4 of the fourth sensor 104, which has sensitivity characteristics different from those of the third sensor 103, as elements. However, this is not limiting. Instead of the fourth sensor 104, an ith sensor 10i (where i is a natural number equal to or less than n and other than 3) having sensitivity characteristics different from those of the third sensor 103 may be used, and the detection data Vi of this ith sensor 10i and the detection data V3 of the third sensor 103 may be used as elements of the second coordinates. In this case, the ith sensor 10i may be the first sensor 101 or the second sensor 102. Setting the ith sensor 10i in this manner reduces the number of sensors required for the gas sensor unit 10 to three, thereby simplifying the configuration of the gas detection device 100.

[0038] In the above description, steps S102 to S111 are executed once using a sensor set consisting of first sensor 101 to fourth sensor 104 in the identification process, but this is not limited to this. Multiple sensor sets having different constituent sensors may be set in advance, and steps S102 to S111 may be executed for each sensor set to prepare multiple third coordinates. An example of such a process is shown in FIG. 9. FIG. 9 is a flow chart for explaining another example of the identification process of the gas detection device according to one embodiment of the present invention.

[0039] In the flow shown in FIG. 9, a total number k of sensor sets are set in advance, and each of these sensor sets is associated with a different sensor set number S1 (a natural number equal to or less than the total number k of sensor sets). When the identification process is started, first, the sensor set number S1 is set to 1 (step S201), and data is acquired from the sensors constituting the sensor set corresponding to sensor set number 1 (step S202). Then, steps S102 to S111 shown in FIG. 2 are executed (step S203). Specifically, a first coordinate and a second coordinate are set, and a first parameter P1 and a second parameter P2 are calculated from these coordinates, and a third coordinate having these parameters as elements is obtained. Next, it is determined whether the sensor set number S1 is equal to the total number k of sensor sets (step S204). If the sensor set number S1 is not equal to the total number k, the sensor set number S1 is incremented by 1 (step S205), and data is acquired from the sensors constituting the sensor set corresponding to sensor set number 2 (step S202), and step S203 is executed. On the other hand, if it is determined in step S204 that the sensor set number S1 is equal to the total number k of sensor sets, odor classification is performed using the multiple third coordinates obtained by the processes up to this point (step S206).

[0040] The combination of sensors selected to set the first and second coordinates preferably has a large difference in sensitivity characteristics, but is not limited to this. It is sufficient if the gases can be appropriately separated in the third plot.

[0041] Furthermore, it is not required that all sensors be able to detect all of the measurement gases. For example, the third sensor 103, which provides the second plot shown in Fig. 7, can clearly detect the first gas G1, but cannot detect the second gas G2 and the third gas G3, as shown in Fig. 10. Even with such sensors, if the first sensor 101 and the second sensor 102 can separately detect the second gas G2 and the third gas G3, the respective gases can be identified with high discriminability by using the third plot showing the third coordinates in step S112.

[0042] Further explaining the characteristics of each sensor, in the first plot shown in Fig. 6, the first coordinate representing the first gas G1 and the first coordinate representing the third gas G3 are located relatively close to each other. However, in the second plot shown in Fig. 7, the second coordinate representing the first gas G1 is located in a different region from the second coordinates of the other gases, and therefore, in the third plot, the third coordinates of the first gas G1 to the third gas G3 are displayed with high distinguishability from each other.

[0043] In the above description, the first parameter P1 is a coordinate transformation value of the first coordinate, but is not limited to this and may be a value derived from the coordinate transformation value. For example, as described above, if the first parameter P1 is an angle coordinate when the first coordinate plotted in a linear rectangular coordinate system is transformed into polar coordinates, the first parameter P1 may be a trigonometric function (e.g., tangent) with the angle coordinate (argument angle θ1) as a variable. Similarly, the second parameter P2 is a coordinate transformation value of the second coordinate, but is not limited to this and may be a value derived from the coordinate transformation value. [Explanation of symbols]

[0044] 100: Gas detection device 10: Gas sensor section 20: Judgment section 101: First sensor 102: Second sensor 103: Third sensor 104: 4th sensor 10i: ith sensor 10n: nth sensor BA1, BA2: Reference axes D1, D2: distance D1s, D2s: threshold G1: First gas G2: Second gas G3: Third gas L1, L2: line R1: 1st area R2: 2nd area R3: 3rd area R4: 4th area S1: Sensor set number θ11, θ12, θ13, θ21, θ22, θ23: Declination angle

Claims

1. A gas detection device comprising a gas sensor unit having first to nth sensors (n is a natural number greater than 3) and a determination unit, the gas detection device determining a type of gas contained in a measurement gas, The determination unit a one-dimensional first parameter P1 calculated based on a first coordinate system including, as elements, the detection data V1 of the first sensor and the detection data V2 of a second sensor having sensitivity characteristics different from that of the first sensor; a one-dimensional second parameter P2 is calculated based on second coordinates independent of the first coordinates, and includes, as elements, the detection data V3 of a third sensor and the detection data Vi of an i-th sensor (i is a natural number equal to or less than n and other than 3) having sensitivity characteristics different from that of the third sensor; determining a gas type based on a third coordinate that includes the first parameter P1 and the second parameter P2 as elements and is independent of both the first coordinate and the second coordinate; the first parameter P1 is a first angular coordinate obtained by converting the first coordinate in a linear rectangular coordinate system into a polar coordinate, or a value derived from the first angular coordinate, the second parameter P2 is a second angular coordinate obtained by converting the second coordinate in a linear rectangular coordinate system into a polar coordinate, or a value derived from the second angular coordinate, the determination unit has a determination coordinate system divided into a plurality of regions, with the first angle coordinate or a value derived from the first angle coordinate as one element and the second angle coordinate or a value derived from the second angle coordinate as the other element, determining the type of gas based on which of the plurality of regions the third coordinate in the determination coordinate system belongs to; A gas detection device characterized by:

2. A gas detection device as described in Claim 1, wherein the first parameter P1 is calculated as a coordinate among the first coordinates whose distance from the origin in a linear Cartesian coordinate system is equal to or greater than a predetermined threshold value.

3. A gas detection device as described in claim 1 or claim 2, wherein the second parameter P2 is calculated as a coordinate among the second coordinates whose distance from the origin in a linear Cartesian coordinate system is greater than or equal to a predetermined threshold value.

Images