Component in gas detection device, component in gas detection method, and component in gas detection program

Abstract

To provide a component in gas detection device, component in gas detection method, component in gas detection program that boost estimation accuracy of density of a component in gas.SOLUTION: A model generation unit 12 is configured to generate a density estimation model to be used in estimating density of a prescribed component by a multivariate analysis on the basis of a measurement result of the density of each component obtained by a plurality of gas sensors 21 to 24 different in a detection principle, and included in testing gas having the density of the prescribed component adjusted. A density calculation unit 13 is configured to calculate the density of the prescribed component using the density estimation model with respect to the measurement result by the gas sensors 21 to 24 of the density of each component included in gas targeted for the density estimation of the prescribed component.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a gas component detection device, a gas component detection method, and a gas component detection program. [Background technology]

[0002] As a power transmission cable system for transmitting electricity to a consumer area, there are long-distance lines connecting substations and short-distance lines such as substation premises interconnection lines. In such a power transmission cable system, power cables that carry electricity, intermediate connections that connect the power cables to each other, and terminal connections that connect the power cables to other facilities are used.

[0003] Termination connections are used for 20 to 30 years, but partial discharges may occur due to aging, etc. Since partial discharges have a high probability of ultimately leading to accidents such as fires, there is a strong demand for accurate detection and response to partial discharges.

[0004] Here, the occurrence of partial discharge accelerates the deterioration of the insulators and silicone oil, and generates decomposition gases. Since the decomposition gases are diffused into the protective metal fittings at the bottom of the terminal connection, it may be possible to determine whether or not partial discharge has occurred by analyzing the gas inside the protective metal fittings. It is known that discharge deterioration of silicone oil generates components such as H2, CO, CH4 (methane), C2H6 (ethane), C2H4 (ethylene), and C2H2 (acetylene). Of these generated components, acetylene is one that is unlikely to be generated by causes other than discharge.

[0005] Therefore, a method of detecting the occurrence of partial discharge by estimating the acetylene concentration in the protective metal fitting of the terminal connection part using a gas sensor capable of detecting various components is considered. There are various types of gas sensors that detect components, such as electrochemical sensors, photoionization sensors, and infrared sensors. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2013-53949 A [Non-patent literature]

[0007] [Non-Patent Document 1] Takahiro Ohno et al., "Acetylene Reduction Behavior in Oil-Immersed Transformers," Journal of Electrical Engineering B, Vol. 119, No. 1, pp54-62, 1999 Summary of the Invention [Problem to be solved by the invention]

[0008] However, in any type of gas sensor, components other than the target component are detected together with the target component. The gas in the protective metal fitting of the termination connection part, which is the target for detecting the occurrence of partial discharge, contains many components with similar chemical properties to acetylene gas, and a general gas sensor will detect similar components such as ethylene as well. Furthermore, the concentration range of each component in the gas is wide, and the acetylene concentration to be estimated is low, at sub-ppm to several ppm. Therefore, similar components become interference gases, and it is difficult for a general gas sensor to properly measure the trace concentration of acetylene. For example, there is a possibility that a detection of another component that is an interference gas is mistaken for a detection of acetylene. In addition, when another component and acetylene are detected together, it is difficult to determine whether the detected value can be regarded as the concentration of acetylene. Therefore, with the conventional technology for detecting components in gas using a single gas sensor, the detection accuracy of the acetylene concentration in the gas in the protective metal fitting of the termination connection part is low, and it is difficult to accurately detect the occurrence of partial discharge in the termination connection part, making it difficult to improve the safety of the power transmission cable system.

[0009] The disclosed technology has been made in consideration of the above, and aims to provide a gas component detection device, a gas component detection method, and a gas component detection program that improve the accuracy of estimating the concentration of gas components. [Means for solving the problem]

[0010] In one aspect of the gas component detection device, gas component detection method, and gas component detection program disclosed herein, a model generation unit generates a concentration estimation model used for estimating the concentration of a predetermined component by multivariate analysis based on measurement results of the concentration of each component contained in a test gas in which the concentration of the predetermined component has been adjusted, obtained by a plurality of gas sensors having different detection principles. A concentration calculation unit calculates the concentration of the predetermined component contained in the target gas using the concentration estimation model generated by the model generation unit for the measurement results of the concentration of each component contained in a target gas in which the concentration of the predetermined component is to be estimated, obtained by the plurality of gas sensors. Effect of the Invention

[0011] According to one aspect, the present invention can improve the accuracy of estimating the concentration of a component in a gas. [Brief description of the drawings]

[0012] [Figure 1] FIG. 1 is a diagram showing an example of a power transmission cable system. [Diagram 2] FIG. 2 is a diagram showing an example of the structure of the termination connection portion. [Diagram 3] FIG. 3 is a block diagram of a gas component detection system according to an embodiment. [Figure 4] FIG. 4 is a diagram showing an example of a score plot with the principal components as axes. [Diagram 5] FIG. 5 is a diagram in which a scale for detecting acetylene concentration is added to the concentration estimation model. [Figure 6] FIG. 6 is a diagram showing the tendency of the output value of each gas sensor. [Figure 7] FIG. 7 is a flowchart of a partial discharge occurrence determination process using an estimation of the acetylene concentration by the gas component detection system according to the embodiment. [Figure 8] FIG. 8 is a diagram showing the hardware configuration of the gas component detection device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Hereinafter, examples of the gas component detection device, gas component detection method, and gas component detection program disclosed in the present application will be described in detail with reference to the drawings. Note that the gas component detection device, gas component detection method, and gas component detection program disclosed in the present application are not limited to the following examples. EXAMPLES

[0014] Fig. 1 is a diagram showing an example of a power transmission cable system. The power transmission cable system in Fig. 1 is an underground power transmission line in an urban area. That is, the power transmission cable system shown in Fig. 1 transmits power between substations.

[0015] For example, the power transmission cable system shown in Fig. 1 transmits power underground using power cables 201 to 204. In the power transmission cable system, power cables 202 and 203 are connected to each other by intermediate junction box 103, and power cables 203 and 204 are connected to each other by intermediate junction box 104. Furthermore, power cables 201, 202 and 204 are connected to a transformer, gas-insulated switchgear, and the like by terminal junction boxes 101, 102 and 105.

[0016] The terminal connection boxes 101, 102, and 105 each have a terminal connection part 110 as shown in Fig. 2 for connecting the power cables 201, 202, and 204, respectively. Fig. 2 is a diagram showing an example of the structure of a terminal connection part. In Fig. 2, a partial cross section is shown to show the internal structure of the terminal connection part 110. The terminal connection part 110 has a protective metal fitting 111 as well as various fixing devices and insulators. When partial discharge occurs in the terminal connection part 110, decomposition product gas is generated and diffuses and accumulates in the protective metal fitting 111.

[0017] Fig. 3 is a block diagram of a gas component detection system according to an embodiment. The gas component detection system 1 estimates the acetylene concentration in the gas retained in the protective metal fitting 111 shown in Fig. 2, and detects the occurrence of partial discharge using the estimation result. The gas component detection system 1 has gas sensors 21-24 and a gas component detection device 10.

[0018] The gas sensors 21 to 24 include sensors that use different component detection methods. For example, in this embodiment, the gas sensor 21 is an electrochemical sensor that detects sulfur compounds as well as acetylene. The gas sensors 22 and 23 are electrochemical sensors that detect different gases than the gas sensor 21. The gas sensor 24 is a photoionization sensor. For example, the gas sensor 24 detects organic compounds in general.

[0019] When creating a concentration estimation model, the gas sensors 21 to 24 measure the concentrations of components contained in a test gas, which is a mixed gas created by adjusting the concentration. In this embodiment, the test gases are C2H2 gas (0.1% in N2), C2H4 gas (99.9%), CO gas (99.9%), H2 gas (99.9%), CH4, C2H6, C3H8, n-C4H 10 , iso-C4H 10 , CO2 (1% each in N2) mixed gas, C2H4, C2H6, C3H6, C3H8, n-C4H 10 , iso-C4H 10 Six types of mixed gases (1% in N2 each) were appropriately mixed and used.

[0020] In addition, when estimating the acetylene concentration in the gas retained in the protective metal fitting 111 of the actual terminal connection portion 110, the gas sensors 21 to 24 measure the concentrations of components contained in the gas obtained from a gas pipe connected to a connector provided on the protective metal fitting 111.

[0021] The gas component detection device 10 generates a concentration estimation model for estimating the acetylene concentration using the detection results of each of a plurality of test gases by the gas sensors 21-24. Then, the gas component detection device 10 acquires the detection results of the gas inside the protective metal fitting 111 by the gas sensors 21-24, and estimates the acetylene concentration in the gas inside the protective metal fitting 111 using the generated concentration estimation model. Thereafter, the gas component detection device 10 detects the occurrence of partial discharge in the termination connection part 110 based on the estimation results of the acetylene concentration.

[0022] The following describes in detail the gas component detection device 10. The gas component detection device 10 has a data acquisition unit 11, a model generation unit 12, a concentration calculation unit 13, a partial discharge occurrence detection unit 14, and a notification unit 15, as shown in FIG.

[0023] The data acquisition unit 11 acquires the detection results of components in the gas obtained by each of the gas sensors 21 to 24. In the case of generating a concentration estimation model, the data acquisition unit 11 outputs the acquired detection results to the model generation unit 12. In the case of estimating the acetylene concentration in the gas retained in the protective metal fitting 111, the data acquisition unit 11 outputs the acquired detection results to the concentration calculation unit 13.

[0024] The model generation unit 12 receives the detection results of each test gas by each of the gas sensors 21-24 from the data acquisition unit 11. Then, the model generation unit 12 uses a multivariate analysis method, which is one of the linear regression analysis methods, on the acquired detection results of each test gas to identify main components that are highly correlated with the acetylene concentration value.

[0025] For example, the model generating unit 12 executes partial least squares regression for a plurality of correlated explanatory variables using the following formula (1) as a multivariate analysis method. Here, in this embodiment, the explanatory variables are output values ​​corresponding to the detection results of the gas sensors 21 to 24. The detection results of the gas sensors 21 to 24 are correlated with the gas concentrations.

[0026]

number

[0027] Here, A is the number of components in partial least squares regression, which corresponds to the number of principal components. ta is the a-th principal component. pa is the a-th loading, i.e., the weighting of the output value of each sensor. E is the residual of X. qa is the a-th coefficient. f is the residual of y.

[0028] Then, the model generating unit 12 sets the principal components such that the variance between the principal components and the acetylene concentration value, which is the objective variable, is maximized. Here, the principal components are variables expressed by a linear equation of the explanatory variables. By classifying the detection results using such principal components, the difference in acetylene concentration can be maximized.

[0029] Next, the model generation unit 12 calculates each detection result using a linear equation of the set principal components, and plots the calculation results in a coordinate space with the set principal components as axes to create a score plot with the set principal components as axes. In this way, the model generation unit 12 generates a concentration estimation model for estimating the acetylene concentration, which is the objective variable.

[0030] 4 is a diagram showing an example of a score plot with the principal components as axes. For example, the model generation unit 12 sets one principal component as Factor-1=0.01×(output value of gas sensor 24)+0.44×(output value of gas sensor 22)+0.87×(output value of gas sensor 21)+0.48×(output value of gas sensor 24). At this time, the output value of each gas sensor uses the average centered value. For Factor-2, a principal component is set again that maximizes the covariance for the residues E and f when Factor-1 is calculated.

[0031] Then, the model generation unit 12 sets a coordinate space shown in Fig. 4 with each of factor-1 and factor-2 as a coordinate axis, plots the detection results of each test gas on the set coordinate space, and generates the concentration estimation model shown in Fig. 4. In Fig. 4, the horizontal axis indicates the value of factor-1, and the vertical axis indicates the value of factor-2.

[0032] Set 301 in FIG. 4 is a set of points with an acetylene concentration of 0 to 0.5 ppm. Set 302 is a set of points with an acetylene concentration of 0.7 to 1.1 ppm. Set 303 is a set of points with an acetylene concentration of 1.4 ppm. Set 304 is a set of points with an acetylene concentration of 2.2 ppm. Set 305 is a set of points with an acetylene concentration of 3.6 ppm. Set 306 is a set of points with an acetylene concentration of 4.3 ppm. Set 307 is a set of points with an acetylene concentration of 5.8 ppm. Set 308 is a set of points with an acetylene concentration of 7.2 ppm. Set 309 is a set of points with an acetylene concentration of 10.1 ppm.

[0033] In this way, in the plot results shown in Figure 4, the points for each acetylene concentration are distributed in order of increasing concentration from the bottom left to the top right on the page. In other words, the concentration estimation model shown in Figure 4 indicates that the acetylene concentration increases toward the top right on the page.

[0034] Fig. 5 is a diagram in which a scale for detecting acetylene concentration is added to the concentration estimation model. The concentration estimation model generated by the model generation unit 12 can express the change in acetylene concentration, for example, with a scale 310 as shown in Fig. 5. The model generation unit 12 can create a concentration estimation model that estimates the acetylene concentration according to, for example, the scale 310 in Fig. 5. In this embodiment, as shown in Fig. 5, the model generation unit 12 can generate a concentration estimation model having a resolution of about 0.5 ppm in the low acetylene concentration region of 0 to 3 ppm.

[0035] Here, loading, which means weighting of the output value of each gas sensor 21-24 with respect to the principal components, will be described. Fig. 6 is a diagram showing the tendency of the output value of each gas sensor. The horizontal axis of Fig. 6 represents the value of factor-1, and the vertical axis represents the value of factor-2. For example, the loading with respect to the output value of gas sensor 21 is indicated by arrow 311. Furthermore, the loading with respect to the output value of gas sensor 22 or the output value of gas sensor 23 is indicated by arrow 312, and the loading with respect to the output value of gas sensor 24 is indicated by arrow 313.

[0036] That is, it can be said that the output value of the gas sensor 21 strongly reflects the acetylene concentration. Also, the output values ​​of the gas sensors 22 and 23 reflect the acetylene concentration and contribute to separating components other than acetylene. Also, it can be said that the output value of the gas sensor 24 strongly contributes to separating components other than acetylene. In this way, the model generating unit 12 determines the loading of each of the gas sensors 21 to 24 according to the tendency of the output value of each of the gas sensors 21 to 24 with respect to the acetylene concentration.

[0037] Here, in this embodiment, the model generation unit 12 generates the concentration estimation model using partial least squares regression. By using partial least squares regression, dimensionality can be reduced, and by compressing to a number of variables less than the number of sensor values, the gas component detection device 10 becomes easier to determine the results. In addition, by using partial least squares regression, the variance of the output can be maximized, and the gas component detection device 10 becomes able to estimate the acetylene concentration in more detail. Furthermore, by using partial least squares regression, the gas component detection device 10 can become less susceptible to the effects of variance in the measurement values.

[0038] However, the model generating unit 12 can also create a concentration estimation model for the acetylene concentration using a multivariate analysis method other than partial least squares regression. For example, the model generating unit 12 can apply various regression methods such as multiple regression analysis, discriminant analysis, canonical discriminant analysis, logistic regression, extended quantification theory type 1, extended quantification theory type 2, factor analysis, principal component regression, or support vector machine (SVM). Even when using these regression methods, the model generating unit 12 can generate a model that estimates the acetylene concentration based on the output values ​​of each of the gas sensors 21 to 24, or a model that sets a predetermined reference value and determines whether the concentration is equal to or greater than the reference value or less.

[0039] Continuing the explanation by returning to Fig. 3, the concentration calculation unit 13 acquires the concentration estimation model generated by the model generation unit 12. Then, the concentration calculation unit 13 receives an input of the detection results of the gas in the protective metal fitting 111 by the gas sensors 21 to 24 from the data acquisition unit 11.

[0040] Then, the concentration calculation unit 13 estimates the acetylene concentration by calculating the acetylene concentration of the gas retained in the protective metal fitting 111 using a concentration estimation model for the acquired detection results. For example, the concentration calculation unit 13 calculates factor-1 and factor-2 from the acquired detection results, plots them in FIG. 5, and estimates the acetylene concentration by reading the corresponding values ​​on the scale 310. Thereafter, the concentration calculation unit 13 outputs the estimated value of the acetylene concentration to the partial discharge occurrence detection unit 14.

[0041] The partial discharge occurrence detection unit 14 has in advance a threshold value for the acetylene concentration in the gas inside the protective metal fitting 111 where partial discharge occurs in the termination connection unit 110. The partial discharge occurrence detection unit 14 receives an input of an estimated value of the acetylene concentration from the concentration calculation unit 13. Then, the partial discharge occurrence detection unit 14 judges whether or not the acquired estimated value of the acetylene concentration is equal to or greater than the threshold value. If the acquired estimated value of the acetylene concentration is equal to or greater than the threshold value, the partial discharge occurrence detection unit 14 notifies the notification unit 15 of the occurrence of partial discharge.

[0042] The notification unit 15 receives a notification of the occurrence of partial discharge from the partial discharge occurrence detection unit 14. Then, the notification unit 15 notifies a manager of the power transmission cable system, etc., that a partial discharge has occurred in the termination connection unit 110 by displaying a message of the occurrence of partial discharge on a monitor (not shown) or by turning on a warning light, etc.

[0043] Fig. 7 is a flowchart of partial discharge occurrence determination processing using acetylene concentration estimation by the gas component detection system according to the embodiment. Next, the flow of partial discharge occurrence determination processing by the gas component detection system 1 including the acetylene concentration estimation processing by the gas component detection device 10 according to the embodiment will be described with reference to Fig. 7.

[0044] The gas sensors 21 to 24 measure the concentration of a test gas, which is a mixed gas prepared by adjusting the acetylene concentration (step S1).

[0045] The data acquisition unit 11 acquires the detection results obtained by each of the gas sensors 21-24. The model generation unit 12 sets principal components that maximize the variance between the principal components and the acetylene concentration value, which is the objective variable, by using partial least squares regression, which is a multivariate analysis method, for the detection results of each test gas by each of the gas sensors 21-24 acquired by the data acquisition unit 11. Next, the model generation unit 12 calculates each detection result using a linear equation of the set principal components, and plots the calculation results in a coordinate space with the set principal components as axes, thereby creating a score plot with the set principal components as axes. As a result, the model generation unit 12 generates a concentration estimation model for estimating the acetylene concentration, which is the objective variable (step S2).

[0046] The gas sensors 21 to 24 measure the concentrations of components contained in the gas obtained from the gas pipe connected to the connector provided on the protective metal fitting 111 (step S3).

[0047] The concentration calculation unit 13 acquires the concentration estimation model generated by the model generation unit 12. Then, the concentration calculation unit 13 calculates the acetylene concentration of the gas in the protective metal fitting 111 by using the concentration estimation model for the detection results of the gas in the protective metal fitting 111 by the gas sensors 21-24 acquired from the data acquisition unit 11 (step S4).

[0048] The partial discharge occurrence detection unit 14 determines whether or not a partial discharge has occurred in the termination connection unit 110 depending on whether or not the estimated value of the acetylene concentration calculated by the concentration calculation unit 13 is equal to or greater than a threshold value (step S5). If a partial discharge has not occurred in the termination connection unit 110 (step S5: No), the partial discharge occurrence detection unit 14 ends the partial discharge occurrence determination process.

[0049] On the other hand, if a partial discharge has occurred in the termination connection unit 110 (step S5: Yes), the notification unit 15 receives a notification of the occurrence of a partial discharge from the partial discharge occurrence detection unit 14. Then, the notification unit 15 notifies the administrator of the power transmission cable system or the like that a partial discharge has occurred in the termination connection unit 110 (step S6).

[0050] As described above, the gas component detection device according to this embodiment generates a concentration estimation model by a multivariate analysis method using the detection results of a test gas in which the acetylene concentration has been adjusted, using multiple gas sensors with different detection methods or components to be detected. The gas component detection device then estimates the acetylene concentration in the protective metal fitting using the generated concentration estimation model for the detection results of the gas in the protective metal fitting of the terminal connection part by the multiple gas sensors, and determines the occurrence of partial discharge from the estimated acetylene concentration. Note that the gas component detection device and gas component detection method according to this embodiment are not limited to the terminal connection boxes 101, 102, and 105 illustrated in FIG. 1, but can also be applied to intermediate connection boxes 103 and 1014 and the like having a similar structure illustrated in FIG. 1.

[0051] This makes it possible to estimate the acetylene concentration in the range of 0-10 ppm with a resolution of about 0.5 ppm even in the presence of many interfering gases with similar components to acetylene. Therefore, it is possible to accurately estimate the low concentration of acetylene in the gas inside the protective metal fittings of the termination connection. Furthermore, by improving the accuracy of estimating the acetylene concentration, it is possible to accurately detect the occurrence of partial discharge in the termination connection, thereby improving the safety of the power transmission cable system.

[0052] In the above description, the concentration estimation model was generated using the measured values ​​from the electrochemical sensor and the photoionization sensor to estimate the acetylene concentration, but the type of gas sensor is not limited to this, and other information can be used as long as it is information that can estimate the acetylene concentration. For example, other gas sensors that can be used include infrared, semiconductor, new ceramic, thermal conduction, optical interference, non-dispersive infrared, differential absorption spectroscopy, diaphragm electrode, arc ultraviolet spectroscopy, and hydrogen flame ionization. Furthermore, humidity, response speed, measurement time difference, and the like can also be used as information used to estimate the acetylene concentration. Even when these measured values ​​are used, the concentration estimation model can be generated by the same procedure as in this embodiment.

[0053] (Hardware configuration) 8 is a hardware configuration diagram of the gas component detector 10. Next, an example of a hardware configuration for realizing each function of the gas component detector 10 will be described with reference to FIG.

[0054] 8, the gas component detection device 10 includes, for example, a CPU (Central Processing Unit) 91, a memory 92, a hard disk 93, and a network interface 94. The CPU 91 is connected to the memory 92, the hard disk 93, and the network interface 94 via a bus.

[0055] The network interface 94 is an interface for communication between the gas component detection device 10 and an external device. The network interface 94 relays communication between the gas sensors 21 to 24 and the CPU 91, for example.

[0056] 1. The hard disk 93 stores various programs including programs for realizing the functions of the data acquiring unit 11, the model generating unit 12, the concentration calculating unit 13, the partial discharge occurrence detecting unit 14, and the notifying unit 15. The hard disk 93 may also store a concentration estimation model generated by the model generating unit 12.

[0057] The memory 92 is a main storage device, and may be, for example, a dynamic random access memory (DRAM).

[0058] The CPU 91 reads out various programs from the hard disk 93, and deploys and executes them in the memory 92. In this way, the CPU 91 realizes the functions of the data acquisition unit 11, the model generation unit 12, the concentration calculation unit 13, the partial discharge occurrence detection unit 14, and the notification unit 15 illustrated in FIG. [Explanation of symbols]

[0059] 1. Gas component detection system 10. Gas component detector 11 Data Acquisition Section 12 Model Generation Unit 13 Concentration calculation section 14 Partial discharge detection section 15. Notification Department 21~24 Gas sensor 101, 102, 105 Termination box 103,104 Intermediate junction box 110 Termination 111 Protective bracket 201~204 Power cables

Claims

1. a model generating unit that generates a concentration estimation model for estimating the concentration of a predetermined component by multivariate analysis based on measurement results of concentrations of each component contained in a test gas in which the concentration of the predetermined component has been adjusted, the measurement results being obtained by a plurality of gas sensors having different detection principles; a concentration calculation unit that calculates the concentration of the predetermined component contained in the target gas using the concentration estimation model generated by the model generation unit based on the measurement results of the concentrations of the respective components contained in the target gas for estimating the concentration of the predetermined component, the measurement results being obtained by the plurality of gas sensors; A gas component detection device comprising:

2. 2. The gas component detection device according to claim 1, wherein the plurality of gas sensors use at least one of an electrochemical sensor and a photoionization sensor, or both of the electrochemical sensor and the photoionization sensor.

3. 2. The gas component detection device according to claim 1, wherein among the plurality of gas sensors, there are a plurality of gas sensors which have the same detection principle but detect different components.

4. 2. The gas component detection device according to claim 1, wherein the model generation unit generates the concentration estimation model using a partial least squares method.

5. 2. The gas component detection device according to claim 1, wherein the model generation unit and the concentration calculation unit use acetylene as the predetermined component.

6. the concentration calculation unit calculates a concentration of acetylene of the target gas accumulated in protective fittings of a terminal connection box and an intermediate connection box that connect a power cable; a partial discharge occurrence detection unit that compares the acetylene concentration calculated by the concentration calculation unit with a predetermined threshold value to detect the occurrence of partial discharge in the termination box; and a notification unit that notifies the occurrence of the partial discharge when the partial discharge occurrence detection unit detects the occurrence of the partial discharge.

6. The gas component detection device according to claim 5.

7. The gas component detection device is generating a concentration estimation model for estimating the concentration of the predetermined component by multivariate analysis based on measurement results of the concentration of each component contained in a test gas in which the concentration of the predetermined component has been adjusted, the measurement results being obtained by a plurality of gas sensors having different detection principles; Calculating the concentration of the predetermined component contained in the target gas using the concentration estimation model for the measurement results of the concentrations of the components contained in the target gas for which the concentration of the predetermined component is to be estimated, which are obtained by the plurality of gas sensors. A method for detecting components in a gas, comprising:

8. generating a concentration estimation model for estimating the concentration of the predetermined component by multivariate analysis based on measurement results of the concentration of each component contained in a test gas in which the concentration of the predetermined component has been adjusted, the measurement results being obtained by a plurality of gas sensors having different detection principles; Calculating the concentration of the predetermined component contained in the target gas using the concentration estimation model for the measurement results of the concentrations of the components contained in the target gas for which the concentration of the predetermined component is to be estimated, which are obtained by the plurality of gas sensors. A gas component detection program that causes a computer to execute a process.

Images