Gas concentration derivation device, gas concentration derivation method, computer program product, and sensor system

Abstract

The present application provides a kind of gas concentration derivation device, gas concentration derivation method, computer program product and sensor system.The gas concentration derivation device has: measurement information acquisition unit, obtains measurement information, the measurement information indicates the measurement value corresponding to the gas concentration in the measurement object environment measured by gas concentration measurement unit;Determination unit, based on the plurality of measurement information in the first period, determine whether the measurement information at the first time point after the first period is used as reference information when updating correction parameter;Gas concentration derivation unit, based on the measurement information and the correction parameter, derives the gas concentration in the measurement object environment;And update unit, update the correction parameter based on at least one reference information.

Description

Technical Field

[0001] This invention relates to a gas concentration extraction device, a gas concentration extraction method, a computer program product, and a sensor system. Background Technology

[0002] Patent document 1 describes a "method for estimating the true value of a measurement object by performing N measurements on a measurement object whose distribution of measurement values ​​is non-normal".

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: International Publication No. 2024 / 004281

[0006] Non-patent literature

[0007] Non-Patent Literature 1: Van der Hoven, I: Power spectrum of horizontal wind speed in the frequency range from 0.0007 to 900 cycles per hour, Journal of Meteorology, Vol. 14, No. 2, pp. 160-164, 1957. Summary of the Invention

[0008] In a first aspect of the present invention, a gas concentration deriving device is provided, comprising: a measurement information acquisition unit that acquires measurement information, the measurement information representing a measurement value corresponding to a gas concentration in a target environment measured by a gas concentration measurement unit; a determination unit that, based on multiple measurement information within a first period, determines whether to use the measurement information at a first time point after the first period as reference information when updating correction parameters; a gas concentration deriving unit that derives the gas concentration in the target environment based on the measurement information and the correction parameters; and an updating unit that updates the correction parameters based on at least one of the reference information.

[0009] The gas concentration extraction device may include: an environmental information acquisition unit that acquires environmental information, the environmental information representing one or more environmental parameters in the environment of the target environment measured by the environmental information measurement unit; and a correlation derivation unit that, based on multiple measurement information and the environmental information within the first period, derives the correlation between the measured value and the environmental information within the first period. The determination unit may, based on the environmental information at the first time point and the correlation within the first period, determine whether to use the measurement information at the first time point as the reference information.

[0010] In any of the gas concentration deriving devices, the correction parameters may include at least one of offset information, gain information, and curvature information.

[0011] In any of the gas concentration exporting devices, the determination unit may, based on the environmental information at a first time point after the first period and the correlation within the first period, export a predicted measurement value at the first time point. If the difference between the predicted measurement value and the measurement value shown by the measurement information at the first time point is below a threshold, the determination unit may use the measurement information at the first time point as the reference information.

[0012] In any of the gas concentration deriving devices, the threshold can be more than three standard deviations of the measured information.

[0013] In any of the gas concentration deriving devices, the correlation deriving unit can derive the correlation between the gas concentration and one or more environmental parameters within the second period, based on multiple measurement information and multiple environmental information within a second period including the period following the first period. The determination unit can determine whether to use the measurement information at the second time point as reference information when updating the correction parameters, based on the measurement information and environmental information at a second time point after the second period and the correlation within the second period.

[0014] In any of the gas concentration deriving devices, the correlation deriving unit can derive a machine learning model as the correlation. This machine learning model is obtained by performing machine learning using environmental information as an explanatory variable and measured values ​​as target variables. The determination unit can use the machine learning model to derive a predicted measured value for the environmental information at the first time point, and based on the predicted measured value and the measured value shown by the measurement information at the first time point, determine whether to use the measurement information at the first time point as reference information when updating the correction parameters.

[0015] In any of the gas concentration export devices, the determination unit may determine that the measurement information at the first time point is used as the reference information if the difference between the predicted measurement value and the measurement value shown in the measurement information at the first time point is below a threshold.

[0016] In any of the gas concentration deriving devices, the updating unit can derive statistical values ​​based on the measured values ​​shown by the plurality of reference information, and update the correction parameters based on the statistical values.

[0017] In any of the gas concentration export devices, the updating unit can export the statistical value based on the measured values ​​shown by the plurality of reference information within a third period shorter than the first period.

[0018] In any of the gas concentration deriving devices, if the determination unit determines that the measurement information during the third period is used as the reference information for a predetermined number of consecutive times, the update unit derives the statistical value based on the measurement value shown by each of the plurality of reference information during the third period.

[0019] In any of the gas concentration export devices, there may also be a notification unit that notifies the outside of information based on the determination result of the determination unit.

[0020] Any of the gas concentration extraction devices may further include: a reliability extraction unit that extracts the reliability of the determination result of the determination unit; and a notification unit that notifies the outside of information based on the reliability.

[0021] In a second aspect of the present invention, a gas concentration derivation method is provided, comprising the following stages: obtaining measurement information, the measurement information representing a measurement value corresponding to a gas concentration in the environment of the measurement target measured by a gas concentration measuring unit; determining, based on multiple measurement information within a first period, whether to use the measurement information at a first time point after the first period as reference information when updating the correction parameter; and updating the correction parameter based on at least one of the reference information.

[0022] In a third aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a computer, causes the computer to function as the following components: a measurement information acquisition unit that acquires measurement information, the measurement information representing a measurement value corresponding to a gas concentration in the target environment measured by a gas concentration measurement unit; a determination unit that, based on multiple measurement information within a first period, determines whether to use the measurement information at a first time point after the first period as reference information when updating the correction parameter; a gas concentration derivation unit that, based on the measurement information and the correction parameter, derives the gas concentration in the target environment; and an update unit that updates the correction parameter based on at least one of the reference information.

[0023] In a fourth aspect of the present invention, a gas concentration deriving device is provided, comprising: a measurement information acquisition unit that acquires measurement information, the measurement information representing a measurement value corresponding to a gas concentration in a target environment measured by a gas concentration measurement unit; an environmental information acquisition unit that acquires environmental information, the environmental information representing one or more environmental parameters in the target environment measured by an environmental information measurement unit; a gas concentration deriving unit that derives the gas concentration in the target environment based on the measurement information, the environmental information, and a correction parameter; a correlation deriving unit that derives a correlation between the gas concentration and the environmental information within a first period based on multiple gas concentrations and multiple environmental information within a first period; and an update unit that derives a predicted primary gas concentration at a first time point after the first period and the correlation within the first period, and updates the correction parameter based on the primary gas concentration derived by the gas concentration deriving unit based on the correction parameter for the measurement information at the first time point and the predicted primary gas concentration at the first time point. In the gas concentration exporting device, the gas concentration exporting unit exports the secondary gas concentration at the first time point based on the measurement information at the first time point, the environmental information at the first time point, and the correction parameters at the first time point updated by the updating unit.

[0024] In the gas concentration output device, the correction parameters may include at least one of offset information, gain information, and curvature information.

[0025] In any of the gas concentration exporting devices, the updating unit may update the correction parameter based on the difference between the predicted primary gas concentration and the primary gas concentration exported by the gas concentration exporting unit at the first time point, if the difference is below a threshold.

[0026] In any of the gas concentration deriving devices, the correlation deriving unit can derive the correlation between the gas concentration and one or more environmental parameters within the second period, based on multiple gas concentrations and multiple environmental information within the second period, including the period following the first period. The updating unit can derive the predicted primary gas concentration at the second time point based on the environmental information at the second time point after the second period and the correlation within the second period, and update the correction parameters based on the primary gas concentration derived by the gas concentration deriving unit for the environmental information at the second time point and the predicted primary gas concentration at the second time point. The gas concentration deriving unit can derive the secondary gas concentration at the second time point based on the measurement information at the second time point, the environmental information at the second time point, and the correction parameters at the second time point updated by the updating unit.

[0027] In any of the aforementioned gas concentration export devices, the correlation export unit can export a machine learning model as the correlation. This machine learning model is obtained by performing machine learning using environmental information as an explanatory variable and the gas concentration exported by the gas concentration export unit as the target variable. The update unit can use the machine learning model to export a predicted primary gas concentration based on the environmental information at the first time point, and update the correction parameters based on the predicted primary gas concentration and the primary gas concentration exported by the gas concentration export unit at the first time point.

[0028] In any of the gas concentration exporting devices, the updating unit may update the correction parameter based on the difference between the predicted primary gas concentration and the primary gas concentration exported by the gas concentration exporting unit at the first time point, if the difference is below a threshold.

[0029] In a fifth aspect of the present invention, a gas concentration derivation method is provided, comprising the following stages: obtaining measurement information, the measurement information representing a measurement value corresponding to a gas concentration in a target environment measured by a gas concentration measuring unit; obtaining environmental information, the environmental information representing one or more environmental parameters in the target environment measured by an environmental information measuring unit; deriving the gas concentration in the target environment based on the measurement information, the environmental information, and predetermined offset information; deriving the correlation between the gas concentration and the environmental information within the first period based on multiple gas concentrations and multiple environmental information within the first period; and deriving a predicted primary gas concentration at the first time point based on the primary gas concentration derived from the measurement information at a first time point after the first period, the environmental information, and the correlation within the first period, based on the stage of deriving the gas concentration from the measurement information at a first time point after the first period, and updating the correction parameters based on the primary gas concentration at the first time point derived from the stage of deriving the gas concentration and the predicted primary gas concentration at the first time point. In the gas concentration derivation method, the stage of deriving the gas concentration includes the following stages: deriving the secondary gas concentration at the first time point based on the measurement information at the first time point, the environmental information at the first time point, and the correction parameters at the first time point after being updated during the update stage.

[0030] In a sixth aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a computer, causes the computer to function as the following components: a measurement information acquisition unit that acquires measurement information, the measurement information representing a measurement value corresponding to a gas concentration in a target environment measured by a gas concentration measurement unit; an environmental information acquisition unit that acquires environmental information, the environmental information representing one or more environmental parameters in the target environment measured by an environmental information measurement unit; a gas concentration derivation unit that derives the gas concentration in the target environment based on the measurement information, the environmental information, and correction parameters; a correlation derivation unit that derives the correlation between the gas concentration and the environmental information within the first period based on multiple gas concentrations and multiple environmental information within the first period; and an update unit that, based on the primary gas concentration derived by the gas concentration derivation unit based on the correction parameters for the measurement information at a first time point after the first period, the environmental information, and the correlation within the first period, derives a predicted primary gas concentration at the first time point, and updates the correction parameters based on the primary gas concentration derived by the gas concentration derivation unit at the first time point and the predicted primary gas concentration at the first time point. In the program, the gas concentration derivation unit derives the secondary gas concentration at the first time point based on the measurement information at the first time point, the environmental information at the first time point, and the correction parameters at the first time point updated by the update unit.

[0031] In a seventh aspect of the present invention, a sensor system is provided, comprising: any of the aforementioned gas concentration exporting devices; the gas concentration measuring unit; and the environmental information measuring unit.

[0032] In an eighth aspect of the present invention, the sensor system may include: a housing; a detection unit including at least the gas concentration measuring unit for taking in air from the environment of the target measurement object; and a ventilation unit for ventilating the interior of the housing. The time until 90% of the volume of the housing is ventilated by diffusion can be 60 seconds or less. The duration of the first period can be longer than the time until the ventilation is performed.

[0033] Furthermore, the above summary of the invention does not list all the features of the invention. Additionally, sub-combinations of these feature groups can also constitute inventions. Attached Figure Description

[0034] Figure 1 An example block diagram showing the structure of sensor system 100 is shown.

[0035] Figure 2An example of a graph showing the relationship between the measured values ​​derived by the gas concentration deriving unit 26 and the predicted values ​​after a period shown by the width of the sliding window 60.

[0036] Figure 3 This is an example of a chart showing the correlation between environmental information and measured values, with the horizontal axis representing environmental information and the vertical axis representing measured values.

[0037] Figure 4 The figure shows an example of the treatment used when the difference between the predicted and measured values ​​is greater than 62.

[0038] Figure 5 The results of a simulation of the response of the gas sensor, with the gas exchange time as a parameter, relative to the housing 50 of the gas sensor, under an environment in which the external gas concentration varies in a period of 180 seconds.

[0039] Figure 6 An example flowchart of a gas concentration derivation method according to an embodiment is shown.

[0040] Figure 7 An example of a graph showing the relationship between the measured values, environmental information, and primary gas concentrations measured by the gas concentration measuring unit 12 is presented.

[0041] Figure 8 An example of a chart showing measurement information, environmental information, primary gas concentration, and secondary gas concentration.

[0042] Figure 9 This is an example of a graph showing the correlation between environmental information and primary gas concentration, with the horizontal axis representing environmental information and the vertical axis representing primary gas concentration.

[0043] Figure 10 An example flowchart of a gas concentration derivation method according to an embodiment is shown.

[0044] Figure 11 An example of a computer 1200 that can embody the solution of this embodiment in whole or in part is shown.

[0045] Label Explanation

[0046] 12 Gas Concentration Measurement Unit

[0047] 14 Environmental Information Measurement Department

[0048] 20 Control Department

[0049] 22 Measurement Information Acquisition Department

[0050] 24 Environmental Information Acquisition Department

[0051] 26 Gas Concentration Drawout Section

[0052] 28. Derivation of Relationships

[0053] 30 Judgment Department

[0054] 32 Update Department

[0055] 34 Reliability Derivation Section

[0056] 36 Notification Department

[0057] 40. Ventilation Section

[0058] 42 Storage Section

[0059] 44 Ministry of Communications

[0060] 50 Housing

[0061] 60 Sliding window

[0062] 62 poor

[0063] 64 Calculation Window

[0064] 100 Sensor Systems

[0065] 1200 computers

[0066] 1210 Main Controller

[0067] 1212 CPU

[0068] 1214 RAM

[0069] 1220 Input / Output Controller

[0070] 1222 Communication Interface

[0071] 1230 ROM Detailed Implementation

[0072] The present invention will now be described through embodiments thereof, but these embodiments do not limit the invention as defined in the claims. Furthermore, the combinations of features described in the embodiments are not necessarily all necessary for the solutions provided by the invention.

[0073] Figure 1 An example block diagram illustrating the structure of a sensor system 100 is shown. The sensor system 100 includes components within a housing 50 for detecting gas concentrations in the target environment. The sensor system 100 includes a gas concentration measuring unit 12, an environmental information measuring unit 14, a control unit 20, a ventilation unit 40, a storage unit 42, and a communication unit 44.

[0074] The gas concentration measuring unit 12 may be a gas sensor that measures the gas concentration of an object in the target environment. The gas concentration measuring unit 12 outputs measurement information, representing the measurement value corresponding to the gas concentration continuously measured at unit measurement intervals, to the control unit 20. The gas concentration measuring unit 12 may include a detection unit for sensing the gas concentration contained in the air of the target environment taken into the housing 50. For example, the unit measurement interval may be 0.3 seconds, 10 seconds, or 60 seconds. The unit measurement interval may be less than one-third or less of the time width of the first period.

[0075] The gas concentration measuring unit 12 can output a signal corresponding to the concentration of the gas to be measured, which is detected as a voltage signal, as measurement information. The signal output by the gas concentration measuring unit 12 can be a voltage signal or a current signal. The gas concentration measuring unit 12 can operate from a commercial power supply or an externally supplied DC constant voltage source or constant current source.

[0076] The sensor system 100 of this embodiment is envisioned for use in an environment where environmental parameters change, including at least one of temperature, humidity, atmospheric pressure, wind direction, wind speed, solar radiation, dust level, noise, vibration, illuminance, electromagnetic waves, X-ray radiation, radiation, ozone concentration, and rainfall.

[0077] The gas detection method performed by the gas concentration measuring unit 12 may include at least one of the following: NDIR (non-dispersive infrared), (metal oxide) semiconductor, OGI (optical gas imaging), TDLAS (tunable diode laser absorption spectroscopy), DIAL (differential absorption LiDAR), TCSPC (time-correlated single photon counting), photoacoustic, solid electrolyte, thermal conductivity, acoustic wave, and electrostatic capacitance methods.

[0078] The gas detected by the gas concentration measuring unit 12 can be CO2 (carbon dioxide), H2O (water vapor), or O2 (oxygen). The gas being detected can be flammable gases such as CH4 (methane), C3H8 (propane), C2H5OH (ethanol), H2 (hydrogen), C2H4 (ethylene), or MCH (methylcyclohexane). Alternatively, the gas being detected can be toxic gases such as CO (carbon monoxide), H2S (hydrogen sulfide), CH2O (formaldehyde), or NH3 (ammonia). Furthermore, the gas being detected can be greenhouse gases such as CO2 (carbon dioxide), N2O (nitric oxide), or refrigerant gases.

[0079] NDIR sensors utilize the absorption characteristics of infrared light. In a sensor system 100 with a light source and detector in the infrared wavelength (frequency) region, the absorption wavelength of the target gas differs. By using a filter suitable for the target gas, the presence of the target gas can be determined. The target gas for NDIR sensors can be CO2 (carbon dioxide) or H2O (water vapor). The target gas can be flammable gases such as CH4 (methane), C3H8 (propane), C2H5OH (ethanol), H2 (hydrogen), C2H4 (ethylene), and MCH (methylcyclohexane). Additionally, the target gas can be toxic gases such as CO (carbon monoxide), H2S (hydrogen sulfide), CH2O (formaldehyde), and NH3 (ammonia). Furthermore, the target gas can be greenhouse gases such as CO2 (carbon dioxide), N2O (nitric oxide), or refrigerant gases. Furthermore, NDIR sensors can be combined with other types of sensors, such as zirconia sensors, to detect O2 (oxygen).

[0080] Metal-oxide-semiconductor (MOS) sensors utilize the contact between the gas being measured and the surface of an element using an oxide such as SnO2. Oxygen ions or gas molecules are adsorbed, causing a change in the semiconductor's conductivity (resistance), which in turn determines the gas concentration. This type of sensor is used for measuring flammable gases (methane, propane, or other hydrocarbons) or CO, etc.

[0081] OGI (Optical Gain Infrared) sensors use infrared cameras (mostly cooled cameras) capable of optically detecting specific wavelengths of infrared light to visualize the absorption or emission of infrared radiation by gases. By utilizing the temperature difference with the background and the difference in the gas's absorption spectrum, OGI sensors can display leaked gas in the camera image as a form of "smoke." This type of sensor is used for the determination of flammable gases (methane, propane, or other hydrocarbons, etc.). It should be noted that OGI sensors, in particular, can be used not for the quantitative determination of gas concentration, but for purposes such as gas detection or leak detection.

[0082] TDLAS sensors utilize a wavelength-tunable semiconductor laser to capture the absorption peaks of gases with a narrow spectral width. By targeting specific absorption lines of gas molecules, TDLAS sensors achieve high selectivity and sensitivity. This type of sensor can be used to measure various gases such as H2O (water vapor), CO, CO2, CH4 (methane), or NH3 (ammonia) by adjusting the laser wavelength.

[0083] The environmental information measurement unit 14 measures one or more environmental parameters in the environment of the target environment. For example, the environmental information measurement unit 14 measures at least one of the following environmental parameters: temperature, humidity, atmospheric pressure, wind direction, wind speed, solar radiation, and rainfall. Especially when the target environment is outdoors, these parameters fluctuate significantly, affecting the measured values ​​of the sensor system 100. The environmental information measurement unit 14 measures the timestamp and these environmental parameters. The environmental information measurement unit 14 outputs a signal representing the measured one or more environmental parameters as environmental information to the control unit 20.

[0084] The control unit 20 is a component that performs control in the sensor system 100. The control unit 20 includes a measurement information acquisition unit 22, an environmental information acquisition unit 24, a gas concentration derivation unit 26, a correlation derivation unit 28, a judgment unit 30, an update unit 32, a reliability derivation unit 34, and a notification unit 36.

[0085] The measurement information acquisition unit 22 acquires measurement information representing the measurement value corresponding to the gas concentration in the environment of the object being measured. The measurement information acquisition unit 22 can acquire the measurement information based on the signal representing the gas concentration measurement value output by the gas concentration measurement unit 12. The measurement information acquisition unit 22 can output the acquired measurement information to the gas concentration derivation unit 26.

[0086] The environmental information acquisition unit 24 acquires environmental information representing one or more environmental parameters in the environment of the target environment. The environmental information acquisition unit 24 can acquire environmental information based on signals representing one or more environmental parameters output by the environmental information measurement unit 14. The environmental information acquisition unit 24 can output the acquired environmental information to the gas concentration export unit 26.

[0087] In measurements performed by the gas concentration measuring unit 12, the measured values ​​shown in the measurement information consistently contain errors. Therefore, even when the target gas is not present, the gas concentration may not be detected as 0. Therefore, the gas concentration derivation unit 26 derives the gas concentration in the target environment based on the measurement information, environmental information, and predetermined correction information representing correction parameters used to compensate for errors. Thus, the gas concentration derivation unit 26 derives the gas concentration by eliminating the influence of such errors from the measurement information.

[0088] However, the correction parameters used to compensate for errors change over time due to variations in the characteristics of the gas sensor, such as the deterioration of its components over the years, or, for example, the deterioration of the optical elements if the gas sensor is an NDIR type sensor. Such variations arise, for example, from the elapsed time since manufacturing or from the cumulative operating time, or from seasonal changes.

[0089] In addition, the calibration parameters include parameters for correcting for sudden degradation in the gas sensor components. Sudden degradation, for example in the case of an NDIR gas sensor, includes sudden changes caused by condensation on the optical elements of the gas sensor. Such changes may occur due to rapid temperature and humidity changes in the environment where the gas sensor is located. Furthermore, the calibration parameters include parameters for correcting for the effects of characteristic changes accompanying environmental variations in the gas sensor components. For example, in the case of an NDIR gas sensor, the characteristics of the gas sensor change over time with changes in the environment of the optical elements of the gas sensor. Such changes may occur, for example, due to temperature and humidity changes in the environment where the gas sensor is located over time.

[0090] Therefore, considering the changes in gas sensor characteristics associated with the deterioration of its components over time, a calibration process (calibration procedure) for updating the calibration parameters is preferred. The calibration process can be, for example, performed in an environment where the target gas is absent. To provide such an environment, one could consider moving the gas sensor to such an environment or stopping the equipment used in the target environment. However, sometimes gas sensors are used in environments where it is difficult to move the gas sensor or stop the equipment used in the target environment. For example, the calibration parameters may include at least one of offset information, gain information, and curvature information.

[0091] For example, offset information represents the difference between the measured value and the theoretical value of the gas sensor output in a zero-point gas environment during calibration processing. When the measured value of the gas sensor output in the measurement environment changes or drifts over time, calibration based on the difference between the measured value and the theoretical value can improve the accuracy of gas concentration.

[0092] For example, gain information represents information used to correct the measured value to a predetermined scale based on the sensitivity characteristics of the gas sensor during calibration. By correcting the gas concentration based on gain information, the accuracy of the gas concentration can be improved, regardless of whether the measured value increases or decreases in scale relative to the measured value during calibration in the measurement environment.

[0093] For example, curvature information refers to information used to correct the nonlinear response characteristics of the output, i.e., the measured value, of a gas sensor for gas concentration. When the nonlinear response characteristics of the measured value change relative to the nonlinear response characteristics at the time of correction, correcting the gas concentration based on curvature information can improve the accuracy of the gas concentration readings.

[0094] Therefore, in this embodiment, a sensor system 100 is provided that can perform calibration processing with high accuracy without requiring the absence of the target gas. In the sensor system 100 of this embodiment, calibration processing is dynamically performed during the measurement of gas concentration in the target environment during the processing of the correlation derivation unit 28, determination unit 30, and update unit 32 described below. As a result, the sensor system 100 can dynamically and continuously update calibration information even when continuous measurements are performed in the measurement environment, and can suppress false detections of gas concentration even without performing calibration processing separately.

[0095] Next, in Figure 1 Based on, refer to Figures 2-4 The functions of the related relationship derivation unit 28, the determination unit 30, and the update unit 32 will be explained. Figure 2 An example graph is shown illustrating the relationship between the measured values ​​obtained by the gas concentration measuring unit 12 and the predicted values ​​after a period indicated by the width of the sliding window 60. Figure 2 The diagram shows the measured values ​​obtained by the gas concentration measuring unit 12 at time t and the measurement information obtained by the measurement information acquisition unit 22.

[0096] The correlation derivation unit 28 derives the correlation between the measurement information and the environmental information within a period indicated by the width of the sliding window 60, based on multiple measurement information and multiple environmental information contained in the sliding window 60 of a predetermined width. The correlation can be a regression curve. When the measurement environment is outdoor, the environmental parameters constituting the environmental information may vary significantly. However, in the absence of the target gas, a predetermined correlation, such as a regression curve, is easily observed between the environmental information and the measurement information.

[0097] On the other hand, when a target gas is present, the measured values ​​included in the measurement information may contain a mixture of measured values ​​indicating that a gas has been detected and deviated significantly from the regression curve, and measured values ​​indicating that a gas has not been detected and the difference between the detected gas and the value predicted according to the regression curve is below a predetermined threshold. In this case, the sensor system 100 excludes the measured values ​​indicating gas detection, selects the measured values ​​indicating undetected gas, derives the average error based on the set of undetected gas measured values, and derives a correction parameter for compensating for the error. The determination unit 30 performs such determination processing of the measured values. Regarding the processing of the determination unit 30, see [reference needed]. Figure 3To be described later.

[0098] Furthermore, the correlation derivation unit 28 can also derive a machine learning model as a correlation by performing machine learning with environmental information as an explanatory variable and measurement information as a target variable. Specifically, the machine learning model used by the correlation derivation unit 28 can be a machine learning model using at least one of the following: the Hotelling method, the subspace method, or the Bayesian method. Additionally, the correlation derivation unit 28 can also determine the correlation by performing frequency analysis, such as Fast Fourier Transform (FFT) or Wavelet Transform, on the environmental information and measurement information obtained as time information. Furthermore, the correlation derivation unit 28 can model the correlation by performing state estimation, such as using a Kalman filter, on the environmental information and measurement information. Finally, the correlation derivation unit 28 can determine the correlation by performing matrix decomposition, such as singular value decomposition, on the environmental information and measurement information within a specified period.

[0099] The T of the Hotelling Law 2 A statistical measure is used to evaluate the degree to which a sample deviates from the mean, assuming a multivariate normal distribution. The Hotelling method is as follows: for a feature vector x of a specified dimension, it uses the sample mean μ and variance-covariance matrix estimated from normal data to evaluate T, defined by the square of the Mahalanobis distance. 2 The statistic separates normal values ​​from outliers. Determining the variance-covariance matrix in the Hotelling method is equivalent to determining the correlation. The environmental parameters and the mean of the normal values ​​of the measured values ​​are plotted on a coordinate space with the environmental parameters as the axes. For the deviation of the eigenvector x relative to the sample mean μ under varying environmental parameters, the components of the variance-covariance matrix and T are evaluated. 2 The contribution of statistics can be used to evaluate the correlation of multidimensional data.

[0100] The subspace method involves estimating a low-dimensional space (subspace) within the normal value set using methods such as principal component analysis. When data for a new sample is obtained, the magnitude of the reconstruction error when projected onto this low-dimensional space is evaluated, thereby determining whether the new sample data is abnormal. In this embodiment, when environmental parameters are the principal components, determining the principal component loading representing the change in the measured value corresponding to the change in the environmental parameters is equivalent to determining the correlation. Furthermore, in this embodiment, the case of "representing an outlier" corresponds to a measured value that deviates from the correlation shown by the normal values ​​through the sensor system 100.

[0101] Furthermore, when using the Bayesian method, the model parameters are treated as a set of unknown parameters θ, and the prior distribution p(θ) is set based on the values ​​that θ is likely to take. Then, it is assumed that relative to the observed data D(x)...i y i (x) i For environmental parameters, y i The regression relationship is given for the measured values. Assuming a linear regression relationship, the measured values ​​y are assumed to be... i Following the average of α+βx i The normal distribution y (where α is the regression intercept and β is the slope) i ~N(α+βx) i , σ 2 )(σ 2 This is a probabilistic model for the error variance. In this case, α, β, σ 2 Since all parameters are unknown, they can be defined as θ = (α, β, σ). 2 ).

[0102] The likelihood p(D|θ) of the observed data D can be defined as each (x i y i The regression independently follows the normal distribution described above. In this case, based on Bayes' theorem, the posterior distribution p(θ|D) = (p(D|θ) × p(θ)) / P(D) is derived, and the high-density intervals in the posterior distribution are derived, thus enabling the evaluation of the reliability of the correlation. Here, P(D) is the value obtained by integrating over θ in the entire space of θ, serving as a normalization constant. In this case, by investigating the high-density intervals in the derived posterior distribution p(θ|D), the strength and reliability of the correlation can be evaluated. In this approach, determining the regression relationship itself is equivalent to determining the correlation, and the significance of the correlation can be quantitatively grasped through the posterior distribution of θ.

[0103] After the period shown by the width of the sliding window 60, when a unit measurement interval has elapsed, the gas concentration measuring unit 12 measures the value corresponding to the new gas concentration. As an example, the environmental information measuring unit 14 measures environmental information at the same time interval.

[0104] The determination unit 30 derives a predicted measurement value at a given time point based on environmental information at a time interval after the period shown by the width of the sliding window 60 and the correlation within that period. Therefore, at that time point, based on environmental information and correlations such as regression curves, a measurement value corresponding to the predicted gas concentration is derived.

[0105] An example of such a regression curve is shown in Figure 3 . Figure 3This is an example of a graph showing the correlation between environmental information and measured values, with the horizontal axis representing environmental information and the vertical axis representing measured values. The graph shows an example where the correlation is a first-order regression curve (i.e., a straight line). However, the case where the correlation is represented by a first-order regression curve is merely illustrative; the correlation can also be represented by higher-order regression curves (second order or higher).

[0106] Here, as described above, environmental information exemplifies environmental parameters such as temperature, humidity, or wind speed, but is not limited to these. As another example, it could also be based on a signal emitted by a reference sensor that monitors the emission level of an LED emitting infrared light for measurement by the gas concentration measuring unit 12. Since the emission level of the LED depends on temperature information, the signal emitted by the reference sensor can be set as information that has acquired a portion of the environmental information.

[0107] Furthermore, the determination unit 30 compares the derived predicted measurement value with the measurement value shown in the measurement information at that time point. Even in the absence of the target gas, a certain correlation can easily arise between environmental factors and the measurement value, even in a correction environment where the influence of environmental factors is not intentionally removed. On the other hand, when the target gas is present in the target environment, outliers, including those with a correlation deviating from the regression curve and where the difference 62 from the regression curve easily increases, can occur. Even when outliers are present in the target environment, by continuously updating the correction information using measurement values ​​determined as normal by the determination unit 30, more accurate gas concentrations can be continuously derived using correction information suitable for the environmental information of the target environment. By using this correction information, the sensor system 100 can measure the gas concentration after removing errors caused by changes in environmental information (temperature, humidity, or wind speed, etc.) or by the deterioration of components of the sensor system 100 over time.

[0108] The determination unit 30 distinguishes between normal and abnormal values ​​from the measured values ​​by determining whether the measured value represents a value close to a predicted measured value based on such a correlation. The sensor system 100 can use the average value of the normal values ​​as a correction parameter for the measured values ​​in the environment of the target object, when the target gas is present in the environment of the target object.

[0109] The period shown by the width of such a sliding window 60 is an example of a "first period". Additionally, a time point after the period shown by the width of the sliding window 60, at which a unit measurement interval has elapsed, is an example of a "first time point". The period shown by the width of the sliding window 60 can be a period that includes environmental changes, such as changes in wind direction or wind speed in the environment of the measured object, and can be an integer multiple of the unit measurement interval, such as 60 seconds, 1 hour, or 3 hours.

[0110] Through this comparison, the determination unit 30 determines, based on the measurement information and environmental information at a time point after a unit measurement interval following the period shown by the width of the sliding window 60, and the correlation within the period shown by the width of the sliding window 60, whether the measurement information at a time point after a unit measurement interval following the period shown by the width of the sliding window 60 should be used as reference information when updating the offset information. Since the correction information is updated based on this reference information, the measurement information that is determined not to be used as reference information is evaluated as a deviation value (outlier) in the measurement and is not used as reference information for updating the correction information.

[0111] Specifically, if the difference 62 between the predicted measurement value and the measurement value shown in the measurement information at a time point after a unit measurement interval following the period shown by the width of the sliding window 60 is below a threshold, the determination unit 30 determines that the measurement information at that time point is used as reference information. On the other hand, if the difference 62 between the predicted measurement value and the measurement value shown in the measurement information at a time point after a unit measurement interval following the period shown by the width of the sliding window 60 is greater than the threshold, the determination unit 30 determines that the measurement information at that time point is not used as reference information. Therefore, by evaluating the difference 62 between the measurement value predicted based on the correlation and the measurement value shown in the measurement information, appropriate reference information is continuously obtained in order to dynamically update the correction information. Furthermore, the period of the sliding window 60 may also include multiple measurement information and environmental information after the first time point.

[0112] Furthermore, a threshold for judgment can be set using a statistic determined based on the correlation for the predicted measured value. The threshold can be a quantity representing the difference in magnitude of the amount determined by the standard deviation of the predicted measured value; more specifically, the threshold can be 3 standard deviations (3σ). The judgment unit 30 can determine whether the difference 62 between the predicted measured value and the measured value shown in the measurement information is within 3σ, that is, whether the measured value is within the range of (predicted measured value ± 3σ). However, the threshold can be set to 1 standard deviation (1σ) or 2 standard deviations (2σ) depending on the desired judgment accuracy.

[0113] Furthermore, the updating unit 32 updates the correction information based on at least one piece of reference information determined as reference information by the determination unit 30. The updating unit 32 can derive statistical values ​​based on the measured values ​​shown by each of the multiple pieces of reference information, and update the correction information based on the statistical values. The statistical values ​​can be the mean, median, maximum, minimum, variance, or moments, etc.

[0114] In addition, the sliding window 60 slides in the positive direction of the time axis as time elapses. The period indicated by the width of the sliding window 60 after sliding is an example of a "second period". For example, the time width between the first period and the second period may be an integer multiple of the unit measurement interval.

[0115] In this case, the correlation derivation unit 28 can derive the correlation between the measured values and one or more environmental parameters within the period indicated by the width of the sliding window 60 after sliding, based on a plurality of measurement information and a plurality of environmental information within the period indicated by the width of the sliding window 60 after sliding, which includes the period after the period indicated by the width of the sliding window 60 before sliding. In addition, the correlation derivation unit 28 can derive the correlation between the gas concentration and one or more environmental information within the period indicated by the width of the sliding window 60 after sliding, based on a plurality of measurement information and a plurality of environmental information within the period indicated by the width of the sliding window 60 after sliding, which includes the period after the period indicated by the width of the sliding window 60.

[0116] The determination unit 30 can determine whether to use the measurement information at the second time point as reference information when updating correction information, based on the measurement information and environmental information at the time point after the elapse of the unit measurement interval following the period indicated by the width of the sliding window 60 after sliding, and the correlation within the period indicated by the width of the sliding window 60 after sliding.

[0117] The time point after the elapse of the unit measurement interval following the period indicated by the width of the sliding window 60 after sliding is an example of a "second time point".

[0118] Here, with reference to Figure 4 , the calculation window 64 will be described. Figure 4 It is a diagram showing an example of processing performed when the difference 62 between the predicted measured value and the actual measured value is large.

[0119] At time t0, the period indicated by the width of the sliding window 60 elapses. In this case, the determination unit 30 predicts the measured value at the time point after the elapse of the unit measurement interval following the period indicated by the width of the sliding window 60.

[0120] Time t1 corresponds to the time point after the unit measurement interval has elapsed from time t0. The determination unit 30 determines at time t1 whether the difference 62 between the predicted measured value and the measured value indicated by the measurement information at time t1 is equal to or less than a threshold value.

[0121] In the example shown in the figure, a case where the difference 62 exceeds the threshold value at time t1 is shown. In this case, the determination unit 30 determines not to add the measured value at time t1 to the reference information.

[0122] At time t2, a time interval further elapsed from time t1, the determination unit 30 determines whether to use the measurement information at time t2 as reference information when updating the correction information, based on the measurement information and environmental information at time t2 and the correlation derived by the correlation derivation unit 28. In this case, the correlation derived by the correlation derivation unit 28 may also be, like the measurement value at time t1, derived from the measurement values ​​included in the calculation window 64 after excluding measurement values ​​where the difference 62 exceeds the threshold.

[0123] That is, the calculation window 64 is a time window with a width shorter than that of the sliding window 60, and it is a time window that only includes the measurement information (determined to represent normal values) that the determination unit 30 determines to be added to the reference information. The calculation window 64 slides in the positive direction of the time axis as time passes. As an example, the measurement information included in the calculation window 64 used by the correlation derivation unit 28 to derive the correlation may be the measurement information of the immediate preceding time points that the determination unit 30 determines will not be added to the reference information. By using such immediate preceding time points, gas concentration correction is performed based on the most recent normal data. As a result, compared with the case where past trend data is mixed in, the sensor system 100 can perform gas concentration correction well. Thus, in the sensor system 100 of this embodiment, gas concentration correction that takes into account time-series data can be performed.

[0124] The width of the calculation window 64 is set to be shorter than the width of the sliding window 60. For example, the regression curve between environmental information and measured values ​​can be derived as a correlation curve based on approximately three or more measurement points.

[0125] The determination unit 30 determines the difference 62 between the predicted measurement value and the measurement value shown in the measurement information at time t2, a time point after a unit measurement interval from time t1. If the difference 62 between the predicted measurement value and the measurement value shown in the measurement information at time t2 is within a threshold, the determination unit 30 determines that the measurement information at time t2 should be used as reference information when updating the correction information. On the other hand, if the difference 62 between the predicted measurement value and the measurement value shown in the measurement information at time t2 exceeds the threshold, the determination unit 30 determines that the measurement information at time t2 should not be used as reference information when updating the correction information.

[0126] Furthermore, the update unit 32 derives statistical values ​​based on the measured values ​​shown in each of the multiple reference information entries, and updates the offset information based on the statistical values. In this case, the update unit 32 can derive statistical values ​​based on the measured values ​​shown in each of the multiple reference information entries within the period indicated by the width of the calculation window 64.

[0127] The updating unit 32 can derive statistical values ​​based on the measurement values ​​shown by each of the multiple reference information within the period shown by the width of the calculation window 64, when the determination unit 30 determines that the measurement information within the period is used as reference information for a predetermined number of consecutive times. In the case where the target gas is present in the environment of the target object, multiple outliers may occur. Therefore, even when the target gas is present in the environment of the target object, the updating unit 32 can update the correction information based on the accumulated normal values ​​as normal values ​​continuously accumulate.

[0128] The period shown by the width of calculation window 64 is an example of a "third period". The period shown by the width of calculation window 64 can be a period with minimal impact from environmental changes, including variations in wind direction or speed within the measured environment; for example, it could be 10 seconds, 60 seconds, or 10 minutes. For instance, the time width between the first and third periods can be an integer multiple of the unit measurement interval, and the period shown by the width within the third period can also include periods outside the range of the first period, excluding the first time point. The time width of the third period can be shorter than the time width of the first period.

[0129] Refer again Figure 1 The functions of the reliability output unit 34, notification unit 36, ventilation unit 40, and storage unit 42 will be explained.

[0130] The reliability derivation unit 34 derives the reliability of the judgment result from the judgment unit 30. The reliability derivation performed by the reliability derivation unit 34 can be based on mathematical statistics. For example, the reliability calculation can be based on the evaluation of the standard deviation relative to the average of the measured values ​​based on bootstrapping sampling, or it can be based on the evaluation of the proportion of measured values ​​that exceed a predetermined threshold from the average value. As another example, the reliability can be determined based on the proportion of measured values ​​that deviate from the percentile width of the range of normal values ​​(corresponding to a threshold for a difference of 62) and are thus judged as outliers. For example, the reliability derivation can be based on residual analysis, F-test, Akaike information criterion, or Bayesian information criterion. For example, the reliability can be derived by the correlation derivation unit 28 using residual analysis to analyze the measurement information and residuals of the correlation and evaluating the reliability of the correlation.

[0131] The notification unit 36 ​​notifies external parties of information based on the determination results of the determination unit. Specifically, the notification unit 36 ​​notifies the manager when the gas concentration value exceeds a threshold. Specifically, the notification unit 36 ​​may notify the manager that an anomaly caused by gas generation has occurred in the environment of the object being measured, or that anomaly determinations have occurred consecutively, if the number of consecutive occurrences of anomaly determinations by the determination unit 30 exceeds a predetermined threshold.

[0132] In addition, the notification unit 36 ​​can notify external parties of information based on reliability. For example, the notification unit 36 ​​can notify the manager if the reliability of the judgment result or concentration value is lower than a specified threshold.

[0133] The ventilation unit 40 performs ventilation by continuously drawing air from the environment of the object being measured into the housing 50. For example, the time until the ventilation unit 40 ventilates 90% of the volume of the housing 50 through diffusion ventilation can be set to within 60 seconds, or it can be set to twice the average or median time of environmental change cycles, including changes in wind speed, wind direction, temperature, and humidity.

[0134] Similarly, the gas sensor measurement cycle can be set to within 60 seconds. For example, when the gas sensor's ventilation time and measurement cycle are within 60 seconds, the gas trapped inside the gas sensor housing is quickly expelled, thus reducing the response delay to changes in gas concentration outside the gas sensor housing and enabling more accurate gas measurement.

[0135] For example, when the gas sensor's ventilation time and measurement cycle are within 60 seconds, measurement and environmental information can be acquired at a faster rate than the change in offset value caused by environmental variations. Therefore, it is possible to correct for the offset value caused by environmental changes and to measure the gas more accurately. As a specific installation example, the ventilation time and gas sensor measurement cycle can be set to be shorter than the duration of a first period that includes the effects of environmental variations. That is, the duration of the first period can be longer than the time until ventilation is performed.

[0136] The ventilation unit 40 can perform ventilation within the housing 50 at intervals less than the environmental change cycle. The ventilation time for the ventilation unit 40 to ventilate 90% of the volume within the housing 50 via diffusion can be designed to be less than this cycle, and can be set to other times. As a specific embodiment, the ventilation time for the ventilation unit 40 to ventilate the entire housing 50 can be set to 10 seconds, 30 seconds, or 60 seconds.

[0137] When the ventilation unit 40 is equipped with a mechanism such as a fan for forced ventilation, the ventilation cycle performed by the ventilation unit 40 within the housing 50 can be less than the cycle of environmental changes. The time taken by the ventilation unit 40 to ventilate 90% of the volume within the housing 50 by forced ventilation can be designed to be other times, as long as it is less than that cycle. As a specific embodiment, the time for the ventilation unit 40 to perform forced ventilation throughout the entire housing 50 can be set to 10 seconds, 30 seconds, or 60 seconds.

[0138] The storage unit 42 stores the information output by the control unit 20. Specifically, the storage unit 42 stores measurement information at each time point, environmental information, calibration information, gas concentration, and correlations within the period shown by the width of the sliding window 60.

[0139] The communication unit 44 communicates with other devices, computers, or computer systems. Thus, the communication unit 44 can send and receive information related to the measurement results of the sensor system 100 with other devices, computers, or computer systems.

[0140] Figure 5 The results of a response simulation are shown, with the gas sensor's ventilation time as a parameter, under an environment where the external gas concentration varies relative to the gas sensor housing 50 with a period of 180 seconds. Here, the horizontal axis represents time (seconds), and the vertical axis represents gas concentration (ppm).

[0141] like Figure 5 As shown, in this embodiment, the effect of the gas sensor's ventilation time on the gas sensor's responsiveness under varying wind direction or speed conditions in an outdoor environment is calculated. Ventilation of the housing 50 is based on diffusion; forced ventilation mechanisms such as fans are not used for ventilation of the housing.

[0142] Non-Patent Document 1 describes that the timescale of turbulence in the atmospheric boundary layer is approximately 10 to 180 seconds. In outdoor environments, gas concentration and gas diffusion direction change within this timescale. Therefore, the gas exchange section 40 of the gas sensor preferably replaces the gas inside the housing 50 within a time shorter than this timescale.

[0143] Based on response simulations, it was confirmed that when the ventilation time exceeds 60 seconds, the gas concentration inside the housing 50 does not reach equilibrium within the timescale of wind speed changes caused by turbulence, and therefore cannot follow changes in gas concentration outside the housing. On the other hand, it was confirmed that when the ventilation time is less than 60 seconds, the gas concentration inside the housing 50 follows changes in external gas concentration well, enabling accurate detection of gas arrival events. Furthermore, it was confirmed that when the ventilation time is less than 30 seconds, the output response time of the gas sensor is consistent with environmental changes without significant hysteresis, making the gas sensor effective for detecting instantaneous leaks outdoors.

[0144] In this way, the time until 90% of the volume of the casing 50 is ventilated through diffusion can be less than 60 seconds. Furthermore, the time until 90% of the volume of the casing 50 is ventilated through diffusion can be less than 30 seconds, in which case changes in gas concentration can be accurately detected on the timescale of wind speed or wind direction changes.

[0145] The gas sensor may include: a ventilation section 40, which is a measurement space for acquiring measurement information; a housing 50, which houses the ventilation section; a ventilation port, which is located on a part of the housing 50 and connects the external space to the measurement space to allow gas flow; and a filter material, which is located at the ventilation port. The gas sensor may also include: an optical path and optical components that pass through the measurement space at least once; a light-emitting section that emits infrared light; a light-receiving section that receives infrared light; and a structure for fixing various components required for the gas sensor.

[0146] For such a gas sensor, a gas exchange test that measures the time it takes for 90% of the volume within the housing 50 to be exchanged through diffusion in the gas exchange section 40 can be conducted, for example, in a case where the gas sensor is installed in a closed space intended to isolate it from natural turbulence. In this case, taking the initial state as a state where the gas exchange section 40 of the gas sensor is sealed with the detection gas and the gas exchange port of the gas exchange section 40 is sealed, the time it takes for the gas in the gas exchange section 40 to naturally diffuse through the gas exchange port into the closed space from the time the gas exchange port is opened and sealed is measured.

[0147] In this ventilation test, the moment when the seal of the ventilation port is opened is set to t0, and the gas concentration of the ventilation section 40 in the initial state is set to C0. Furthermore, when 90% of the gas in the ventilation section 40 is replaced after a certain period of time from time t0 is set to t1, and the gas concentration in the ventilation section 40 at time t1 is set to C1, the following relationship (1) holds.

[0148] (C0-C1) / C0=0.9 ···(1)

[0149] In this case, the elapsed time between time t0 and time t1 corresponds to the gas exchange time of the gas sensor.

[0150] Furthermore, in the ventilation test, if it is difficult to seal the ventilation port of the gas sensor and enclose the gas in the ventilation section 40, the gas sensor can be installed in a second enclosed space that is sufficiently small relative to the enclosed space (relative to the "second enclosed space" described in this paragraph, which corresponds to the "first enclosed space") and larger than the gas sensor, and the gas can be enclosed in the second enclosed space and the ventilation section 40 of the gas sensor. Here, it is preferable that the second enclosed space has a sufficiently large ventilation port, and the time it takes for the gas to naturally diffuse into the enclosed space through the second enclosed space and the gas sensor ventilation port from the moment the ventilation port of the second enclosed space is opened is measured.

[0151] The ratio (V / S) of the volume V of the ventilation section to the area S of the ventilation port (with the dimension of length) can be at least 200 or less, or it can be 100 or less. Here, when the time for 90% gas exchange in the ventilation section by diffusion is set as τ, and the flow velocity of the gas through the ventilation port is set as v, the following formula holds.

[0152] τ=V / (S·v) …(2)

[0153] The vent can be disposed on at least two or more surfaces of the housing 50. Furthermore, when the vent is disposed on at least two or more surfaces of the housing 50, these two surfaces can be disposed on opposite surfaces within the housing 50. This suppresses pressure rise in the internal space of the venting section 40 during air exchange and allows for rapid discharge of gas from the internal space of the venting section 40.

[0154] Furthermore, when the vents are disposed on at least two or more surfaces of the housing 50, the positions of the vents can be arranged asymmetrically with respect to the geometric center of the internal space of the ventilation section 40. This allows for the suppression of spaces within the internal space of the ventilation section 40 where airflow through the ventilation section 40 does not pass (thus creating "dead zones" relative to the airflow path), and it also suppresses localized stagnation of the detection gas within the internal space of the ventilation section 40 by generating turbulence.

[0155] The direction of the airflow flowing from the gas sensor's vent into the ventilation section 40, which serves as the measurement space, or the opening direction of the vent, can intersect with the direction of the infrared light path within the ventilation section 40, which serves as the measurement space. This allows for efficient gas replacement within the measurement space, and the gas concentration within the light path can rapidly follow changes in the external atmosphere. Specifically, the angle between the direction of the airflow flowing into the ventilation section 40 and the direction of the infrared light path within the ventilation section 40 can be in the range of 60° to 120°, more preferably in the range of 75° to 105°.

[0156] Regarding the top-view geometry of the ventilation section 40, the distance along the longitudinal direction from one side of the ventilation section 40 to the opposite side can be at least 1.5 times the distance along the width direction from the other side of the ventilation section 40 to the opposite side. Furthermore, in this top view, the ventilation opening can be positioned approximately perpendicular to the width direction of the ventilation section 40. This allows for a thinner profile in the direction of the top-view ventilation section 40. Moreover, in the in-plane direction when the ventilation section 40 is viewed from above, for example, if the infrared light path length is set in the longitudinal direction, the infrared light path length can be set to be relatively long.

[0157] The air resistance R of the filter material used as the air exchange port in the air exchange section 40 can be at least 50 kPa·s / m or less, or even 10 kPa·s / m or less. For example, the filter material can be a porous material with dustproof, waterproof, and explosion-proof properties, such as non-woven fabric, membrane filter, or metal mesh. Furthermore, the lower the air resistance R of the filter material, the higher the air exchange capacity and the better the responsiveness of the gas sensor.

[0158] In the gas sensor, at least one air exchange port of the air exchange section 40 is preferably located on a surface different from the heating element of the gas sensor. This allows for efficient exhaust of the gas within the air exchange section 40 from the air exchange port when convection caused by heat generated by the heating element occurs in the internal space of the air exchange section 40.

[0159] Furthermore, the gas sensor of this embodiment may also have a thermally insulating layer on a portion of the outer periphery of the ventilation section 40. This allows it to block the influence of external temperature changes and stably obtain measurement information.

[0160] Furthermore, as an embodiment different from that without forced ventilation, this embodiment of the gas sensor, aimed at shortening the ventilation time of the gas sensor, may include an intake fan for introducing gas from the external environment into the ventilation section 40 and / or an exhaust fan for discharging gas. The airflow Q of the ventilation fan or intake fan can have a Q / V ratio greater than 0.013 relative to the volume V of the ventilation internal space. Therefore, the gas sensor of this embodiment can exhaust more than 90% of the gas inside the housing 50 within 180 seconds, enabling rapid gas replacement of the measurement environment. With this structure, the gas sensor of this embodiment can maintain a small ventilation port area, ensuring sufficient ventilation performance and gas sensor response performance without compromising dustproof, waterproof, and explosion-proof properties.

[0161] The gas sensor can have multiple housings 50. The gas concentration measuring unit 12 and / or the environmental information measuring unit 14 can be housed in one of the multiple housings 50. The gas concentration measuring unit 12 and the environmental information measuring unit 14 can be housed in the same housing 50 or in different housings 50. As an example, the gas concentration measuring unit 12 and / or the environmental information measuring unit 14 can be housed inside a second housing among the multiple housings 50. The second housing can be a structure that ensures dustproof, waterproof, explosion-proof, or thermal insulation relative to the external environment and has a gas flow path inside. The gas concentration measuring unit 12 and / or the environmental information measuring unit 14 are located as part of the gas flow path, and can be positioned opposite to the airflow introduced from the vent communicating with the external atmosphere, or positioned in the direction in which the airflow crosses the optical path. This shortens the response time of the gas sensor and suppresses measurement errors caused by the influence of the external measurement environment.

[0162] The temperature, humidity, and atmospheric pressure sensors of the environmental information measuring unit 14 can be installed inside the ventilation unit 40. Furthermore, other sensors, such as those for wind direction, wind speed, solar radiation, or rainfall, which are also present in the gas measuring sensor, can be installed outside the ventilation unit 40. For example, when the gas sensor is of the NDIR type, by installing the temperature, humidity, or atmospheric pressure sensors inside the ventilation unit 40, it is possible to more accurately correct for changes in offset values ​​due to temperature variations, changes in the amount of light reaching the light-receiving part due to humidity variations, or changes in gas concentration due to changes in optical elements or atmospheric pressure. On the other hand, by installing sensors for wind direction, wind speed, solar radiation, or rainfall outside the ventilation unit 40, the gas sensor of this embodiment can more accurately measure the flow or state of gas in the measurement environment.

[0163] The ventilation port can be composed of at least three through holes. That is, the ventilation section 40 can be a perforated plate with multiple through holes. In this case, the area of ​​the opening portion of each through hole in the ventilation port can be 1 mm². 2 The above. In this case, the ventilation port can also be sealed with a filter material that is dustproof, waterproof, or explosion-proof. Thus, a high opening ratio, structural strength of the outer edge of the ventilation port, and a tight seal between the ventilation port and the filter material can coexist in the ventilation port.

[0164] The ventilation port can be designed such that the opening area on the external space side of the ventilation section 40 is larger than the opening area on the internal space side of the ventilation section 40. This allows airflow from the external space of the ventilation section 40 to be more effectively drawn into the internal space of the ventilation section 40.

[0165] Furthermore, the sampling frequency or the time width of the sliding window 60 can be updated when the measurement information within the sliding window 60 determines that the calibration information parameters do not include the measurement information at the first time point. Alternatively, the sampling frequency or the time width of the sliding window 60 can be updated when the measurement information within the sliding window 60 determines that the calibration information parameters continuously include the measurement information at the first time point. This optimizes the sampling frequency and sampling time, reducing power consumption in the driving or communication of the gas sensor.

[0166] Figure 6 An example flowchart of a gas concentration derivation method according to an embodiment is shown. The gas concentration derivation method includes steps S102 to S122.

[0167] The control unit 20 sets a sliding window (S102). Next, the gas concentration measuring unit 12 measures measurement information that corresponds to the gas concentration in the target environment. The environmental information measuring unit 14 measures environmental information that represents one or more environmental parameters in the target environment.

[0168] Measurement information acquisition unit 22 acquires measurement information within a sliding window, and environmental information acquisition unit 24 acquires environmental information (S104). Next, based on the acquired measurement information and environmental information, environmental information acquisition unit 24 derives a regression curve between the measurement information and environmental information (S106).

[0169] Measurement information acquisition unit 22 acquires measurement information at the measurement time point following the sliding window 60, and environmental information acquisition unit 24 acquires environmental information (S108). Here, the measurement time point following the sliding window 60 can be a time point after a unit measurement interval has elapsed following the period indicated by the width of the sliding window 60. Correlation derivation unit 28 derives the correlation between the measured values ​​and the environmental information within the period indicated by the width of the sliding window 60, based on multiple measurement information and multiple environmental information in the sliding window 60 with a specified width.

[0170] Based on the correlation, the determination unit 30 derives predicted measurement information for measurement time points that have passed a unit measurement interval after the period indicated by the width of the sliding window 60 (S110). Then, the determination unit 30 determines whether the difference 62 between the measured value and the predicted measured value at the measurement time point is below a threshold (S112). If the difference 62 is below the threshold, the process proceeds to S114. On the other hand, if the difference 62 exceeds the threshold, the process proceeds to S116.

[0171] If the difference between the measured value and the predicted measured value at the measurement time point is less than 62, the measurement information at the measurement time point is registered as reference information (S114).

[0172] On the other hand, if the difference 62 between the measured value and the predicted measured value at the measurement time point exceeds a threshold, the control unit 20 can delete the registered reference information (S116). In addition, the notification unit 36 ​​notifies the outside of the anomaly (S118).

[0173] Then, the control unit 20 determines whether the registered reference information has reached the benchmark number (S120). If the registered reference information has reached the benchmark number, the process proceeds to S122. On the other hand, if the registered reference information has not reached the benchmark number, the process returns to S102.

[0174] The updating unit 32 updates the correction information based on the registered reference information (S122). The gas concentration derivation method processing is now complete.

[0175] By using the sensor system 100 or the gas concentration derivation method of this embodiment, even if the measured value is erroneous due to changes in environmental information (temperature, humidity, or wind speed, etc.) or deterioration over time, the measured value can be continuously measured without individual correction processing. This allows for correction of the gas concentration derived from the measured value using only the measured value representing a normal value. Therefore, in the sensor system 100 or the gas concentration derivation method of this embodiment, there is no need for correction processing that requires moving the gas sensor to an environment where the target gas is absent or stopping the equipment in the target environment. Even in environments where gas may be present, the concentration of the target gas can be measured, eliminating errors caused by changes in environmental information (temperature, humidity, or wind speed, etc.) or deterioration of the components of the sensor system 100 over time.

[0176] In the example described above, a determination was made regarding the correlation between multiple measurement information derived by the correlation derivation unit 28 and multiple environmental information within the period shown in the width of the sliding window 60. The determination unit 30 uses the correlation to determine whether to update the reference information, which is used to update the correction parameters at a time point after a unit measurement interval has elapsed following the period shown in the width of the sliding window 60.

[0177] In another embodiment, the determination unit 30 sometimes determines whether to use the measurement information as reference information for correcting the correction parameters based solely on frequency analysis of multiple measurement information within a period indicated by the width of the sliding window 60. The determination unit 30 of the control unit 20 can perform frequency analysis on the measurement information obtained as time information using methods such as Fast Fourier Transform (FFT) or Wavelet Transform.

[0178] However, this analysis can also be performed through other structures of the control unit 20. For example, the correlation derivation unit 28 is described as performing frequency analysis on environmental information and measurement information to derive correlation, but the correlation derivation unit 28 may also perform frequency analysis on the measurement information for the determination unit 30 without deriving correlation.

[0179] When performing frequency analysis on the measurement information, the effects of factors such as the deterioration of various elements of the sensor system 100 over the years or the drift of measurement values ​​caused by environmental changes are more likely to manifest as low-frequency signals compared to signals indicating the presence of the gas to be measured. Therefore, the control unit 20 can derive correction parameters from the low-frequency signals when the gas to be measured is absent based on frequency analysis, and extract information about the presence or absence of the gas to be measured from the high-frequency signals. Consequently, the determination unit 30 can determine, based on multiple measurement information, whether to use the measurement information at a time point after a unit measurement interval has elapsed following the period indicated by the width of the sliding window 60 as reference information. Therefore, the determination unit 30 can determine, based on multiple measurement information within a first period, whether to use the measurement information at a first time point after the first period as reference information when updating the correction parameters.

[0180] In this way, the sensor system 100, based on multiple measurement information obtained in a time series, does not perform correction processing that would involve moving the gas sensor to an environment where the target gas is not present or stopping the equipment in the target environment. Even in an environment where the gas may be present, it can measure the concentration of the target gas, removing errors caused by changes in environmental information (temperature, humidity, or wind speed, etc.) or by the deterioration of the components of the sensor system 100 over the years.

[0181] Next, refer to Figure 7 The functions of the sensor system 100 in different implementations will be described. Figure 7 An example of a graph showing the relationship between the measured values, environmental information, and primary gas concentrations measured by the gas concentration measuring unit 12 is presented.

[0182] In the embodiment shown in the figure, the gas concentration derivation unit 26 also derives the gas concentration based on the measurement information measured by the gas concentration measuring unit 12 and acquired by the measurement information acquisition unit 22, and the environmental information measured by the environmental information measuring unit 14 and acquired by the environmental information acquisition unit 24. Here, the gas concentration derivation unit 26 derives a primary gas concentration from predetermined correction information based on the measurement information and the environmental information. In this embodiment, further, after deriving the primary gas concentration, the correlation derivation unit 28 derives the correlation between the primary gas concentration and the environmental information, and the update unit 32 updates the correction information representing the correction parameters based on the correlation.

[0183] Then, based on the updated correction information from the gas concentration derivation unit 26, the secondary gas concentration is derived. Figure 8 An example of a chart showing measurement information, environmental information, primary gas concentration, and secondary gas concentration.

[0184] Here, it is described that in the measurement information representing the measured values ​​measured by the gas concentration measuring unit 12, even when the concentration of the gas being measured is 0, measurement information may sometimes indicate an error where the measured value is not zero. Correction parameters used to compensate for this error are set before shipment, and the gas concentration deriving unit 26 uses these correction parameters to derive the gas concentration. However, the correction parameters may sometimes change based on elapsed time since manufacturing, cumulative operating time, or seasonal variations. These correction parameters include parameters generated based on elapsed time since manufacturing or cumulative operating time. For example, by storing information in the storage unit 42 for evaluating correction information such as elapsed time since manufacturing, cumulative operating time, or seasonal information, it is sometimes possible to multiply these by a coefficient for quantitative evaluation.

[0185] However, the reasons for the error are difficult to evaluate simply by multiplying by coefficients such as the elapsed time since manufacturing, cumulative operating time, or seasonal information. Specifically, environmental information such as moisture absorption due to humidity and the deterioration of the detector section of the gas concentration measuring unit 12 due to dryness over the years, or the effect of strain on the reflector in the detector section of the gas concentration measuring unit 12 due to temperature changes, appear relatively slowly in the environment in which the sensor system 100 is used. These are also considered to be the slow effect of the time constant generated by the environmental information on the sensor system 100, but such effects can manifest differently depending on the environment in which the sensor system 100 is used, so it becomes difficult to evaluate simply by multiplying the information stored in the storage unit 42 by coefficients.

[0186] In the sensor system 100 of this embodiment, the update unit 32 updates the correction information based on the primary gas concentration and derives the secondary gas concentration using the updated correction information based on the primary gas concentration. This allows for the use of correction parameters derived from the average of normal values ​​to eliminate errors. Specifically, in the sensor system 100, the primary gas concentration itself is not used as the measured gas concentration. Instead, the update unit 32 updates the correction information when deriving the secondary gas concentration from the primary gas concentration, and the gas concentration deriving unit 26 derives the updated secondary gas concentration, which is then used as the measured gas concentration. Therefore, the sensor system 100 can set appropriate correction parameters in the current time context without distinguishing between various errors caused by different factors, thereby reducing errors with higher accuracy.

[0187] Next, referring to the diagram showing the relationship between primary gas concentration and environmental information... Figure 9 The functions of the constituent elements of the sensor system 100 in this embodiment will be explained. Figure 9 This is an example of a graph showing the correlation between environmental information and primary gas concentration, with the horizontal axis representing environmental information and the vertical axis representing primary gas concentration.

[0188] The correlation derivation unit 28 derives the correlation between gas concentrations and environmental information within the period shown by the width of the sliding window 60, based on multiple gas concentrations and multiple environmental information within that period. Thus, the correlation derivation unit 28 derives correlations such as regression curves between environmental information and primary gas concentration within the period shown by the width of the sliding window 60.

[0189] The updating unit 32 derives the predicted primary gas concentration at a time point after a unit measurement interval has elapsed since the period shown by the width of the sliding window 60, based on environmental information and the correlation within the period shown by the width of the sliding window 60. The updating unit 32 also updates and corrects the primary gas concentration derived by the gas concentration derivation unit based on offset information for the measurement information at the time point after a unit measurement interval has elapsed since the period shown by the width of the sliding window 60, and the predicted primary gas concentration at that time point.

[0190] If the difference 66 between the predicted primary gas concentration and the primary gas concentration derived by the gas concentration derivation unit 26 at a time point after a unit measurement interval following the period shown by the width of the sliding window 60 is below a threshold, the update unit 32 can update the correction information based on the difference 66.

[0191] The gas concentration derivation unit 26 derives the secondary gas concentration at a time point after a unit measurement interval has elapsed since the period shown by the width of the sliding window 60, based on measurement information at that time point, environmental information at that time point, and correction information at that time point updated by the update unit 32.

[0192] And, with Figures 2-4 Similarly, in this embodiment, the correlation derivation unit 28 can also make the sliding window 60 slide over time.

[0193] In this case, the correlation derivation unit 28 can derive the correlation between gas concentration and one or more environmental parameters within the period shown by the width of the sliding window 60, based on multiple gas concentrations and multiple environmental information within the period shown by the width of the sliding window 60, including the period after the period shown by the width of the sliding window 60.

[0194] The updating unit 32 can derive the predicted primary gas concentration at a time point after a unit measurement interval following the period shown by the width of the sliding window 60, based on environmental information at that time point and the correlation within the period shown by the width of the sliding window 60. Furthermore, the updating unit 32 can update the correction information based on the primary gas concentration derived by the gas concentration derivation unit 26 for the environmental information at that time point after a unit measurement interval following the period shown by the width of the sliding window 60, and the predicted primary gas concentration at that time point.

[0195] Furthermore, the gas concentration deriving unit 26 can derive the secondary gas concentration at a time point after a unit measurement interval has elapsed since the period shown by the width of the sliding window 60 has passed, based on the measurement information at that time point, the environmental information at that time point, and the correction information at that time point updated by the updating unit 32.

[0196] Figure 10 An example flowchart of a gas concentration derivation method according to an embodiment is shown. The gas concentration derivation method includes steps S202 to 220.

[0197] The control unit 20 sets the sliding window 60 (S202). Next, the gas concentration measuring unit 12 measures measurement information that corresponds to the gas concentration in the target environment. The environmental information measuring unit 14 measures environmental information that represents one or more environmental parameters in the target environment.

[0198] Measurement information acquisition unit 22 acquires measurement information within the sliding window, and environmental information acquisition unit 24 acquires environmental information (S204). Correlation derivation unit 28 derives the correlation between gas concentration and environmental information based on the gas concentration and environmental information within the sliding window 60 (S206). The correlation can be a regression curve.

[0199] The measurement information acquisition unit 22 acquires measurement information at the measurement time point following the sliding window 60. The environmental information acquisition unit 24 acquires environmental information at the measurement time point following the sliding window 60. In addition, the control unit 20 acquires the stored offset information from the storage unit 42. Thus, the control unit 20 acquires the measurement information, environmental information, and correction information at the measurement time point following the sliding window 60 (S208).

[0200] The gas concentration derivation unit 26 derives the primary gas concentration based on the measurement information, environmental information, and calibration information at the measurement time points following the sliding window 60 (S210). The update unit 32 derives the predicted primary gas concentration at the measurement time points based on the regression curve (S212).

[0201] The update unit 32 determines whether the difference 66 between the predicted primary gas concentration at the measurement time point and the derived primary gas concentration at the measurement time point is below a threshold (S214). If the difference 66 is below the threshold, the process proceeds to S216. On the other hand, if the difference 66 exceeds the threshold, the process proceeds to S220.

[0202] If the difference 66 between the predicted primary gas concentration at the measurement time point and the derived primary gas concentration at the measurement time point is below a threshold, the updating unit 32 updates the correction information based on the difference 66 (S216). The gas concentration derivation unit 26 derives the secondary gas concentration at the measurement time point based on the measurement information, environmental information, and updated correction information at the measurement time point following the sliding window 60 (S218). After that, the gas concentration derivation method processing ends.

[0203] If the difference 66 between the predicted primary gas concentration at the measurement time point and the derived primary gas concentration at the measurement time point exceeds a threshold, the gas concentration deriving unit 26 derives the primary gas concentration at the measurement time point as the secondary gas concentration at the measurement time point (S220). Afterwards, the gas concentration deriving method processing merges with the processing after S118 and ends.

[0204] By using the sensor system 100 or the gas concentration derivation method of this embodiment, even if the measured gas concentration deviates due to changes in environmental information (temperature, humidity, or wind speed, etc.) or deterioration over time, the gas concentration can be corrected using only data representing normal values. In particular, in the sensor system 100 or the gas concentration derivation method of this embodiment, an appropriate correction parameter corresponding to the error at the current time point can be set without distinguishing between various errors caused by different factors. By using this correction parameter, the error can be reduced and the gas concentration can be measured more accurately.

[0205] Figure 11 An example of a computer 1200 capable of embodying the entirety or part of the embodiments of this invention is shown. A program installed on the computer 1200 enables the computer 1200 to function as an operation associated with an apparatus of an embodiment of the present invention, or as one or more "parts" of that apparatus. Alternatively, the program enables the computer 1200 to perform that operation or the one or more "parts". The program enables the computer 1200 to execute a process or a stage of a process of an embodiment of the present invention. Such a program can be executed by a CPU (Central Processing Unit) 1212 to cause the computer 1200 to perform specific operations associated with several or all of the blocks in the flowcharts and block diagrams described in this specification.

[0206] The computer 1200 of this embodiment includes a CPU 1212 and RAM 1214, which are interconnected via a main controller 1210. The computer 1200 also includes a communication interface 1222 and an input / output unit, which are connected to the main controller 1210 via an input / output controller 1220. The computer 1200 also includes a ROM 1230. The CPU 1212 operates according to programs stored in the ROM 1230 and RAM 1214, thereby controlling the various units.

[0207] Communication interface 1222 communicates with other electronic devices via a network. The hard disk drive can store programs and data used by the CPU 1212 in the computer 1200. ROM 1230 stores boot programs and / or programs dependent on the hardware of the computer 1200 that are executed by the computer 1200 when activated. Programs are provided via computer-readable storage media such as CR-ROM (Read-Only Optical Disc), USB memory, or IC cards, or via a network. Programs are installed in RAM 1214 or ROM 1230, examples of computer-readable recording media, and executed by the CPU 1212. The information processing described within these programs is read by the computer 1200, resulting in cooperation between the programs and the various types of hardware resources described above. The apparatus or method can be configured to perform information manipulation or processing according to the use of the computer 1200.

[0208] For example, when communication is performed between computer 1200 and an external device, CPU 1212 can execute a communication program loaded into RAM 1214, and based on the processing described in the communication program, command communication processing is performed on communication interface 1222. Under the control of CPU 1212, communication interface 1222 reads transmission data stored in the transmission buffer area provided in a recording medium such as RAM 1214 or USB memory, and sends the read transmission data to the network, or writes received data received from the network to the receive buffer area provided on the recording medium.

[0209] In addition, the CPU 1212 can enable the RAM 1214 to read all or necessary portions of files or databases stored on external recording media such as USB flash drives, and perform various types of processing on the data in the RAM 1214. Next, the CPU 1212 can write the processed data back to the external recording media.

[0210] Various types of information, such as various types of programs, data, tables, and databases, can be stored in recording media and can undergo information processing. CPU 1212 can perform various types of processing on data read from RAM 1214 and write the results back to RAM 1214. These various types of processing include operations, information processing, conditional judgments, conditional branches, unconditional branches, information retrieval / replacement, etc., specified by a sequence of program instructions as described throughout this disclosure. Furthermore, CPU 1212 can retrieve information from files, databases, etc., within the recording medium. For example, if the recording medium stores multiple entries, each having an attribute value of a first attribute associated with an attribute value of a second attribute, CPU 1212 can retrieve from these multiple entries an entry that specifies the attribute value of the first attribute and matches the condition, read the attribute value of the second attribute stored in that entry, and thereby obtain the attribute value of the second attribute associated with the first attribute that satisfies a predetermined condition.

[0211] The programs or software modules described above can be stored on or near the computer 1200 in a computer-readable storage medium. Alternatively, a recording medium such as a hard disk or RAM provided in a server system connected to a dedicated communication network or the Internet can be used as a computer-readable storage medium, thereby providing the program to the computer 1200 via the network.

[0212] Computer-readable media can include any tangible device capable of storing instructions executable by a suitable device. Therefore, a computer-readable medium having instructions stored therein includes a product comprising instructions that can be executed to create means for performing operations specified in a flowchart or block diagram. Examples of computer-readable media include electronic storage media, magnetic storage media, optical storage media, electromagnetic storage media, semiconductor storage media, etc. More specific examples of computer-readable media include floppy disks (registered trademark), magnetic disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), electrically erasable programmable read-only memory (EEPROM (registered trademark)), static random access memory (SRAM), optical disc read-only memory (CD-ROM), digital versatile disc (DVD), Blu-ray (RTM) disc, memory stick, integrated circuit card, etc.

[0213] Computer-readable instructions may include any of source code or object code described in any combination of one or more programming languages. Source code or object code includes conventional procedural programming languages. Conventional procedural programming languages ​​may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or object-oriented programming languages ​​such as Smalltalk, Java, and C++, as well as the "C" programming language or similar programming languages. Computer-readable instructions may be provided locally or via a wide area network (WAN) such as a local area network (LAN) or the Internet to the processor or programmable circuitry of a general-purpose computer, special-purpose computer, or other programmable data processing device. The processor or programmable circuitry may execute the computer-readable instructions to create means for performing the operations specified in a flowchart or block diagram. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, etc.

[0214] The present invention has been described above using embodiments, but the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various modifications or improvements can be made to the above embodiments. As can be seen from the claims, such modifications or improvements are also included within the technical scope of the present invention.

[0215] It should be noted that the execution order of actions, processes, steps, and stages in the apparatus, system, program, and method shown in the claims, specification, and drawings can be implemented in any order, as long as it is not specifically expressed as "before" or "prior to" and the output of an earlier process is not used in a later process. Even if terms such as "firstly" or "next" are used to describe the flow of actions in the claims, specification, and drawings for convenience, it does not mean that the actions must be performed in that order.

Claims

1. A gas concentration extraction device, comprising: The measurement information acquisition unit acquires measurement information, which represents a measurement value corresponding to the gas concentration in the environment of the object to be measured, as measured by the gas concentration measurement unit. The determination unit determines, based on multiple measurement information within the first period, whether to use the measurement information at the first time point after the first period as reference information when updating the correction parameters. The gas concentration extraction unit, based on the measurement information and the correction parameters, extracts the gas concentration in the environment of the object being measured; and The updating unit updates the correction parameters based on at least one of the reference information.

2. The gas concentration extraction device according to claim 1, wherein, The gas concentration extraction device includes: The environmental information acquisition unit acquires environmental information, which represents one or more environmental parameters in the environment of the measurement target measured by the environmental information measurement unit; and The correlation derivation unit, based on multiple measurement data and environmental information within the first period, derives the correlation between the measurement values ​​and the environmental information within the first period. The determination unit determines whether to use the measurement information at the first time point as the reference information based on the environmental information at the first time point and the correlation during the first period.

3. The gas concentration extraction device according to claim 1, wherein, The correction parameters include at least one of offset information, gain information, and curvature information.

4. The gas concentration extraction device according to claim 2, wherein, The determination unit derives the predicted measurement value at the first time point based on the environmental information at the first time point after the first period and the correlation within the first period. If the difference between the predicted measurement value and the measurement value shown by the measurement information at the first time point is below a threshold, it determines that the measurement information at the first time point is used as the reference information.

5. The gas concentration extraction device according to claim 4, wherein, The threshold is three standard deviations or more of the measured information.

6. The gas concentration extraction device according to claim 2, wherein, The correlation derivation unit derives the correlation between the gas concentration and one or more environmental parameters within the second period, based on multiple measurement data and multiple environmental data within a second period that includes the period following the first period. The determination unit determines whether to use the measurement information at the second time point as reference information when updating the correction parameters, based on the measurement information at the second time point after the second period, the environmental information, and the correlation within the second period.

7. The gas concentration extraction device according to claim 2, wherein, The correlation derivation section derives a machine learning model as the correlation. This machine learning model is obtained by performing machine learning using environmental information as an explanatory variable and measured values ​​as the target variable. The determination unit uses the machine learning model to derive a predicted measurement value for the environmental information at the first time point, and based on the predicted measurement value and the measurement value shown by the measurement information at the first time point, determines whether to use the measurement information at the first time point as reference information when updating the correction parameters.

8. The gas concentration extraction device according to claim 7, wherein, If the difference between the predicted measurement value and the measurement value shown in the measurement information at the first time point is below a threshold, the determination unit determines that the measurement information at the first time point should be used as the reference information.

9. The gas concentration extraction device according to claim 1, wherein, The updating unit derives statistical values ​​based on the measured values ​​shown by each of the multiple reference information, and updates the correction parameters based on the statistical values.

10. The gas concentration extraction device according to claim 9, wherein, The updating unit derives the statistical value based on the measured values ​​shown by the plurality of reference information within a third period shorter than the first period.

11. The gas concentration extraction device according to claim 10, wherein, If the determination unit determines that the measurement information during the third period is used as the reference information for a predetermined number of consecutive times, the update unit derives the statistical value based on the measurement value shown by each of the plurality of reference information during the third period.

12. The gas concentration extraction device according to claim 1, wherein, The gas concentration exporting device also includes a notification unit, which notifies the outside of information based on the determination result of the determination unit.

13. The gas concentration extraction device according to claim 1, wherein, The gas concentration extraction device also includes: The reliability derivation unit derives the reliability of the determination result from the determination unit; and The notification department notifies external parties of information based on the aforementioned reliability.

14. A method for deriving gas concentration, comprising the following stages: Obtain measurement information, which represents the measurement value corresponding to the gas concentration in the environment of the test object as measured by the gas concentration measuring unit; Based on multiple measurement information during the first period, determine whether to use the measurement information at the first time point after the first period as reference information when updating the correction parameters; Based on the measurement information and the correction parameters, the gas concentration in the environment of the measured object is derived; and The correction parameters are updated based on at least one of the reference information.

15. A computer program product comprising a computer program that, when executed by a computer, causes the computer to function as a component: The measurement information acquisition unit acquires measurement information, which represents a measurement value corresponding to the gas concentration in the environment of the object to be measured, as measured by the gas concentration measurement unit. The determination unit determines, based on multiple measurement information within the first period, whether to use the measurement information at the first time point after the first period as reference information when updating the correction parameters. The gas concentration extraction unit, based on the measurement information and the correction parameters, extracts the gas concentration in the environment of the object being measured; and The updating unit updates the correction parameters based on at least one of the reference information.

16. A gas concentration extraction device, comprising: The measurement information acquisition unit acquires measurement information, which represents a measurement value corresponding to the gas concentration in the environment of the object to be measured, as measured by the gas concentration measurement unit. The environmental information acquisition unit acquires environmental information, which represents one or more environmental parameters in the environment of the object of measurement measured by the environmental information measurement unit. The gas concentration derivation unit derives the gas concentration in the environment of the measured object based on the measurement information, the environmental information, and the correction parameters. The correlation derivation unit derives the correlation between the gas concentration and the environmental information during the first period based on multiple gas concentrations and multiple environmental information during the first period; as well as The updating unit, based on the environmental information at a first time point after the first period and the correlation within the first period, derives the predicted primary gas concentration at the first time point, and updates the correction parameters based on the primary gas concentration derived by the gas concentration derivation unit based on the correction parameters for the measurement information at the first time point and the predicted primary gas concentration at the first time point. The gas concentration derivation unit derives the secondary gas concentration at the first time point based on the measurement information at the first time point, the environmental information at the first time point, and the correction parameters at the first time point updated by the update unit.

17. The gas concentration extraction device according to claim 16, wherein, The correction parameters include at least one of offset information, gain information, and curvature information.

18. The gas concentration extraction device according to claim 16, wherein, If the difference between the predicted primary gas concentration and the primary gas concentration derived by the gas concentration derivation unit at the first time point is below a threshold, the updating unit updates the correction parameter based on the difference.

19. The gas concentration extraction device according to claim 16, wherein, The correlation derivation unit derives the correlation between the gas concentration and the one or more correction parameters within the second period, based on multiple gas concentrations and multiple environmental information within a second period that includes the period following the first period. The updating unit derives the predicted primary gas concentration at the second time point based on the environmental information at the second time point after the second period and the correlation within the second period. It then updates the correction parameters based on the primary gas concentration derived by the gas concentration derivation unit for the environmental information at the second time point and the predicted primary gas concentration at the second time point. The gas concentration derivation unit derives the secondary gas concentration at the second time point based on the measurement information at the second time point, the environmental information at the second time point, and the correction parameters at the second time point updated by the update unit.

20. The gas concentration extraction device according to claim 19, wherein, The correlation derivation unit derives a machine learning model as the correlation. The machine learning model is obtained by performing machine learning using environmental information as an explanatory variable and the gas concentration derived by the gas concentration derivation unit as the target variable. The updating unit uses the machine learning model to derive a predicted primary gas concentration for the environmental information at the first time point, and updates the correction parameters based on the predicted primary gas concentration and the primary gas concentration derived by the gas concentration deriving unit at the first time point.

21. The gas concentration extraction device according to claim 20, wherein, If the difference between the predicted primary gas concentration and the primary gas concentration derived by the gas concentration derivation unit at the first time point is below a threshold, the updating unit updates the correction parameter based on the difference.

22. A method for deriving gas concentration, comprising the following stages: Obtain measurement information, which represents the measurement value corresponding to the gas concentration in the environment of the test object as measured by the gas concentration measuring unit; Obtain environmental information, which represents one or more environmental parameters in the environment of the object of measurement measured by the environmental information measurement unit; Based on the measurement information, the environmental information, and the correction parameters, the gas concentration in the environment of the measured object is derived. Based on multiple gas concentrations and multiple environmental information during the first period, the correlation between the gas concentrations and the environmental information during the first period is derived. as well as Based on the primary gas concentration derived from the measurement information at a first time point after the first period, the primary gas concentration derived from the correction parameters, the environmental information, and the correlation within the first period, a predicted primary gas concentration at the first time point is derived. The correction parameters are then updated based on the primary gas concentration at the first time point derived during the gas concentration derivation phase and the predicted primary gas concentration at the first time point. The stage of deriving the gas concentration includes the following stages: deriving the secondary gas concentration at the first time point based on the measurement information at the first time point, the environmental information at the first time point, and the correction parameters at the first time point after being updated during the update stage.

23. A computer program product comprising a computer program that, when executed by a computer, causes the computer to function as a component: The measurement information acquisition unit acquires measurement information, which represents a measurement value corresponding to the gas concentration in the environment of the object to be measured, as measured by the gas concentration measurement unit. The environmental information acquisition unit acquires environmental information, which represents one or more environmental parameters in the environment of the object of measurement measured by the environmental information measurement unit. The gas concentration derivation unit derives the gas concentration in the environment of the measured object based on the measurement information, the environmental information, and the correction parameters. The correlation derivation unit derives the correlation between the gas concentration and the environmental information during the first period based on multiple gas concentrations and multiple environmental information during the first period; as well as The updating unit, based on the primary gas concentration derived by the gas concentration derivation unit based on the measurement information at a first time point after the first period, the environmental information, and the correlation within the first period, derives the predicted primary gas concentration at the first time point, and updates the correction parameters based on the primary gas concentration derived by the gas concentration derivation unit at the first time point and the predicted primary gas concentration at the first time point. The gas concentration derivation unit derives the secondary gas concentration at the first time point based on the measurement information at the first time point, the environmental information at the first time point, and the correction parameters at the first time point updated by the update unit.

24. A sensor system comprising: Gas concentration extraction device according to any one of claims 1 to 13 or claims 16 to 21; The gas concentration output section; and The environmental information measurement unit measures environmental information, which represents one or more environmental parameters in the environment of the object being measured.

25. The sensor system according to claim 24, wherein, The sensor system includes: The housing includes at least a detection unit of the gas concentration measuring unit, which takes in air from the environment of the object being measured; and The ventilation section ventilates the interior of the housing. The time until 90% of the volume of the housing is ventilated by diffusion is less than 60 seconds, and the duration of the first period is longer than the time until the ventilation is carried out.

Images