A gas concentration detection method based on double model fitting correction

Abstract

The application discloses a kind of gas concentration detection methods based on double model fitting correction, it is related to integrated optoelectronic field, specifically includes: S1: in the demodulation process of tunable semiconductor laser absorption spectroscopy technology, background noise second harmonic signal is fitted, and background noise second harmonic function is established;S2: using Lambert-Beer law, establish theoretical gas absorption second harmonic function;S3: obtain original gas absorption second harmonic signal;S4: in combination with background noise second harmonic function in S1 and theoretical gas absorption second harmonic function in S2, further fitting is carried out to original gas absorption second harmonic signal, and corrected gas absorption second harmonic signal is obtained;S5: repeat detection different gas concentration, realize the calibration of signal amplitude and gas concentration relationship.This method separates background noise and gas absorption signal by software algorithm, significantly improves detection precision, stability and universality, while reducing hardware dependence.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of integrated optoelectronics, and in particular to a gas concentration detection method based on double model fitting correction. BACKGROUND

[0002] The optical gas sensor based on Tunable Diode Laser Absorption Spectroscopy (TDLAS) technology can detect the composition and concentration of gas in the environment according to the absorption characteristics of gas molecules to specific wavelength light, and has the advantages of high sensitivity, good selectivity, fast response speed, etc. However, when using Wavelength Modulation Spectroscopy (WMS) technology for trace gas detection, the interference of the second harmonic background signal will reduce the measurement accuracy. Moreover, during long-term continuous detection, the second harmonic background signal will change with the test environment, affecting the accuracy of gas concentration inversion and the stability of the system, limiting the stability and reliability of the TDLAS-WMS system in practical applications.

[0003] The existing TDLAS gas absorption signal correction method has the following problems: (1) it relies on a complex measurement system, which is costly; (2) the software correction effect is limited, and the accuracy improvement is not obvious; (3) some methods are only suitable for specific measurement systems and gas types, and the application scene is limited, lacking universality. SUMMARY

[0004] To solve the above-mentioned problems, the present application provides a gas concentration detection method based on double model fitting correction, which overcomes the problems of hardware dependence, high cost, limited software correction accuracy, limited application scene and lack of universality in traditional gas absorption signal correction methods, and realizes the overall improvement of gas concentration detection in precision, stability, universality and real-time, which has important significance for industrial emission monitoring, environmental air quality monitoring and gas sensing needs in complex scenes.

[0005] To achieve the above-mentioned purpose, the present application provides a gas concentration detection method based on double model fitting correction, which specifically includes the following steps:

[0006] Step S1: In the Tunable Diode Laser Absorption Spectroscopy (TDLAS) demodulation process, the background noise second harmonic signal is fitted, and a background noise second harmonic function is established;

[0007] Step S2: Using the Lambert-Beer law, a theoretical gas absorption second harmonic function is established;

[0008] Step S3: Obtain the original gas absorption second harmonic signal;

[0009] Step S4: further fitting the original gas absorption second harmonic signal by combining the background noise second harmonic function in step S1 and the theoretical gas absorption second harmonic function in step S2, to obtain a corrected gas absorption second harmonic signal;

[0010] Step S5: repeating the detection of different gas concentrations to realize the calibration of the relationship between the signal amplitude and the gas concentration.

[0011] Preferably, in step S1, the following steps are specifically included:

[0012] S11: using tunable semiconductor laser absorption spectroscopy technology, under wavelength modulation, multiple measurements are performed to obtain direct absorption spectrum signals in the case of no target detection gas and to perform lock-in amplification, to obtain multiple sets of background noise second harmonic signals;

[0013] S12: function fitting is performed on the multiple sets of background noise second harmonic signals obtained in step S11, to establish a background noise second harmonic function S noise (a i ,b i ,c i ); wherein a i , b i and c i represent amplitude, angular frequency and phase respectively.

[0014] Preferably, in step S2, the theoretical gas absorption second harmonic function S theory (m) is established, wherein m is the gas concentration.

[0015] Preferably, in step S3, tunable semiconductor laser absorption spectroscopy technology is used, under wavelength modulation, to obtain direct absorption spectrum signals in the case of target detection gas and to perform lock-in amplification, to obtain an original gas absorption second harmonic signal S int .

[0016] Preferably, in step S4, the background noise second harmonic function S noise (a i ,b i ,c i ) in step S1 and the theoretical gas absorption second harmonic function S theory (m) in step S2 are used to fit the original gas absorption second harmonic signal S int in step S3 within a determined a i , b i and c i value range, S int =k*S noise (a i ,b i ,ci )+l*S theory (m), wherein k and l are weight values, and a i , b i , c i , k, l, m, and the final corrected gas absorption second harmonic signal is represented as S correct =S theory (m).

[0017] Preferably, in step S5, steps S1-S4 are repeated to obtain corrected gas absorption second harmonic signals S correct at different gas concentrations, to obtain the gas concentration m.

[0018] Preferably, in step S11, the wavelength range of the direct absorption spectrum signal is in the visible, infrared, or terahertz range.

[0019] Preferably, in steps S11 and S3, the phase-locked amplification operation adopts any one of hardware phase-locked or software phase-locked, analog phase-locked or digital phase-locked, and single-phase phase-locked or quadrature phase-locked.

[0020] Preferably, in step S12, the function fitting process adopts one or a combination of exponential function, sine wave function, and cosine wave function.

[0021] Preferably, in step S4, the absorption spectral line of the gas molecule used in the calculation of the established gas absorption second harmonic model is one or a combination of Lorentz line type, Gaussian line type, and Voigt line type.

[0022] Therefore, the gas concentration detection method based on double model fitting correction has the following beneficial effects:

[0023] (1) The gas concentration detection method based on double model fitting correction has high precision and high stability. On the one hand, by respectively constructing a background noise second harmonic model and a gas absorption second harmonic model, and combining a double model fitting correction algorithm to correct the original gas absorption signal, the background noise interference is effectively eliminated, and the accuracy of gas concentration detection is significantly improved. On the other hand, by respectively modeling the background noise and the gas absorption signal independently, the measurement error caused by the background signal drift in the traditional method is avoided, further enhancing the stability and reliability of the system in long-time continuous detection.

[0024] (2) The present application does not need to rely on complex hardware devices, thereby reducing the system cost and maintenance difficulty, and the separation and correction of the background noise and the gas absorption signal are realized through a software algorithm, reducing the dependence on expensive hardware, making the system more economical and efficient.

[0025] (3) The application has strong universality and is suitable for detection of various gas types and different waveband ranges. The method supports direct absorption spectrum signal collection in the visible light, infrared and terahertz wavebands, and is compatible with various phase-locked amplification methods, so it can be widely applied to different types of gas detection scenes. In addition, the background noise second harmonic function and the gas absorption second harmonic function both adopt a flexible function fitting strategy, which can select appropriate model combinations according to actual needs, thereby adapting to different environmental conditions and gas characteristics, and showing high universality.

[0026] (4) The gas concentration detection method proposed in the application can dynamically respond to changes in the test environment, thereby improving the long-term stability of the system. In actual application, factors such as temperature, pressure and light source state of the test environment may change over time, thereby affecting the distribution characteristics of the background noise signal. By fitting the background noise second harmonic model in real time and combining the calculated gas absorption second harmonic model for correction, the interference caused by these environmental changes can be dynamically compensated, ensuring the consistency and reliability of the measurement results. This method is particularly suitable for application scenarios that require long-term online monitoring.

[0027] (5) The application helps to promote the development of integrated optoelectronic sensing technology. The double-model fitting correction algorithm used in the method is easy to integrate with existing TDLAS-WMS systems, and its software-led design concept lays the foundation for the research and development of future intelligent and small-sized gas sensors. By transferring complex signal processing tasks to the algorithm level, it provides a new idea for the popularization and promotion of high-performance gas sensing technology.

[0028] The technical solutions of the application will be described in further detail below with the aid of the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 is a flowchart of a gas concentration detection method based on double-model fitting correction according to the application;

[0030] Figure 2 is a TDLAS-WMS system structure diagram in Example 1 of the application;

[0031] Figure 3 is a background noise second harmonic signal amplitude and background noise second harmonic signal function fitting result diagram in Example 1 of the application;

[0032] Figure 4 is a raw gas absorption second harmonic signal amplitude and double-model fitting result diagram in Example 1 of the application;

[0033] Figure 5 is a corrected gas absorption second harmonic signal amplitude and feature point extraction diagram in Example 1 of the application;

[0034] Figure 6 A relationship diagram of the corrected gas absorption second harmonic signal characteristic value and the target gas concentration in the embodiment one of the present application;

[0035] Figure 7 A comparison diagram of the Allan variance of the gas concentration detection method without using and using the results of the double model fitting correction in the embodiment one of the present application;

[0036] Figure 8 A background noise second harmonic signal amplitude and background noise second harmonic signal function fitting result diagram in the embodiment two of the present application;

[0037] Figure 9 A relationship diagram of the corrected gas absorption second harmonic signal characteristic value and the target gas concentration in the embodiment two of the present application;

[0038] Figure 10 A comparison diagram of the Allan variance of the gas concentration detection method without using and using the results of the double model fitting correction in the embodiment two of the present application. DETAILED DESCRIPTION

[0039] The technical solutions of the present application are further described below by means of the accompanying drawings and embodiments.

[0040] Unless otherwise defined, the technical terms or scientific terms used in the present application shall be understood as the usual meanings understood by those skilled in the art to which the present application belongs.

[0041] The "including" or "containing" and similar words used in the present application mean that the elements before the word cover the elements listed after the word, and do not exclude the possibility of also covering other elements. The terms "in", "out", "up", "down" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. When the absolute position of the described object changes, the relative positional relationship may also change accordingly. In the present application, unless otherwise specified and limited, the term "attached" and the like should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or it can be integrated; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. For those skilled in the art, the specific meaning of the above-mentioned terms in the present application can be understood according to the specific circumstances.

[0042] EMBODIMENT

[0043] As Figure 1As shown, a gas concentration detection method based on double model fitting correction, specifically comprising the following steps:

[0044] Step S1: In the demodulation process of tunable semiconductor laser absorption spectrum technology, the background noise second harmonic signal is fitted to establish the background noise second harmonic function;

[0045] In step S1, specifically comprising the following steps:

[0046] S11: Using tunable semiconductor laser absorption spectrum technology, under wavelength modulation, multiple measurements are taken to obtain direct absorption spectrum signals without target detection gas and lock-in amplification to obtain multiple sets of background noise second harmonic signals; In step S11, the waveband range of the direct absorption spectrum signal is visible, infrared or terahertz waveband.

[0047] S12: Function fitting is performed on the multiple sets of background noise second harmonic signals obtained in step S11 to establish the background noise second harmonic function S noise (a i ,b i ,c i ); wherein a i , b i and c i represent amplitude, angular frequency and phase respectively. In step S12, one or more combinations of exponential function, sine wave function and cosine wave function are used in the function fitting process.

[0048] Step S2: Establishing a theoretical gas absorption second harmonic function using the Lambert-Beer law;

[0049] In step S2, the theoretical gas absorption second harmonic function S theory (m), wherein m is the gas concentration.

[0050] Step S3: Obtain the original gas absorption second harmonic signal;

[0051] In step S3, using tunable semiconductor laser absorption spectrum technology, under wavelength modulation, the direct absorption spectrum signal with target detection gas is obtained and lock-in amplified to obtain the original gas absorption second harmonic signal S int .

[0052] In step S11 and step S3, the operation of lock-in amplification uses any one of hardware lock-in or software lock-in, analog lock-in or digital lock-in, single-phase lock-in or quadrature lock-in.

[0053] Step S4: combining the background noise second harmonic function in step S1 with the theoretical gas absorption second harmonic function in step S2, the original gas absorption second harmonic signal is further fitted to obtain a corrected gas absorption second harmonic signal;

[0054] In step S4, the background noise second harmonic function S noise (a i ,b i ,c i ) in step S1 and the theoretical gas absorption second harmonic function S theory (m) in step S2 are used to fit the original gas absorption second harmonic signal S i in step S3 in the determined a i , b i and c int value range, S int =k*S noise (a i ,b i ,c i )+l*S theory (m), wherein k and l are weight values, to obtain the values of a i , b i , c i , k, l, m, and the final corrected gas absorption second harmonic signal is represented as S correct =S th eory(m).

[0055] In step S4, the absorption spectral line of the gas molecule used in the established gas absorption second harmonic model is one or a combination of more than one of the Lorentz line type, the Gaussian line type and the Voigt line type.

[0056] Step S5: repeating the detection of different gas concentrations to realize the calibration of the relationship between the signal amplitude and the gas concentration. In step S5, steps S1-S4 are repeated, and under different gas concentrations, the corrected gas absorption second harmonic signal is obtained as S correct , and the gas concentration is m.

[0057] Example One

[0058] As Figure 2The experiment was carried out at room temperature, and a distributed feedback (DFB) laser with a center wavelength of 1651 nm was selected as the light source of the TDLAS system. The light signal output by the laser driver board was a high-frequency sinusoidal wave with a period of 0.26 seconds and a frequency of 7820 Hz. The light signal passed through a collimating lens and entered a Herriott-type multi-reflection gas absorption cell with an optical path length of 3 meters. Then the light signal was received by a photodetector and converted into an electrical signal. The target gas was 3.5% methane gas, and pure nitrogen was used as the zero gas to provide an interference-free background reference. The signal output by the photodetector was collected by a data acquisition card at a sampling rate of 80 KHz and transmitted to a computer for subsequent processing.

[0059] To simulate the detection conditions in the actual application environment, first, nitrogen was introduced into a Herriott-type multi-reflection gas absorption cell with an optical path length of 3 meters, and the DFB laser controlled by the laser driver board was turned on. The direct absorption spectrum signal under the zero gas state was read by the photodetector for 60 seconds, and transmitted to the computer end by the data acquisition card. On the computer end, the signal was traversed to find the starting index that meets the threshold condition (the signal amplitude exceeds 0.02 mV), and time calibration was performed to make the time axis start from zero. Then, the long-time signal was divided and aligned by period. The time stamps of the even-numbered crossing points were extracted, and the time difference between adjacent crossing points was calculated to determine the signal period. According to the period division result, the long-time background signal was divided into several period segments, the shortest period length was found, and the sampling points of each period were intercepted to ensure that all period signals were aligned.

[0060] Based on the MATLAB platform, the second harmonic signal extraction was realized. The specific steps are as follows: read the signal and obtain the frequency (f) of the direct absorption spectrum signal by fast Fourier transform, and construct a sinusoidal reference signal (frequency 2f) synchronized with the input signal. Multiply the input signal and the sinusoidal reference signal to obtain the multiplied signal, and perform low-pass filtering (use FIR filter with Kaiser window function) to extract the low-frequency component of the signal. After lock-in amplification, the second harmonic signal of the background noise is obtained, denoted as S noise (t).

[0061] To extract the periodic characteristics, a multi-frequency sinusoidal superposition function was used to fit the background signal. In this embodiment, the mathematical function model of the background signal is defined as a multi-frequency sinusoidal superposition form:

[0062]

[0063] where n represents the number of trigonometric functions used for fitting (n=7 in this embodiment), a i represents the amplitude of the i-th frequency component, and bi represents the angular frequency of the i-th frequency component, c i represents the phase shift of the i-th frequency component. Considering the slight changes of each group of a i , b i , c i , the slight perturbations of the second harmonic signal of the background noise will be caused, so that the fitting can accurately obtain the second harmonic signal of the background noise in the current state. According to the experimental experience and ensuring that the fitting result has physical meaning, the appropriate initial parameters and constraint conditions are defined, and the lsqcurvefit function in MATLAB is used to perform nonlinear least squares fitting on the background signal Noise(t), and the result is shown in Figure 3 . The specific values of the parameters a i , b i and c i obtained by fitting are shown in Table 1. Subsequently, the fitting result is verified, and the result shows that the adjusted R 2 value is 0.99873, the root mean square error (RMSE) is 1.7953x10 -5 , and the mean absolute error (MAE) is 1.4636x10 -5 . These indicators show that the fitting result has very high precision and reliability. Finally, through the above fitting process, the periodic characteristics of the background signal are successfully extracted.

[0064] Table 1

[0065]

[0066] In the presence of target gas, the tunable semiconductor laser absorption spectroscopy technology is used again, and the signal-to-noise ratio is further improved through wavelength modulation and lock-in amplification processing, so as to obtain the original gas absorption second harmonic signal S int (t). This signal processing method is consistent with the processing method of the background noise signal, including time calibration, period segmentation and alignment, and second harmonic signal extraction steps. Based on the Lambert-Beer law, combined with the modulation laser absorption spectroscopy technology, a theoretical gas absorption second harmonic function S theory (t) is established to describe the absorption characteristics of the target gas at different concentrations. The absorption spectral line of the gas molecule used in this embodiment is Lorentz line type, and its mathematical expression is as follows:

[0067]

[0068] where I0represents the initial light intensity of the incident laser, a represents the absorption line strength of the target gas molecules, L represents the effective optical path length of the gas absorption cell, P represents the number density of the target gas molecules, g represents the full width at half maximum of the Lorentzian line shape, f(t) represents the variation of the laser output wavenumber with time, and f0represents the center wavenumber of the target gas absorption peak. Further, the original gas absorption second-harmonic signal S int satisfies the following expression:

[0069] S int (t) = S noise (t) + S theory (t) (3)

[0070] The periodic characteristics of the background noise and the absorption characteristics of the target gas are comprehensively considered, which can comprehensively reflect the physical essence of the experimental data. Similarly, according to experimental experience and ensuring that the fitting results have physical meaning, appropriate initial parameters and constraint conditions are defined, and the initial second-harmonic signal S int (t) of the target gas is nonlinear least squares fitted using the lsqcurvefit function in MATLAB. In the fitting process, S int (t) is decomposed into a background noise component S noise (t) and a target gas absorption component S theory (t), so as to remove the influence of the background noise, extract the pure target gas absorption signal, and the results are shown in Figure 4 . Figure 4 The average value curve of a representative original gas absorption second-harmonic signal and its corresponding fitting curve are plotted in FIG. 2. The specific values of the parameters a i , b i and c i obtained by fitting are shown in Table 2. The fitting results show that the adjusted R 2 value is 0.999, the root mean square error (RMSE) is 3.382 x 10 -5 , and the mean absolute error (MAE) is 2.701 x 10 -5 . These indicators show that the fitting results have very high precision and reliability, verifying the effectiveness of the method in removing background noise and extracting target gas absorption signals.

[0071] Table 2

[0072]

[0073] After the above fitting operation is completed, the background noise signal is subtracted from the test signal using the fitting result S noise_fit (t) to obtain the corrected absorption signal:

[0074] S correct (t) = Stheory_fit (t) = S int_fit (t) - S noise_fit (t) (4)

[0075] where S correct (t) represents the corrected gas absorption second-harmonic signal, effectively removing the interference of background noise, and retaining the absorption characteristics of the target gas.

[0076] To further analyze S correct (t), feature point extraction was performed. Specifically, by traversing the S correct (t) signal, the local maximum (i.e., peak) within each cycle was extracted, denoted as Max, and its left and right local minimum (i.e., valley), denoted as Left min and Right min , were also extracted, and their amplitudes and time positions were recorded, as shown in Figure 5 Subsequently, according to the extracted feature points, the characteristic value AMP of each cycle was calculated:

[0077]

[0078] Next, a series of methane gas samples with different known concentrations were configured by the mass flow controller, and the characteristic value AMP of each cycle was measured under the same conditions. Specifically, by adjusting the concentration of methane gas entering the gas absorption cell, the AMP value at each concentration was measured from low to high. Pure nitrogen was used to flush the system before each measurement to ensure the consistency of the background signal. The AMP values corresponding to different concentrations were recorded, and a data set containing multiple data points was constructed, as shown in Table 3. Then, a scatter plot of these data points was drawn with the methane gas concentration m as the horizontal coordinate and the characteristic value AMP as the vertical coordinate. It was found through analysis that these data points showed a clear linear relationship. Based on the linear relationship in the above scatter plot, a best-fit straight line was fitted using the least squares method, which reflected the quantitative relationship between the characteristic value AMP and the concentration of methane gas, as shown in Figure 6 The linear fitting goodness was greater than 0.99. This calibration curve not only provides a direct conversion method from the characteristic value to the concentration of methane gas, but also allows the actual concentration of methane gas in the unknown sample to be accurately calculated according to the new characteristic value. For the known concentration of 686 ppm of methane gas passed in, the average value of the characteristic value AMP was 0.00448, and the corresponding measured concentration reading was 683.818 ppm, indicating the accuracy of the detection method.

[0079] Table 3

[0080]

[0081] Finally, the comparison of the Allan variance results obtained by using the gas concentration detection method without and with the dual-model fitting correction method is shown, such as Figure 7 As shown in the figure, Allan variance is an important indicator for measuring the stability and accuracy of a measurement system. The smaller its value, the lower the system noise level and the higher the detection accuracy. For traditional detection methods (i.e., methods that do not use dual-model fitting correction), the failure to effectively separate background noise signals results in a low signal-to-noise ratio. In particular, its sensitivity and accuracy are significantly limited in low-concentration gas detection. The minimum value of its Allan variance occurs at an integration time of 2.354 seconds, corresponding to a value of 0.596 ppm. However, after introducing dual-model fitting correction, the performance of the Allan variance is significantly improved. Specifically, the minimum value of the Allan variance after dual-model fitting correction drops to 0.162 ppm at an integration time of 7.325 seconds, a reduction of 72.91% compared to the traditional method. This comparison shows that the method based on dual-model fitting correction has made a significant contribution to improving the accuracy and sensitivity of gas concentration detection. This method not only effectively suppresses noise interference, but also significantly reduces the detection limit, thereby more accurately capturing subtle changes in gas concentration.

[0082] Example 2

[0083] The experiment was carried out at room temperature, with the same structure as the TDLAS system used in Example 1. The TDLAS system uses a DFB laser with a central wavelength of 2004nm as a light source. It is controlled by the laser driver board and outputs an optical signal with a period of 0.1 seconds and a high-frequency sine wave frequency of 2000Hz. After passing through a collimating lens, the optical signal enters a Herriott-type multi-reflection gas absorption cell with an optical path length of 3 meters, and is then received by a photodetector and converted into an electrical signal. The target gas is pure carbon dioxide (concentration of 100%), and pure nitrogen is used as the zero gas to provide an interference-free background reference. The signal output by the photodetector is collected by a data acquisition card at a sampling rate of 100kHz and transmitted to the computer for subsequent processing.

[0084] To simulate the detection conditions in actual application environments, nitrogen was first introduced into the gas absorption cell and the DFB laser was turned on. A photodetector was used to read the direct absorption spectrum signal in the zero-gas state for 20 seconds and transmitted it to the computer via a data acquisition card. On the computer, the signal was time-calibrated, period-divided, and aligned. Then, the second harmonic signal was extracted based on the MATLAB platform to obtain the second harmonic signal S of the background noise. noise (t).

[0085] To extract the periodic features, a multi-frequency sinusoidal superposition function is used to fit the background signal. In this embodiment, the mathematical model of the background signal is consistent with that of Embodiment 1, and a multi-frequency sinusoidal superposition form is used, as shown in equation (1). Wherein, n represents the number of trigonometric functions used for fitting (n = 8 in this embodiment), a i , b i , and c i represent the amplitude, angular frequency, and phase shift, respectively. The specific values of the fitted parameters a i , b i , and c i are shown in Table 4, and the fitting result is shown in Figure 8 . The fitting result shows that the adjusted R 2 value is 0.9997, the root mean square error (RMSE) is 1.6806 x 10 -6 , and the mean absolute error (MAE) is 8.571 x 10 -7 . These indicators show that the fitting result has very high precision and reliability.

[0086] Table 4

[0087]

[0088] In the presence of the target gas, the TDLAS technology is used again to obtain the original gas absorption second harmonic signal S int (t) through wavelength modulation and lock-in amplification processing. Based on the Lambert-Beer law, combined with the modulation laser absorption spectrum technology, a theoretical gas absorption second harmonic function S theory (t) is established to describe the absorption characteristics of carbon dioxide gas at different concentrations (the same as in Embodiment 1, the expression is shown in equation (2)). In addition, the original gas absorption second harmonic signal S int is defined, which is consistent with that in Embodiment 1, and the expression is shown in equation (3).

[0089] Similarly, in order to ensure that the fitting result has physical meaning, and combined with experimental experience, the lsqcurvefit function in MATLAB is used to perform nonlinear least squares fitting on the initial second harmonic signal S int (t) of the target gas. After the above fitting operation is completed, the fitting result S noise_fit (t) is used to subtract the background noise signal from the test signal to obtain the corrected absorption signal S correct (t) (consistent with Embodiment 1, the expression is shown in equation (4)). In order to further analyze S correct(t), the feature points were extracted, and the feature value AMP of each cycle was calculated (consistent with Embodiment 1, the expression is shown in equation (5)). Then, a series of carbon dioxide gas samples with different known concentrations were configured by the mass flow controller, and the feature value AMP of each cycle was measured under the same conditions. The AMP values corresponding to different concentrations were recorded, and a data set containing multiple data points was constructed, as shown in Table 5.

[0090] Table 5

[0091]

[0092] The scatter plot of these data points was drawn with the carbon dioxide gas concentration m as the abscissa and the feature value AMP as the ordinate. Analysis found that these data points showed a clear linear relationship. Based on the linear relationship in the above scatter plot, a best straight line was fitted using the least squares method, which reflected the quantitative relationship between the feature value AMP and the carbon dioxide gas concentration, as shown in equation (6). Figure 9

[0093] Finally, the Allan variance results obtained by the gas concentration detection method without and with the double-model fitting correction were compared, as shown in equation (7). Figure 10 For the traditional detection method, the minimum value of the Allan variance appeared when the integration time was 1 second, and the corresponding numerical value was 3.2952 ppm. However, after introducing the double-model fitting correction, the performance of the Allan variance was significantly improved. Specifically, the minimum value of the Allan variance after the double-model fitting correction decreased to 1.0378 ppm at an integration time of 1.1 seconds, which was reduced by 68.51% compared with the traditional method.

[0094] Therefore, the present application adopts the above-mentioned gas concentration detection method based on double-model fitting correction, which comprehensively improves the precision, stability, universality and real-time performance of gas concentration detection, and has important significance for industrial emission monitoring, environmental air quality monitoring and gas sensing needs in complex scenarios.

[0095] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application and not to limit them, although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that: it can still modify or equivalently replace the technical solutions of the present application, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present application.​

Claims

1. A method for detecting a gas concentration based on a dual-model fitting correction, characterized in that: Specifically comprising the following steps: Step S1: in the demodulation process of tunable semiconductor laser absorption spectroscopy technology, the background noise second harmonic signal is fitted, and a background noise second harmonic function is established; Step S2: using Lambert-Beer law, a theoretical gas absorption second harmonic function is established; Step S3: obtaining the original gas absorption second harmonic signal; Step S4: combining the background noise second harmonic function in step S1 with the theoretical gas absorption second harmonic function in step S2, the original gas absorption second harmonic signal is further fitted to obtain the corrected gas absorption second harmonic signal; In step S4, the background noise second harmonic function in step S1 and the theoretical gas absorption second harmonic function in step S2 are fitted to the raw gas absorption second harmonic signal in step S3 , and in a determined , , and numerical range, to obtain , , the numerical value of the final corrected gas absorption second harmonic signal is represented as ; wherein , and represent the amplitude, angular frequency and phase, is the gas concentration; Step S5: repeating the detection of different gas concentrations to realize the calibration of the relationship between signal amplitude and gas concentration.

2. The method according to claim 1, wherein the method is characterized by: In step S1, the following steps are specifically included: S11: using tunable semiconductor laser absorption spectroscopy technology, under wavelength modulation, multiple measurements are performed to obtain direct absorption spectrum signals in the absence of target detection gas and are subjected to lock-in amplification to obtain multiple sets of background noise second harmonic signals; S12: function fitting is performed on the plurality of sets of background noise second harmonic signals obtained in step S11 to establish a background noise second harmonic function . 3.The gas concentration detection method based on double model fitting correction according to claim 2, characterized in that: In step S3, the tunable semiconductor laser absorption spectrum technology is adopted, in the wavelength modulation case, the direct absorption spectrum signal of the target detection gas is obtained and is subjected to lock-in amplification, to obtain the original gas absorption second harmonic signal .

4. The gas concentration detection method based on double model fitting correction according to claim 3, characterized in that: In step S5, steps S1-S4 are repeated to obtain the corrected gas absorption second harmonic signal as a function of the gas concentration , for different gas concentrations .

5. The method according to claim 2, wherein the method is characterized by: In step S11, the wavelength range of the direct absorption spectrum signal is visible light, infrared or terahertz band.

6. The method according to claim 4, wherein the method is characterized by: In steps S11 and S3, the operation of lock-in amplification adopts any one of hardware lock-in or software lock-in, analog lock-in or digital lock-in, single-phase lock-in or quadrature lock-in.

7. The method according to claim 2, wherein the method is characterized by: In step S12, the function fitting process adopts one or a combination of exponential function, sine wave function and cosine wave function.

8. The gas concentration detection method based on double model fitting correction according to claim 5, characterized in that: In step S4, the absorption spectral line of the gas molecule used in the calculation of the established gas absorption second harmonic model is one or a combination of Lorentz line type, Gaussian line type and Voigt line type.

Images